sanchayanmaity/gem5
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity
gem5/arch/alpha/isa/mem.isa

685 lines
19 KiB
Plaintext
Raw Normal View History

Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
// -*- mode:c++ -*-
// Copyright (c) 2003-2005 The Regents of The University of Michigan
// All rights reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met: redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer;
// redistributions in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer in the
// documentation and/or other materials provided with the distribution;
// neither the name of the copyright holders nor the names of its
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
output header {{
/**
* Base class for general Alpha memory-format instructions.
*/
class Memory : public AlphaStaticInst
{
protected:
/// Memory request flags. See mem_req_base.hh.
unsigned memAccessFlags;
/// Pointer to EAComp object.
const StaticInstPtr<AlphaISA> eaCompPtr;
/// Pointer to MemAcc object.
const StaticInstPtr<AlphaISA> memAccPtr;
/// Constructor
Memory(const char *mnem, MachInst _machInst, OpClass __opClass,
StaticInstPtr<AlphaISA> _eaCompPtr = nullStaticInstPtr,
StaticInstPtr<AlphaISA> _memAccPtr = nullStaticInstPtr)
: AlphaStaticInst(mnem, _machInst, __opClass),
memAccessFlags(0), eaCompPtr(_eaCompPtr), memAccPtr(_memAccPtr)
{
}
std::string
generateDisassembly(Addr pc, const SymbolTable *symtab) const;
public:
const StaticInstPtr<AlphaISA> &eaCompInst() const { return eaCompPtr; }
const StaticInstPtr<AlphaISA> &memAccInst() const { return memAccPtr; }
};
/**
* Base class for memory-format instructions using a 32-bit
* displacement (i.e. most of them).
*/
class MemoryDisp32 : public Memory
{
protected:
/// Displacement for EA calculation (signed).
int32_t disp;
/// Constructor.
MemoryDisp32(const char *mnem, MachInst _machInst, OpClass __opClass,
StaticInstPtr<AlphaISA> _eaCompPtr = nullStaticInstPtr,
StaticInstPtr<AlphaISA> _memAccPtr = nullStaticInstPtr)
: Memory(mnem, _machInst, __opClass, _eaCompPtr, _memAccPtr),
disp(MEMDISP)
{
}
};
/**
* Base class for a few miscellaneous memory-format insts
* that don't interpret the disp field: wh64, fetch, fetch_m, ecb.
* None of these instructions has a destination register either.
*/
class MemoryNoDisp : public Memory
{
protected:
/// Constructor
MemoryNoDisp(const char *mnem, MachInst _machInst, OpClass __opClass,
StaticInstPtr<AlphaISA> _eaCompPtr = nullStaticInstPtr,
StaticInstPtr<AlphaISA> _memAccPtr = nullStaticInstPtr)
: Memory(mnem, _machInst, __opClass, _eaCompPtr, _memAccPtr)
{
}
std::string
generateDisassembly(Addr pc, const SymbolTable *symtab) const;
};
}};
output decoder {{
std::string
Memory::generateDisassembly(Addr pc, const SymbolTable *symtab) const
{
return csprintf("%-10s %c%d,%d(r%d)", mnemonic,
flags[IsFloating] ? 'f' : 'r', RA, MEMDISP, RB);
}
std::string
MemoryNoDisp::generateDisassembly(Addr pc, const SymbolTable *symtab) const
{
return csprintf("%-10s (r%d)", mnemonic, RB);
}
}};
def format LoadAddress(code) {{
iop = InstObjParams(name, Name, 'MemoryDisp32', CodeBlock(code))
header_output = BasicDeclare.subst(iop)
decoder_output = BasicConstructor.subst(iop)
decode_block = BasicDecode.subst(iop)
exec_output = BasicExecute.subst(iop)
}};
def template LoadStoreDeclare {{
/**
* Static instruction class for "%(mnemonic)s".
*/
class %(class_name)s : public %(base_class)s
{
protected:
/**
* "Fake" effective address computation class for "%(mnemonic)s".
*/
class EAComp : public %(base_class)s
{
public:
/// Constructor
EAComp(MachInst machInst);
%(BasicExecDeclare)s
};
/**
* "Fake" memory access instruction class for "%(mnemonic)s".
*/
class MemAcc : public %(base_class)s
{
public:
/// Constructor
MemAcc(MachInst machInst);
%(BasicExecDeclare)s
};
public:
/// Constructor.
%(class_name)s(MachInst machInst);
%(BasicExecDeclare)s
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(InitiateAccDeclare)s
%(CompleteAccDeclare)s
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
};
}};
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
def template InitiateAccDeclare {{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * initiateAcc(%(CPU_exec_context)s *, Trace::InstRecord *) const;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
}};
def template CompleteAccDeclare {{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * completeAcc(uint8_t *, %(CPU_exec_context)s *, Trace::InstRecord *) const;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
}};
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
def template LoadStoreConstructor {{
/** TODO: change op_class to AddrGenOp or something (requires
* creating new member of OpClass enum in op_class.hh, updating
* config files, etc.). */
inline %(class_name)s::EAComp::EAComp(MachInst machInst)
: %(base_class)s("%(mnemonic)s (EAComp)", machInst, IntAluOp)
{
%(ea_constructor)s;
}
inline %(class_name)s::MemAcc::MemAcc(MachInst machInst)
: %(base_class)s("%(mnemonic)s (MemAcc)", machInst, %(op_class)s)
{
%(memacc_constructor)s;
}
inline %(class_name)s::%(class_name)s(MachInst machInst)
: %(base_class)s("%(mnemonic)s", machInst, %(op_class)s,
new EAComp(machInst), new MemAcc(machInst))
{
%(constructor)s;
}
}};
def template EACompExecute {{
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault *
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(class_name)s::EAComp::execute(%(CPU_exec_context)s *xc,
Trace::InstRecord *traceData) const
{
Addr EA;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * fault = NoFault;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(fp_enable_check)s;
%(op_decl)s;
%(op_rd)s;
%(code)s;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(op_wb)s;
xc->setEA(EA);
}
return fault;
}
}};
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
def template LoadMemAccExecute {{
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault *
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(class_name)s::MemAcc::execute(%(CPU_exec_context)s *xc,
Trace::InstRecord *traceData) const
{
Addr EA;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * fault = NoFault;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(fp_enable_check)s;
%(op_decl)s;
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(op_rd)s;
EA = xc->getEA();
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
fault = xc->read(EA, (uint%(mem_acc_size)d_t&)Mem, memAccessFlags);
%(code)s;
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(op_wb)s;
}
return fault;
}
}};
def template LoadExecute {{
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * %(class_name)s::execute(%(CPU_exec_context)s *xc,
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
Trace::InstRecord *traceData) const
{
Addr EA;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * fault = NoFault;
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(fp_enable_check)s;
%(op_decl)s;
%(op_rd)s;
%(ea_code)s;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
fault = xc->read(EA, (uint%(mem_acc_size)d_t&)Mem, memAccessFlags);
%(memacc_code)s;
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(op_wb)s;
}
return fault;
}
}};
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
def template LoadInitiateAcc {{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * %(class_name)s::initiateAcc(%(CPU_exec_context)s *xc,
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
Trace::InstRecord *traceData) const
{
Addr EA;
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * fault = NoFault;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(mem_acc_type)s Mem = 0;
%(fp_enable_check)s;
%(op_src_decl)s;
%(op_rd)s;
%(ea_code)s;
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
if (fault == NoFault) {
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
fault = xc->read(EA, (uint%(mem_acc_size)d_t &)Mem, memAccessFlags);
}
return fault;
}
}};
def template LoadCompleteAcc {{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * %(class_name)s::completeAcc(uint8_t *data,
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(CPU_exec_context)s *xc,
Trace::InstRecord *traceData) const
{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * fault = NoFault;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(mem_acc_type)s Mem = 0;
%(fp_enable_check)s;
%(op_dest_decl)s;
memcpy(&Mem, data, sizeof(Mem));
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
if (fault == NoFault) {
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(memacc_code)s;
}
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
if (fault == NoFault) {
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(op_wb)s;
}
return fault;
}
}};
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
def template StoreMemAccExecute {{
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault *
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(class_name)s::MemAcc::execute(%(CPU_exec_context)s *xc,
Trace::InstRecord *traceData) const
{
Addr EA;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * fault = NoFault;
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
uint64_t write_result = 0;
%(fp_enable_check)s;
%(op_decl)s;
%(op_rd)s;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
EA = xc->getEA();
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(code)s;
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
fault = xc->write((uint%(mem_acc_size)d_t&)Mem, EA,
memAccessFlags, &write_result);
if (traceData) { traceData->setData(Mem); }
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(postacc_code)s;
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(op_wb)s;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}
return fault;
}
}};
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
def template StoreExecute {{
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * %(class_name)s::execute(%(CPU_exec_context)s *xc,
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
Trace::InstRecord *traceData) const
{
Addr EA;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * fault = NoFault;
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
uint64_t write_result = 0;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(fp_enable_check)s;
%(op_decl)s;
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(op_rd)s;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(ea_code)s;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(memacc_code)s;
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
fault = xc->write((uint%(mem_acc_size)d_t&)Mem, EA,
memAccessFlags, &write_result);
if (traceData) { traceData->setData(Mem); }
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(postacc_code)s;
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(op_wb)s;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}
return fault;
}
}};
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
def template StoreInitiateAcc {{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * %(class_name)s::initiateAcc(%(CPU_exec_context)s *xc,
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
Trace::InstRecord *traceData) const
{
Addr EA;
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * fault = NoFault;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
uint64_t write_result = 0;
%(mem_acc_type)s Mem = 0;
%(fp_enable_check)s;
%(op_src_decl)s;
%(op_rd)s;
%(ea_code)s;
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
if (fault == NoFault) {
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(memacc_code)s;
}
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
if (fault == NoFault) {
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
fault = xc->write((uint%(mem_acc_size)d_t&)Mem, EA,
memAccessFlags, &write_result);
if (traceData) { traceData->setData(Mem); }
}
return fault;
}
}};
def template StoreCompleteAcc {{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * %(class_name)s::completeAcc(uint8_t *data,
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(CPU_exec_context)s *xc,
Trace::InstRecord *traceData) const
{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * fault = NoFault;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
uint64_t write_result = 0;
%(fp_enable_check)s;
%(op_dest_decl)s;
memcpy(&write_result, data, sizeof(write_result));
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
if (fault == NoFault) {
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(postacc_code)s;
}
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
if (fault == NoFault) {
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(op_wb)s;
}
return fault;
}
}};
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
def template MiscMemAccExecute {{
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * %(class_name)s::MemAcc::execute(%(CPU_exec_context)s *xc,
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
Trace::InstRecord *traceData) const
{
Addr EA;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * fault = NoFault;
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(fp_enable_check)s;
%(op_decl)s;
%(op_rd)s;
EA = xc->getEA();
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(code)s;
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
return NoFault;
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
}
}};
def template MiscExecute {{
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * %(class_name)s::execute(%(CPU_exec_context)s *xc,
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
Trace::InstRecord *traceData) const
{
Addr EA;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
Fault * fault = NoFault;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(fp_enable_check)s;
%(op_decl)s;
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(op_rd)s;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
%(ea_code)s;
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
if (fault == NoFault) {
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
%(memacc_code)s;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}
Changed the fault enum into a class, and fixed everything up to work with it. Next, the faults need to be pulled out of all the other code so that they are only used to communicate between the CPU and the ISA. SConscript: The new faults.cc file in sim allocates the system wide faults. When these faults are generated through a function interface in the ISA, this file may go away. arch/alpha/alpha_memory.cc: Changed Fault to Fault * and took the underscores out of fault names. arch/alpha/alpha_memory.hh: Changed Fault to Fault *. Also, added an include for the alpha faults. arch/alpha/ev5.cc: Changed the fault_addr array into a fault_addr function. Once all of the faults can be expected to have the same type, fault_addr can go away completely and the info it provided will come from the fault itself. Also, Fault was changed to Fault *, and underscores were taken out of fault names. arch/alpha/isa/decoder.isa: Changed Fault to Fault * and took the underscores out fault names. arch/alpha/isa/fp.isa: Changed Fault to Fault *, and took the underscores out of fault names. arch/alpha/isa/main.isa: Changed Fault to Fault *, removed underscores from fault names, and made an include of the alpha faults show up in all the generated files. arch/alpha/isa/mem.isa: Changed Fault to Fault * and removed underscores from fault names. arch/alpha/isa/unimp.isa: arch/alpha/isa/unknown.isa: cpu/exec_context.hh: cpu/ozone/cpu.hh: cpu/simple/cpu.cc: dev/alpha_console.cc: dev/ide_ctrl.cc: dev/isa_fake.cc: dev/pciconfigall.cc: dev/pcidev.cc: dev/pcidev.hh: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: Changed Fault to Fault *, and removed underscores from fault names. arch/alpha/isa_traits.hh: Changed the include of arch/alpha/faults.hh to sim/faults.hh, since the alpha faults weren't needed. cpu/base_dyn_inst.cc: Changed Fault to Fault *, and removed underscores from fault names. This file probably shouldn't use the Unimplemented Opcode fault. cpu/base_dyn_inst.hh: Changed Fault to Fault * and took the underscores out of the fault names. cpu/exec_context.cc: cpu/o3/alpha_dyn_inst.hh: cpu/o3/alpha_dyn_inst_impl.hh: cpu/o3/fetch.hh: dev/alpha_console.hh: dev/baddev.hh: dev/ide_ctrl.hh: dev/isa_fake.hh: dev/ns_gige.hh: dev/pciconfigall.hh: dev/sinic.hh: dev/tsunami_cchip.hh: dev/tsunami_io.hh: dev/tsunami_pchip.hh: dev/uart.hh: dev/uart8250.hh: Changed Fault to Fault *. cpu/o3/alpha_cpu.hh: Changed Fault to Fault *, removed underscores from fault names. cpu/o3/alpha_cpu_impl.hh: Changed Fault to Fault *, removed underscores from fault names, and changed the fault_addr array to the fault_addr function. Once all faults are from the ISA, this function will probably go away. cpu/o3/commit_impl.hh: cpu/o3/fetch_impl.hh: dev/baddev.cc: Changed Fault to Fault *, and removed underscores from the fault names. cpu/o3/regfile.hh: Added an include for the alpha specific faults which will hopefully go away once the ipr stuff is moved, changed Fault to Fault *, and removed the underscores from fault names. cpu/simple/cpu.hh: Changed Fault to Fault * dev/ns_gige.cc: Changed Fault to Fault *, and removdd underscores from fault names. dev/sinic.cc: Changed Fault to Fault *, and removed the underscores from fault names. dev/uart8250.cc: Chanted Fault to Fault *, and removed underscores from fault names. kern/kernel_stats.cc: Removed underscores from fault names, and from NumFaults. kern/kernel_stats.hh: Changed the predeclaration of Fault from an enum to a class, and changd the "fault" function to work with the classes instead of the enum. Once there are no system wide faults anymore, this code will simplify back to something like it was originally. sim/faults.cc: This allocates the system wide faults. sim/faults.hh: This declares the system wide faults. sim/syscall_emul.cc: sim/syscall_emul.hh: Removed the underscores from fault names. --HG-- rename : arch/alpha/faults.cc => sim/faults.cc rename : arch/alpha/faults.hh => sim/faults.hh extra : convert_revision : 253d39258237333ae8ec4d8047367cb3ea68569d
2006-02-16 07:22:51 +01:00
return NoFault;
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}
}};
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
def template MiscInitiateAcc {{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * %(class_name)s::initiateAcc(%(CPU_exec_context)s *xc,
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
Trace::InstRecord *traceData) const
{
Addr EA;
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * fault = NoFault;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(fp_enable_check)s;
%(op_decl)s;
%(op_rd)s;
%(ea_code)s;
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
if (fault == NoFault) {
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(memacc_code)s;
}
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
return NoFault;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
}
}};
def template MiscCompleteAcc {{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
Fault * %(class_name)s::completeAcc(uint8_t *data,
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
%(CPU_exec_context)s *xc,
Trace::InstRecord *traceData) const
{
Some changes which weren't needed before doing a bk pull were needed afterwards, for some reason. arch/alpha/ev5.cc: Took out the unnecessary check for a null Fault pointer. arch/alpha/isa/mem.isa: Changed Fault to Fault *, and removed underscores from fault names. --HG-- extra : convert_revision : 367a58a375f911185ddcc5fc826034af96427461
2006-02-16 08:08:13 +01:00
return NoFault;
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
}
}};
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
// load instructions use Ra as dest, so check for
// Ra == 31 to detect nops
def template LoadNopCheckDecode {{
{
AlphaStaticInst *i = new %(class_name)s(machInst);
if (RA == 31) {
i = makeNop(i);
}
return i;
}
}};
// for some load instructions, Ra == 31 indicates a prefetch (not a nop)
def template LoadPrefetchCheckDecode {{
{
if (RA != 31) {
return new %(class_name)s(machInst);
}
else {
return new %(class_name)sPrefetch(machInst);
}
}
}};
let {{
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
def LoadStoreBase(name, Name, ea_code, memacc_code, mem_flags, inst_flags,
postacc_code = '', base_class = 'MemoryDisp32',
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
decode_template = BasicDecode, exec_template_base = ''):
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
# Make sure flags are in lists (convert to lists if not).
mem_flags = makeList(mem_flags)
inst_flags = makeList(inst_flags)
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
# add hook to get effective addresses into execution trace output.
ea_code += '\nif (traceData) { traceData->setAddr(EA); }\n'
# generate code block objects
ea_cblk = CodeBlock(ea_code)
memacc_cblk = CodeBlock(memacc_code)
postacc_cblk = CodeBlock(postacc_code)
# Some CPU models execute the memory operation as an atomic unit,
# while others want to separate them into an effective address
# computation and a memory access operation. As a result, we need
# to generate three StaticInst objects. Note that the latter two
# are nested inside the larger "atomic" one.
# generate InstObjParams for EAComp object
ea_iop = InstObjParams(name, Name, base_class, ea_cblk, inst_flags)
# generate InstObjParams for MemAcc object
memacc_iop = InstObjParams(name, Name, base_class, memacc_cblk, inst_flags)
# in the split execution model, the MemAcc portion is responsible
# for the post-access code.
memacc_iop.postacc_code = postacc_cblk.code
# generate InstObjParams for unified execution
cblk = CodeBlock(ea_code + memacc_code + postacc_code)
iop = InstObjParams(name, Name, base_class, cblk, inst_flags)
iop.ea_constructor = ea_cblk.constructor
iop.ea_code = ea_cblk.code
iop.memacc_constructor = memacc_cblk.constructor
iop.memacc_code = memacc_cblk.code
iop.postacc_code = postacc_cblk.code
if mem_flags:
s = '\n\tmemAccessFlags = ' + string.join(mem_flags, '|') + ';'
iop.constructor += s
memacc_iop.constructor += s
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
# select templates
memAccExecTemplate = eval(exec_template_base + 'MemAccExecute')
fullExecTemplate = eval(exec_template_base + 'Execute')
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
initiateAccTemplate = eval(exec_template_base + 'InitiateAcc')
completeAccTemplate = eval(exec_template_base + 'CompleteAcc')
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
# (header_output, decoder_output, decode_block, exec_output)
return (LoadStoreDeclare.subst(iop), LoadStoreConstructor.subst(iop),
decode_template.subst(iop),
EACompExecute.subst(ea_iop)
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
+ memAccExecTemplate.subst(memacc_iop)
Gives separate methods for initiating and completing a memory access, which will be helpful for the merged memory model. arch/alpha/isa/mem.isa: Include methods that allow a memory operation to be split between the part that initiates the access, and the part that completes the access. In these functions the Mem variable is explicitly declared; in the default execute functions, the Mem variable is still handled through %(op_decl)s. arch/isa_parser.py: Include recording the type of the memory access variable so that it can be used if it needs to be explicitly declared in a template. Have memory operands consider themselves neither a source nor a destination to avoid including themselves on the op_src_decl list or the op_dest_decl list. Record op_src_decl and op_dest_decl lists to allow for declaring only source or destination operands. This is needed for the split memory access methods. --HG-- extra : convert_revision : f674f7a2f747ae40ba8c3a0933b0337c87ee0b6c
2006-02-15 19:05:21 +01:00
+ fullExecTemplate.subst(iop)
+ initiateAccTemplate.subst(iop)
+ completeAccTemplate.subst(iop))
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}};
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
def format LoadOrNop(memacc_code, ea_code = {{ EA = Rb + disp; }},
mem_flags = [], inst_flags = []) {{
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
(header_output, decoder_output, decode_block, exec_output) = \
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
LoadStoreBase(name, Name, ea_code, memacc_code, mem_flags, inst_flags,
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
decode_template = LoadNopCheckDecode,
exec_template_base = 'Load')
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}};
// Note that the flags passed in apply only to the prefetch version
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
def format LoadOrPrefetch(memacc_code, ea_code = {{ EA = Rb + disp; }},
mem_flags = [], pf_flags = [], inst_flags = []) {{
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
# declare the load instruction object and generate the decode block
(header_output, decoder_output, decode_block, exec_output) = \
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
LoadStoreBase(name, Name, ea_code, memacc_code, mem_flags, inst_flags,
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
decode_template = LoadPrefetchCheckDecode,
exec_template_base = 'Load')
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
# Declare the prefetch instruction object.
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
# Make sure flag args are lists so we can mess with them.
mem_flags = makeList(mem_flags)
pf_flags = makeList(pf_flags)
inst_flags = makeList(inst_flags)
pf_mem_flags = mem_flags + pf_flags + ['NO_FAULT']
pf_inst_flags = inst_flags + ['IsMemRef', 'IsLoad',
'IsDataPrefetch', 'MemReadOp']
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
(pf_header_output, pf_decoder_output, _, pf_exec_output) = \
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
LoadStoreBase(name, Name + 'Prefetch', ea_code,
'xc->prefetch(EA, memAccessFlags);',
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
pf_mem_flags, pf_inst_flags, exec_template_base = 'Misc')
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
header_output += pf_header_output
decoder_output += pf_decoder_output
exec_output += pf_exec_output
}};
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
def format Store(memacc_code, ea_code = {{ EA = Rb + disp; }},
mem_flags = [], inst_flags = []) {{
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
(header_output, decoder_output, decode_block, exec_output) = \
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
LoadStoreBase(name, Name, ea_code, memacc_code, mem_flags, inst_flags,
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
exec_template_base = 'Store')
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}};
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
def format StoreCond(memacc_code, postacc_code,
ea_code = {{ EA = Rb + disp; }},
mem_flags = [], inst_flags = []) {{
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
(header_output, decoder_output, decode_block, exec_output) = \
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
LoadStoreBase(name, Name, ea_code, memacc_code, mem_flags, inst_flags,
postacc_code, exec_template_base = 'Store')
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}};
// Use 'MemoryNoDisp' as base: for wh64, fetch, ecb
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
def format MiscPrefetch(ea_code, memacc_code,
mem_flags = [], inst_flags = []) {{
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
(header_output, decoder_output, decode_block, exec_output) = \
Add keyword parameters and list-valued arguments to instruction format functions in ISA description language. Take advantage of these features to clean up memory instruction definitions in Alpha. arch/alpha/isa/decoder.isa: arch/alpha/isa/mem.isa: arch/alpha/isa/pal.isa: Take advantage of new keyword parameters to disambiguate instruction vs. memory-request flags, and to provide a default EA calculation for memory ops (since 99% of them are the same). arch/isa_parser.py: Add two new features to instruction format functions: - Keyword parameters, a la Python. - List-valued arguments. Also export makeList() function to Python code blocks, as this is handy for dealing with flags. --HG-- extra : convert_revision : 99bbbaa2e765230aa96b6a06ed193793325f9fb0
2006-02-11 21:11:00 +01:00
LoadStoreBase(name, Name, ea_code, memacc_code, mem_flags, inst_flags,
Change how memory operands are handled in ISA descriptions. Should enable implementation of split-phase timing loads with new memory model. May create slight timing differences under FullCPU, as I believe we were not handling software prefetches correctly before when the split MemAcc/Exec model was used. I haven't looked into this in any detail though. arch/alpha/isa/decoder.isa: HwLoadStore format split into separate HwLoad and HwStore formats. Copy instructions now fall under MiscPrefetch format. Mem_write_result is now just write_result in store conditionals. arch/alpha/isa/mem.isa: Split MemAccExecute and LoadStoreExecute templates into separate templates for loads and stores; now that memory operands are handled differently from registers, it's impossible to have a single template serve both. Also unified the handling of "regular" prefetches (loads to r31) and "misc" prefetches (e.g., wh64) under the new scheme. It looks like SW prefetches were not handled correctly in FullCPU up til now, since we generated an execute() method for the outer instruction but didn't generate a proper method for MemAcc::execute() (instead getting a default no-op method for that). arch/alpha/isa/pal.isa: Split HwLoadStore into separate HwLoad and HwStore formats to select proper template (see change to mem.isa in this changeset). arch/isa_parser.py: Stop trying to treat memory operands like register operands, since we never used them in a uniform way anyway, and it made it impossible to do split-phase loads as needed for the new CPU model. Now there's no more 'op_mem_rd', 'op_nonmem_rd', etc.: 'op_rd' just does register operands, and the template code is responsible for formulating the call to the memory system. Right now the only thing exported by InstObjParams is a new attribute 'mem_acc_size' which gives the memory access size in bits, though more attributes can be added if needed. Also moved code in findOperands() method to OperandDescriptorList.__init__(), which is where it belongs. --HG-- extra : convert_revision : 6d53d07e0c5e828455834ded4395fa40f9146a34
2006-02-10 15:12:55 +01:00
base_class = 'MemoryNoDisp', exec_template_base = 'Misc')
Split Alpha ISA description into multiple files (thanks to Gabe's include feature!). arch/alpha/isa/main.isa: Split out into multiple .isa files. --HG-- extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
2006-02-10 05:02:38 +01:00
}};
Reference in a new issue Copy permalink