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

346 lines
9.1 KiB
C++
Raw Normal View History

Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
/*
* Copyright (c) 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.
*/
#include <string>
#include "arch/alpha/isa_traits.hh"
#include "arch/alpha/stacktrace.hh"
#include "arch/alpha/vtophys.hh"
#include "base/bitfield.hh"
#include "base/trace.hh"
#include "cpu/base.hh"
#include "cpu/exec_context.hh"
using namespace std;
ProcessInfo::ProcessInfo(ExecContext *_xc)
: xc(_xc)
{
Addr addr = 0;
if (!xc->system->kernelSymtab->findAddress("thread_info_size", addr))
panic("thread info not compiled into kernel\n");
thread_info_size = *(int32_t *)vtomem(xc, addr, sizeof(int32_t));
if (!xc->system->kernelSymtab->findAddress("task_struct_size", addr))
panic("thread info not compiled into kernel\n");
task_struct_size = *(int32_t *)vtomem(xc, addr, sizeof(int32_t));
if (!xc->system->kernelSymtab->findAddress("thread_info_task", addr))
panic("thread info not compiled into kernel\n");
task_off = *(int32_t *)vtomem(xc, addr, sizeof(int32_t));
if (!xc->system->kernelSymtab->findAddress("task_struct_pid", addr))
panic("thread info not compiled into kernel\n");
pid_off = *(int32_t *)vtomem(xc, addr, sizeof(int32_t));
if (!xc->system->kernelSymtab->findAddress("task_struct_comm", addr))
panic("thread info not compiled into kernel\n");
name_off = *(int32_t *)vtomem(xc, addr, sizeof(int32_t));
}
Addr
ProcessInfo::task(Addr ksp) const
{
Addr base = ksp & ~0x3fff;
if (base == ULL(0xfffffc0000000000))
return 0;
Addr task;
CopyOut(xc, &task, base + task_off, sizeof(task));
return task;
}
int
ProcessInfo::pid(Addr ksp) const
{
Addr task = this->task(ksp);
if (!task)
return -1;
uint16_t pid;
CopyOut(xc, &pid, task + pid_off, sizeof(pid));
return pid;
}
string
ProcessInfo::name(Addr ksp) const
{
Addr task = this->task(ksp);
if (!task)
return "console";
char comm[256];
CopyString(xc, comm, task + name_off, sizeof(comm));
if (!comm[0])
return "startup";
return comm;
}
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
StackTrace::StackTrace()
: xc(0), stack(64)
{
}
StackTrace::StackTrace(ExecContext *_xc, StaticInstPtr<TheISA> inst)
: xc(0), stack(64)
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
{
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
trace(_xc, inst);
}
StackTrace::~StackTrace()
{
}
void
StackTrace::trace(ExecContext *_xc, bool is_call)
{
xc = _xc;
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
bool usermode = (xc->regs.ipr[AlphaISA::IPR_DTB_CM] & 0x18) != 0;
Addr pc = xc->regs.npc;
bool kernel = xc->system->kernelStart <= pc && pc <= xc->system->kernelEnd;
if (usermode) {
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
stack.push_back(user);
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
return;
}
if (!kernel) {
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
stack.push_back(console);
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
return;
}
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
SymbolTable *symtab = xc->system->kernelSymtab;
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
Addr ksp = xc->regs.intRegFile[TheISA::StackPointerReg];
Addr bottom = ksp & ~0x3fff;
Addr addr;
if (is_call) {
if (!symtab->findNearestAddr(pc, addr))
panic("could not find address %#x", pc);
stack.push_back(addr);
pc = xc->regs.pc;
}
Addr ra;
int size;
while (ksp > bottom) {
if (!symtab->findNearestAddr(pc, addr))
panic("could not find symbol for pc=%#x", pc);
assert(pc >= addr && "symbol botch: callpc < func");
stack.push_back(addr);
if (isEntry(addr))
return;
if (decodePrologue(ksp, pc, addr, size, ra)) {
if (!ra)
return;
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
if (size <= 0) {
stack.push_back(unknown);
return;
}
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
pc = ra;
ksp += size;
} else {
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
stack.push_back(unknown);
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
return;
}
bool kernel = xc->system->kernelStart <= pc &&
pc <= xc->system->kernelEnd;
if (!kernel)
return;
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
if (stack.size() >= 1000)
panic("unwinding too far");
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
}
panic("unwinding too far");
}
bool
StackTrace::isEntry(Addr addr)
{
if (addr == xc->regs.ipr[AlphaISA::IPR_PALtemp12])
return true;
if (addr == xc->regs.ipr[AlphaISA::IPR_PALtemp7])
return true;
if (addr == xc->regs.ipr[AlphaISA::IPR_PALtemp11])
return true;
if (addr == xc->regs.ipr[AlphaISA::IPR_PALtemp21])
return true;
if (addr == xc->regs.ipr[AlphaISA::IPR_PALtemp9])
return true;
if (addr == xc->regs.ipr[AlphaISA::IPR_PALtemp2])
return true;
return false;
}
bool
StackTrace::decodeStack(MachInst inst, int &disp)
{
// lda $sp, -disp($sp)
//
// Opcode<31:26> == 0x08
// RA<25:21> == 30
// RB<20:16> == 30
// Disp<15:0>
const MachInst mem_mask = 0xffff0000;
const MachInst lda_pattern = 0x23de0000;
const MachInst lda_disp_mask = 0x0000ffff;
// subq $sp, disp, $sp
// addq $sp, disp, $sp
//
// Opcode<31:26> == 0x10
// RA<25:21> == 30
// Lit<20:13>
// One<12> = 1
// Func<11:5> == 0x20 (addq)
// Func<11:5> == 0x29 (subq)
// RC<4:0> == 30
const MachInst intop_mask = 0xffe01fff;
const MachInst addq_pattern = 0x43c0141e;
const MachInst subq_pattern = 0x43c0153e;
const MachInst intop_disp_mask = 0x001fe000;
const int intop_disp_shift = 13;
if ((inst & mem_mask) == lda_pattern)
disp = -sext<16>(inst & lda_disp_mask);
else if ((inst & intop_mask) == addq_pattern)
disp = -int((inst & intop_disp_mask) >> intop_disp_shift);
else if ((inst & intop_mask) == subq_pattern)
disp = int((inst & intop_disp_mask) >> intop_disp_shift);
else
return false;
return true;
}
bool
StackTrace::decodeSave(MachInst inst, int &reg, int &disp)
{
// lda $stq, disp($sp)
//
// Opcode<31:26> == 0x08
// RA<25:21> == ?
// RB<20:16> == 30
// Disp<15:0>
const MachInst stq_mask = 0xfc1f0000;
const MachInst stq_pattern = 0xb41e0000;
const MachInst stq_disp_mask = 0x0000ffff;
const MachInst reg_mask = 0x03e00000;
const int reg_shift = 21;
if ((inst & stq_mask) == stq_pattern) {
reg = (inst & reg_mask) >> reg_shift;
disp = sext<16>(inst & stq_disp_mask);
} else {
return false;
}
return true;
}
/*
* Decode the function prologue for the function we're in, and note
* which registers are stored where, and how large the stack frame is.
*/
bool
StackTrace::decodePrologue(Addr sp, Addr callpc, Addr func,
int &size, Addr &ra)
{
size = 0;
ra = 0;
for (Addr pc = func; pc < callpc; pc += sizeof(MachInst)) {
MachInst inst;
CopyOut(xc, (uint8_t *)&inst, pc, sizeof(MachInst));
int reg, disp;
if (decodeStack(inst, disp)) {
if (size) {
// panic("decoding frame size again");
return true;
}
size += disp;
} else if (decodeSave(inst, reg, disp)) {
if (!ra && reg == ReturnAddressReg) {
CopyOut(xc, (uint8_t *)&ra, sp + disp, sizeof(Addr));
if (!ra) {
// panic("no return address value pc=%#x\n", pc);
return false;
}
}
}
}
return true;
}
#if TRACING_ON
void
StackTrace::dump()
{
StringWrap name(xc->cpu->name());
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
SymbolTable *symtab = xc->system->kernelSymtab;
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
DPRINTFN("------ Stack ------\n");
string symbol;
for (int i = 0, size = stack.size(); i < size; ++i) {
Addr addr = stack[size - i - 1];
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
if (addr == user)
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
symbol = "user";
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
else if (addr == console)
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
symbol = "console";
Cleanup the StackTrace interfaces and profile interfaces so they are more efficient and reduce the number of new/delete calls arch/alpha/stacktrace.cc: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers - switch to using kernelSymtab instead of allSymtab which will be going away - if the stack adjustment doesn't make any sense, exit and push unknown so we don't get into an infinite loop or record garbage. - check to see if we've made too many iterations through the stack and panic to avoid an infinite loop arch/alpha/stacktrace.hh: - Change the StackTrace code so that the class can more easily be cleaned out and reused to avoid extra allocations. - Allow trace() to accept a static instruction pointer so it can determine if the instruction is worth tracing. This is moved from the CPU. - provide constants for special meaning PCs (user, console, unknown), instead of magic numbers cpu/base.cc: only clear the profile if we have one include profile.hh here since base.hh doesn't do it anymore cpu/base.hh: no need to include cpu/profile.hh here cpu/profile.cc: use ProfileNode pointers instead of objects in the ChildList Consume a vector of addresses since that's really all we care about. cpu/profile.hh: Keep pointers to ProfileNodes to reduce the size of these structures keep a StackTrace around so that we may reuse it. provide consume functions that use the new StackTrace trace interface one consume function is inline and tries to fastpath the no trace condition, it calls the outlined consume function if a trace is generated. cpu/simple/cpu.cc: include cpu/profile.hh here since base.hh no longer does use the new FunctionProfile::consume interface (which contains the tracing functions) --HG-- extra : convert_revision : 5a1d9265289a75f67a497b322926be1f8c2d8eb3
2005-11-20 23:44:58 +01:00
else if (addr == unknown)
Add new function profiling stuff, wrap the pc_sample stuff into it. SConscript: Get rid of the pc_sample stuff and move to the new profiling stuff base/traceflags.py: DPRINTF Stack stuff cpu/base.cc: cpu/base.hh: cpu/exec_context.cc: cpu/exec_context.hh: cpu/simple/cpu.cc: Add profiling stuff kern/kernel_stats.hh: Use a smart pointer sim/system.cc: sim/system.hh: Create a new symbol table that has all of the symbols for a particular system util/stats/categories.py: change around the categories, add categories for function profiling stuff util/stats/profile.py: No profile parsing and display code to deal with function profiling stuff, graph, dot, and text outputs. --HG-- extra : convert_revision : b3de0cdc8bd468e42647966e2640ae009bda9eb8
2005-10-19 01:07:42 +02:00
symbol = "unknown";
else
symtab->findSymbol(addr, symbol);
DPRINTFN("%#x: %s\n", addr, symbol);
}
}
#endif
Reference in a new issue Copy permalink