minor change to syntax of 'def operands' in ISA descriptions.
arch/alpha/isa/main.isa:
arch/mips/isa/operands.isa:
arch/sparc/isa/operands.isa:
Change 'def operands' statement to work with new
isa_parser changes.
arch/isa_parser.py:
Merge OperandTraits and OperandDescriptor objects into a
unified hierarchy of Operand objects.
Required a change in the syntax of the 'def operands'
statement in the ISA description.
--HG--
extra : convert_revision : cb43f1607311497ead88ba13953d410ab5bc6a37
arch/isa_parser.py:
Minor cleanup of operand type and traits code:
- build operand size map right away when types are defined
instead of waiting to do it lazily
- check that operand types have been defined before operands
- don't use 'type' as a variable name
- use isinstance() instead of checking for types directly
--HG--
extra : convert_revision : 099c1ee8d490f9c38316749bf87209388c55c971
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
wrong
arch/alpha/alpha_linux_process.cc:
fix #if. I wonder why my compiler had no issues
--HG--
extra : convert_revision : 880a0442b28811db5ec548ce940060d4b26ec634
arch/alpha/alpha_linux_process.cc:
Add endian conversions to fstat
sim/byteswap.hh:
for some reason I don't understand g++ really wanted a long version defined
even though int32_t should be the same.
--HG--
extra : convert_revision : 5bfe9d3f0b31824fa5a7ae3f51fd0be5ed4d555d
arch/alpha/alpha_linux_process.cc:
arch/alpha/alpha_tru64_process.cc:
fixup for bsd hosts. Some headers are included by default which means that
more variables need TGT_ prefixes and there isn't a stat call (everything
is a stat64 call) so we have to work around that a bit
base/intmath.hh:
base/socket.cc:
this is no longer needed with mac os 10.4
cpu/inst_seq.hh:
just use a uint64_t instead of long long
cpu/o3/inst_queue_impl.hh:
I much cleaner way to get max int
sim/syscall_emul.hh:
fix stat64 problems on *BSD
--HG--
extra : convert_revision : 9eef5f896e083ae1774e818a9765dd83e0305942
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
(thanks to Gabe's include feature!).
arch/alpha/isa/main.isa:
Split out into multiple .isa files.
--HG--
extra : convert_revision : 30d8edf74ea194d4a208febf1e66edc72a7dbd5d
arch/isa_parser.py:
Clean up ##include code a bit.
arch/sparc/isa/formats.isa:
arch/sparc/isa/main.isa:
Fix include paths.
--HG--
extra : convert_revision : 0689963c2948e5f1088ecbf2cf6018d29bdaceff
arch/isa_parser.py:
Get rid of "munged name" for operands in C++ code.
That is, "Ra.uq" will now be known in the C++ as "Ra"
rather than "Ra_uq". It wasn't legal to use different
type extensions for the same operand at the same time
anyway, and now it will be easier to refer to explicit
operands in template code if necessary.
--HG--
extra : convert_revision : 9ff41e0201aeefe761743084ecdb34f4b9c84fdb
arch/mips/isa/decoder.isa:
Code for di,ei,seb,seh,clz,and clo ....
Every instruction has a format now (of course these are initial formats are still subject to change!)
arch/mips/isa/formats/branch.isa:
Format Branch in MIPS similar to Alpha Format
--HG--
extra : convert_revision : 2118a1d9668610b1e9f1dea66d878b7b36c1ac7e
arch/alpha/alpha_linux_process.cc:
arch/alpha/alpha_tru64_process.cc:
Replaced the namespace declaration with including arch/alpha/isa_traits.hh
--HG--
extra : convert_revision : 07cb73a9f30f0e165809668f9baff6a3e3f94580
SConscript:
Moved some files out of targetarch. The either no longer need to be there, never needed to be there, or should be referred to directly in arch/alpha due to there strictly alpha content.
arch/alpha/isa_traits.hh:
Added alpha's endianness to it's isa_traits.hh
arch/mips/isa_traits.hh:
Added MIPS endianness to it's isa_traits.hh
arch/sparc/isa_traits.hh:
Added SPARCs endianess to it's isa_traits.hh
build/SConstruct:
Added MIPS as a valid architecture
cpu/exec_context.hh:
Included arch/isa_traits.hh to bring in the endianness of the system.
cpu/o3/alpha_cpu.hh:
Included arch/isa_traits.hh to bring in the systems endianness, and removed the hardcoding of little endianness
cpu/o3/fetch_impl.hh:
kern/freebsd/freebsd_system.cc:
Included arch/isa_traits.hh to bring in the systems endianness, and removed the hardcoding to little endianness.
sim/system.cc:
Included arch/isa_traits.hh to bring in the systems endianness, and removed the hardcoding to little endian.
--HG--
extra : convert_revision : b1ab34b7569db531cd1c74f273b24222e63f9007
The extra class is needed because of the necessisty of an immediate member variable.
Also, added some 'very modest' python code to choose between the IntOp and
the IntImmOp based on the instruction name ...
--HG--
extra : convert_revision : f109c12418202a99b40e270360134e8335739836
arch/mips/isa/formats/int.format:
Looks like Integer Ops with Immediates may not need their own separate class because all those instructions are distinct from
their reg-reg counterparts
--HG--
rename : arch/mips/isa/bitfields.def => arch/mips/isa/bitfields.isa
rename : arch/mips/isa/decoder.def => arch/mips/isa/decoder.isa
rename : arch/mips/isa/formats.def => arch/mips/isa/formats.isa
rename : arch/mips/isa/includes.h => arch/mips/isa/includes.isa
rename : arch/mips/isa/operands.def => arch/mips/isa/operands.isa
extra : convert_revision : 8e354b4232b28c0264d98d333d55ef8b5a6589cc
Output,Format, & Template code needs to be adjusted to correctly take these "decoder.h" inputs ...
--HG--
extra : convert_revision : 3dcde1f2f587e2766fd61231a93d34d1d7727356
arch/alpha/alpha_linux_process.cc:
arch/alpha/alpha_tru64_process.cc:
Added the endianness namespace. This may change.
cpu/exec_context.hh:
Changed the include path for byteswap, and forced LittleEndianness for lack of a better solution.
cpu/o3/alpha_cpu.hh:
Forced LittleEndianness, for lack of a better solution.
cpu/o3/alpha_cpu_impl.hh:
Cleared away some commented out code.
cpu/o3/fetch_impl.hh:
Changed the include patch for byteswap, and forced LittleEndianness for lack of a better solution.
cpu/simple/cpu.cc:
Added an include for byteswap.hh, and fixed the SimpleCPU to LittleEndian. This cpu only does alpha, so that's fine.
dev/disk_image.cc:
Changed the include path of byteswap.hh
kern/freebsd/freebsd_system.cc:
kern/linux/linux_system.cc:
Added an include for byteswap.hh, and forced LittleEndianness for lack of a better solution.
sim/system.cc:
Forced LittleEndianness for lack of a better solution.
--HG--
extra : convert_revision : b95d3e1265a825e04bd77622a3ac09fbac6bd206
All of the code literals are empty as of now. As much as possible instructions were organized into relevant "formats"
and also references to the tables I used from the MIPS manual were noted via appropriate comments.
--HG--
extra : convert_revision : 9b44fb40e900061a4cdb290b6a5aaceb9750ae13
arch/mips/isa_desc/bitfields.h:
Change from table names to actual bitfield name ...
--HG--
extra : convert_revision : ead69065eb9c3e9c4ea4f67587a6fb07091898ed
- this will decode the instructions but not doing anything to create the C++ object yet
(the 1st of many steps!)
arch/mips/isa_desc/bitfields.h:
initial bitfield constants ... copied some from original alpha bitfields
arch/mips/isa_desc/decoder.h:
decoder function skeleton pt.1
- this will decode the instructions but not doing anything to create the C++ object yet
(the 1st of many steps!)
--HG--
extra : convert_revision : 2b9a0f8160c78b17f9d3d5eaf5af5a4d2f074761
SConscript:
There is a new SConscript in the arch/alpha directory which has the alpha specific files. To add files for an arch, a similar file should be created.
arch/isa_parser.py:
The isa parser now supports include directives. These are done with ##include
build/SConstruct:
The target directory is passed on so that the architecture specific SConscript can have it. Also, sparc was added as a valid architecture type.
arch/alpha/SConscript:
This SConscript adds the alpha specific source
arch/sparc/isa_desc/operands.h:
This sets up the operand types that the sparc isa uses
arch/sparc/isa_traits.cc:
Implementation of sparc specific things, like a register file with windows
build/build_options/default/SPARC_SE:
The default options for a sparc syscall emulation build.
--HG--
extra : convert_revision : 1afedae61dc8cae0d59d3bf1d41420d929be2efd
arch/sparc/isa_desc/bitfields.h:
This file defines the bit fields used by the isa description system
arch/sparc/isa_desc/decoder.h:
This file describes the decoder for the isa description system
arch/sparc/isa_desc/formats.h:
This file declares the instruction formats
arch/sparc/isa_desc/formats/basic.format:
This file implements the "basic" instruction format
arch/sparc/isa_desc/formats/branch.format:
This file implements the "branch" instruction format
arch/sparc/isa_desc/formats/integerop.format:
This file implements the "integerop" instruction format
arch/sparc/isa_desc/formats/mem.format:
This file implements the "mem" instruction format
arch/sparc/isa_desc/formats/noop.format:
This file implements the "noop" instruction format
arch/sparc/isa_desc/formats/trap.format:
This file implements the "trap" instruction format
arch/sparc/isa_desc/includes.h:
This file is all of the inclues that are used by the isa description system
--HG--
extra : convert_revision : 12a2ffe949317b8b57d83263a4261131b9432c2a
arch/alpha/alpha_linux_process.cc:
sim/syscall_emul.cc:
sim/syscall_emul.hh:
Apply patch for syscall emulation provided by Antti Miettinen (apm@brigitte.dna.fi).
--HG--
extra : convert_revision : 37fbc78a927110b7798343afd2c5f37a269e42b4
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
running SPEC FP codes).
arch/alpha/isa_desc:
Don't warn about non-standard trapping modes more than
once per static instruction. (Had the flag to suppress
these but forgot to check it!)
build/SConstruct:
Add USE_SSE2 option to enable compiling w/SSE2 (important
for getting IEEE-compliant FP on x86).
--HG--
extra : convert_revision : eac69efb28cce7b48035480d8b7cb004782969f4
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
Seems to avoid the significant problems on platforms w/o fenv.h.
arch/alpha/isa_desc:
Explicitly handle rounding on FP-to-integer conversions.
Seems to avoid the significant problems on platforms w/o fenv.h.
Get rid of FP "Fast" vs "General" distinction... more headache than
it's worth.
arch/isa_parser.py:
Fix bug with "%s" in C++ templates (must escape properly to
pass through Python string interpolation).
--HG--
extra : convert_revision : de964d764e67e0934ac0ef535f53c974640731fb
