Ali Saidi [Sun, 5 Mar 2006 02:17:24 +0000 (21:17 -0500)]
Merge zizzer:/bk/m5
into zeep.eecs.umich.edu:/z/saidi/work/m5.head
--HG--
extra : convert_revision :
faf71f4ed7d97734835fc7cd552e21b356b8d864
Steve Reinhardt [Sun, 5 Mar 2006 02:06:40 +0000 (21:06 -0500)]
Emulated syscalls should return -errno on failure
(not plain errno). Thanks to Jos Delbar.
--HG--
extra : convert_revision :
bd15ba1c99f25b8560938566d3922734172ab905
Ali Saidi [Sun, 5 Mar 2006 01:45:13 +0000 (20:45 -0500)]
Merge zizzer:/bk/m5
into zeep.eecs.umich.edu:/z/saidi/work/m5.head
--HG--
extra : convert_revision :
405920a23690089a02a5b75cfe947885d8749dc4
Ali Saidi [Sun, 5 Mar 2006 01:45:01 +0000 (20:45 -0500)]
move alpha specific code into arch/alpha
System -> AlphaSystem
SConscript:
Move some more files around
arch/alpha/freebsd/system.cc:
arch/alpha/freebsd/system.hh:
System -> AlphaSystem
arch/alpha/linux/aligned.hh:
arch/alpha/linux/hwrpb.hh:
arch/alpha/linux/system.cc:
arch/alpha/linux/thread_info.hh:
arch/alpha/linux/threadinfo.hh:
move alpha specific code into arch/alpha
python/m5/objects/System.py:
Split off arch specific parts of system into AlphaSystem
--HG--
rename : kern/freebsd/freebsd_system.cc => arch/alpha/freebsd/system.cc
rename : kern/freebsd/freebsd_system.hh => arch/alpha/freebsd/system.hh
rename : kern/linux/aligned.hh => arch/alpha/linux/aligned.hh
rename : kern/linux/hwrpb.hh => arch/alpha/linux/hwrpb.hh
rename : kern/linux/linux_system.cc => arch/alpha/linux/system.cc
rename : kern/linux/linux_system.hh => arch/alpha/linux/system.hh
rename : kern/linux/thread_info.hh => arch/alpha/linux/thread_info.hh
rename : kern/linux/linux_threadinfo.hh => arch/alpha/linux/threadinfo.hh
rename : kern/tru64/tru64_system.cc => arch/alpha/tru64/system.cc
rename : kern/tru64/tru64_system.hh => arch/alpha/tru64/system.hh
extra : convert_revision :
1c496efb823326a4b629574bd9b5f30e99a0aa40
Ali Saidi [Sat, 4 Mar 2006 20:15:47 +0000 (15:15 -0500)]
properly align the uid and size fields in the pre_F54_stat structure.
Patch from Jos Delbar, modified to use variable attributes.
--HG--
extra : convert_revision :
12c9ebbb109a789dd4230b7606b1f8ded00c21fc
Ali Saidi [Sat, 4 Mar 2006 17:35:13 +0000 (12:35 -0500)]
we REALLY don't want to use system headers for libelf...
use "" not <>
targetarch->arch
arch/alpha/system.cc:
targetarch -> arch
base/loader/elf_object.cc:
we REALLY don't want to use system headers for libelf...
use "" not <>
--HG--
extra : convert_revision :
d0e1672dc655dca3081e36b2158e7b19ef54af40
Gabe Black [Sat, 4 Mar 2006 08:19:58 +0000 (03:19 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
--HG--
extra : convert_revision :
cb6de8d050ac5abb8f61eb5780deeb06e93b1a0c
Gabe Black [Sat, 4 Mar 2006 08:16:16 +0000 (03:16 -0500)]
General small SPARC fixups
arch/sparc/SConscript:
Fixed up the SPARC SConscript to use the name of the SPARC source files.
arch/sparc/linux_process.cc:
Changed the namespace name from SPARCISA to SparcISA.
--HG--
extra : convert_revision :
c340fbf78200bf7d962e7578a3f2998860224c53
Gabe Black [Sat, 4 Mar 2006 08:10:34 +0000 (03:10 -0500)]
Removed the commented out declaration of the Addr type.
--HG--
extra : convert_revision :
70612a3efc877febbff0c8fb9065dea7e1d12cd8
Gabe Black [Sat, 4 Mar 2006 08:09:23 +0000 (03:09 -0500)]
Filled out the object file loader so it can load object files for several OSs and architectures.
SConscript:
Added ./libelf as an include search directory. There might be a better spot for this than where I put it.
arch/SConscript:
Combined the linux_process.h and tru64_process.h into process.h. This allows each ISA to support processes from arbitrary OSs.
arch/alpha/SConscript:
Added process.cc as a source file. It provides an implementation of createProcess, which takes an object_file object and creates the appropriate process object, or dies.
base/loader/elf_object.cc:
Actually extract the OS and architecture from the elf file, rather than always guessing Alpha and Linux.
base/loader/object_file.hh:
Added constants for SPARC, MIPS, and Solaris, and changed the include for the Addr type.
sim/process.cc:
Pushed creation of specific process objects into the ISA specific code.
--HG--
extra : convert_revision :
b4754e7ca8328672d07e1394c4d162e199606b53
Ali Saidi [Fri, 3 Mar 2006 19:25:26 +0000 (14:25 -0500)]
Merge zizzer:/bk/m5
into zeep.eecs.umich.edu:/z/saidi/work/m5.head
--HG--
extra : convert_revision :
8f463682e32a457cd9171c63dcec5209cb850b56
Ali Saidi [Fri, 3 Mar 2006 19:24:15 +0000 (14:24 -0500)]
First cut at moving alpha specefic stuff out of /sim/system* into
arch/alpha/system*.
SConscript:
dev/alpha_console.cc:
dev/alpha_console.hh:
kern/freebsd/freebsd_system.cc:
kern/freebsd/freebsd_system.hh:
kern/kernel_stats.cc:
kern/linux/linux_system.cc:
kern/linux/linux_system.hh:
kern/tru64/tru64_system.cc:
kern/tru64/tru64_system.hh:
sim/pseudo_inst.cc:
sim/system.cc:
sim/system.hh:
First cut at moving alpha specefic stuff out of /sim/system*
--HG--
extra : convert_revision :
eba58f537fc04e605af4fc216e184b074a44c8a0
Nathan Binkert [Fri, 3 Mar 2006 19:17:48 +0000 (14:17 -0500)]
Ethernet devices have an RSS option to tell the driver to
use Receive side scaling
dev/ns_gige.cc:
dev/ns_gige.hh:
dev/ns_gige_reg.h:
dev/sinic.cc:
dev/sinic.hh:
dev/sinicreg.hh:
add support for setting the RSS flag to notify the driver
to use RSS
--HG--
extra : convert_revision :
5f0c11668ae976634b3bf0caad669a9464a4c041
Gabe Black [Fri, 3 Mar 2006 02:44:43 +0000 (21:44 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
--HG--
extra : convert_revision :
f6db244a66431dd6b8c5ba251ed02d76cd509cff
Ali Saidi [Thu, 2 Mar 2006 22:49:12 +0000 (17:49 -0500)]
Fix fault handling, dereferincing a null pointer is bad.
--HG--
extra : convert_revision :
a4f658bfd39b7f1d219ad8929b8f45457f49299f
Gabe Black [Thu, 2 Mar 2006 00:03:42 +0000 (19:03 -0500)]
Got rid of the multiple inheritance in the Fault classes, and the base MachineCheck and Alignment faults.
--HG--
extra : convert_revision :
4b1429ef5c7a11c25e28f2bc868cc1481a08c5d1
Ali Saidi [Wed, 1 Mar 2006 17:43:04 +0000 (12:43 -0500)]
Merge zizzer:/bk/m5
into zeep.eecs.umich.edu:/z/saidi/work/m5.q
--HG--
extra : convert_revision :
4855ecfb3de652456a8be413c9e538c91709dcb6
Gabe Black [Wed, 1 Mar 2006 10:28:26 +0000 (05:28 -0500)]
Begain setting up syscall emulation for sparc linux.
arch/sparc/linux_process.cc:
Began modifying this file for sparc, and replaced the alpha syscall numbers with the sparc ones.
arch/sparc/linux_process.hh:
Begain modifying this file for sparc.
--HG--
extra : convert_revision :
b1ed8ac50a35591521c5feaf1ea9c4b8ccc26dc8
Gabe Black [Wed, 1 Mar 2006 10:26:08 +0000 (05:26 -0500)]
Cleaned up some of the Fault system.
arch/alpha/ev5.cc:
Commented out the intr_post function since it's not used. If this really -is- needed, it should be moved into the fault class.
arch/alpha/faults.cc:
arch/alpha/faults.hh:
Moved the fault invocation code into the fault class fully, and got rid of the need for isA.
cpu/exec_context.cc:
cpu/exec_context.hh:
Removed the trap function from the ExecContext. The faults will execute normally in full system mode, but always panic in syscall emulation mode.
cpu/ozone/cpu.hh:
cpu/simple/cpu.hh:
Changed the execution context executing a fault to a fault executing on the execution context.
sim/faults.cc:
If not in full system mode, trying to invoke a fault causes a panic.
sim/faults.hh:
Removed the isA function.
--HG--
extra : convert_revision :
894dc8f0755c8efc4b7ef5a09fb2cf7373042395
Gabe Black [Wed, 1 Mar 2006 10:21:16 +0000 (05:21 -0500)]
Removed the targetarch code.
--HG--
extra : convert_revision :
ab0c8ed2e8d7359b353f34b443e84094ca40c5e6
Gabe Black [Wed, 1 Mar 2006 10:20:47 +0000 (05:20 -0500)]
Made sparc copies of some architecture specific files.
arch/sparc/linux_process.cc:
arch/sparc/linux_process.hh:
Formatting & doxygen.
arch/sparc/linux_process.cc:
Added some more ioctl commands to ignore.
Set unlink and rename to the new functions.
Ignore setrlimit, times and rt_sigaction.
Should eventually provide a function for times.
arch/sparc/linux_process.cc:
Fixes for Linux emulation:
- stat struct alignment
- osf_{get,set}sysinfo return values
- additional syscall numbers
- initialize $r0 to 0
arch/sparc/linux_process.cc:
arch/sparc/linux_process.hh:
Updated Copyright
arch/sparc/linux_process.cc:
Alphabetize includes.
arch/sparc/linux_process.cc:
Changed all syscalls to use syscall return object
arch/sparc/linux_process.cc:
Removed unecessary constructor call at each return.
arch/sparc/linux_process.cc:
Rename sim/universe.{cc,hh} to root.{cc,hh}.
arch/sparc/linux_process.cc:
shuffle files around for new directory structure
arch/sparc/linux_process.cc:
Remove RCS Id string
arch/sparc/linux_process.cc:
arch/sparc/linux_process.hh:
Update copyright dates and author list
arch/sparc/linux_process.hh:
remove $Id$ string
arch/sparc/linux_process.cc:
Apply patch for syscall emulation provided by Antti Miettinen (apm@brigitte.dna.fi).
arch/sparc/linux_process.cc:
Added the endianness namespace. This may change.
arch/sparc/linux_process.cc:
Replaced the namespace declaration with including arch/alpha/isa_traits.hh
arch/sparc/linux_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
arch/sparc/linux_process.cc:
Add endian conversions to fstat
arch/sparc/linux_process.cc:
fix #if. I wonder why my compiler had no issues
arch/sparc/linux_process.cc:
add endian conversions for fstat functions
arch/sparc/linux_process.cc:
move architecture independent code into kern/linux/linux.(hh|cc)
arch/sparc/linux_process.hh:
Add getDesc function
arch/sparc/linux_process.cc:
Added using directive for AlphaISA namespace
arch/sparc/linux_process.cc:
Add pipeFunc.
arch/alpha/linux_process.cc:
Renamed arch files to remove alpha prefix, and changed alpha_memory.hh and cc to a more accurate tlb.hh and cc
arch/sparc/linux_process.cc:
bk cp ../alpha/linux_process.cc linux_process.cc
arch/sparc/linux_process.hh:
bk cp ../alpha/linux_process.hh linux_process.hh
--HG--
extra : convert_revision :
a8eca1c5f848383263a2326b0e3cde4bcd85f14b
Gabe Black [Wed, 1 Mar 2006 06:23:10 +0000 (01:23 -0500)]
Changed the name of the fault's invocation method from ev5_trap to invoke.
--HG--
extra : convert_revision :
b7ab14ac644f6a38c69aaa5372b3002b21f34af0
Gabe Black [Wed, 1 Mar 2006 05:14:09 +0000 (00:14 -0500)]
Used the Fault constructor to create Interrupt faults. Using the pointer directly prevents reference counting from working correctly.
cpu/simple/cpu.cc:
Used the Fault constructor to create an Interrupt fault.
--HG--
extra : convert_revision :
ec2ef7e3cd04ee9a44628877dc6e770fb907104b
Gabe Black [Wed, 1 Mar 2006 05:09:08 +0000 (00:09 -0500)]
moved ev5_trap fully into the fault class.
--HG--
extra : convert_revision :
182cdec9a4e05f55edff0c6a114844b9ad2ca8db
Ali Saidi [Tue, 28 Feb 2006 23:41:04 +0000 (18:41 -0500)]
Add quiesceNs, quiesceTime, quiesceCycles, and m5panic pseudo ops.
This changeset removes a check that prevents quiescing when an
interrupt is pending. *** You should only call quiesce if that
isn't a problem. ***
arch/alpha/isa/decoder.isa:
sim/pseudo_inst.cc:
sim/pseudo_inst.hh:
Add quiesceNs, quiesceCycles, quisceTime and m5panic pseudo ops.
These quiesce for a number of ns, cycles, report how long
we were quiesced for, and panic the simulator respectively.
The latter is added to the panic() function in the console and linux
kernel instead of executing an infinite loop until someone notices.
cpu/exec_context.cc:
cpu/exec_context.hh:
Add a quiesce end event to the execution contexted which upon
executing wakes up a CPU for quiesceCycles/quiesceNs.
util/m5/Makefile:
Make the makefile more reasonable
util/m5/m5.c:
update the m5op executable to use the files from the linux tree
util/m5/m5op.S:
update m5op.S from linux tree
util/m5/m5op.h:
update m5op.h from linux tree
--HG--
rename : util/m5/m5op.s => util/m5/m5op.S
extra : convert_revision :
3be18525e811405b112e33f24a8c4e772d15462d
Kevin Lim [Tue, 28 Feb 2006 20:16:24 +0000 (15:16 -0500)]
Merge ktlim@zizzer:/bk/m5
into zamp.eecs.umich.edu:/z/ktlim2/clean/m5-clean
arch/alpha/ev5.cc:
cpu/o3/regfile.hh:
Hand merge.
--HG--
rename : arch/alpha/alpha_memory.cc => arch/alpha/tlb.cc
extra : convert_revision :
c941dd2198851398820b38a66471372ed8454891
Gabe Black [Tue, 28 Feb 2006 11:28:09 +0000 (06:28 -0500)]
Corrected some mistakes in the hand merge
--HG--
extra : convert_revision :
84d852e625f754c4414403e8c3e26bda80500d4f
Gabe Black [Tue, 28 Feb 2006 11:17:57 +0000 (06:17 -0500)]
Hand merged
--HG--
extra : convert_revision :
0fd5ecbba5dc6299dd046926ec27595ca4d97ee7
Gabe Black [Tue, 28 Feb 2006 11:13:35 +0000 (06:13 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
--HG--
extra : convert_revision :
0b3ffc0605c9043d7f5bf6c15f4a3c68846a732a
Gabe Black [Tue, 28 Feb 2006 11:03:57 +0000 (06:03 -0500)]
Merge zizzer.eecs.umich.edu:/z/m5/Bitkeeper/m5
into zizzer.eecs.umich.edu:/z/m5/Bitkeeper/multiarch
--HG--
extra : convert_revision :
3a3ff8fcf472db385219a00ae434d1f3fea43b18
Gabe Black [Tue, 28 Feb 2006 11:02:18 +0000 (06:02 -0500)]
Cleaned up and slightly reorganized the Fault class heirarchy.
arch/alpha/ev5.cc:
Changed c style casts of Faults to dynamic_casts
arch/alpha/faults.cc:
AlphaFault is now an abstract class.
arch/alpha/faults.hh:
AlphaFault is now an abstract class. Also, AlphaMachineCheckFault and AlphaAlignmentFault multiply inherit from both AlphaFault and from MachineCheckFault and AlignmentFault respectively. These classes get their name from the generic classes.
cpu/o3/alpha_cpu_impl.hh:
Changed a c style cast to a dynamic_cast for a Fault
sim/faults.hh:
All generic Fault classes are now abstract. Also, MachineCheckFault and AlignmentFault inherit FaultBase as a virtual base class to help resolve ambiguities when they are multiply inherited in ISA specific classes. The override the isMachineCheckFault and isAlignmentFault functions appropriately, and provide a standard name for these faults.
--HG--
extra : convert_revision :
2cb906708e3eaec4a12587484c09e50ed6ef88fc
Gabe Black [Tue, 28 Feb 2006 04:26:13 +0000 (23:26 -0500)]
Changed ev5_trap from a function of the execution context to a function of the fault. The actual function still resides in the execution context.
--HG--
extra : convert_revision :
56e33536cdd9079ace03896b85ea3c84b6eb4e57
Gabe Black [Mon, 27 Feb 2006 21:46:00 +0000 (16:46 -0500)]
Moved the _stat for MachineCheckFault and AlignmentFault into the isa specific classes to prevent instantiation of the generic classes.
--HG--
extra : convert_revision :
0378261342df008c2bf6d260295ef21b15f119a8
Gabe Black [Mon, 27 Feb 2006 21:27:01 +0000 (16:27 -0500)]
Got rid of the fault_addr function.
--HG--
extra : convert_revision :
deb54cd82db47abb6d9bac76e072f2a4b1c883b2
Kevin Lim [Mon, 27 Feb 2006 17:09:08 +0000 (12:09 -0500)]
Merge ktlim@zizzer:/bk/m5
into zamp.eecs.umich.edu:/z/ktlim2/clean/m5-clean
--HG--
extra : convert_revision :
97c345f0715a347ce34f9cabd994485f30f2e171
Kevin Lim [Mon, 27 Feb 2006 16:44:35 +0000 (11:44 -0500)]
Changes to put all the misc regs within the misc reg file. This includes the FPCR, Uniq, lock flag, lock addr, and IPRs.
They are now accessed by calling readMiscReg()/setMiscReg() on the XC. Old IPR accesses are supported by using readMiscRegWithEffect() and setMiscRegWithEffect() (names may change in the future).
arch/alpha/alpha_memory.cc:
Change accesses to IPR to go through the XC.
arch/alpha/ev5.cc:
Change accesses for IPRs to go through the misc regs.
arch/alpha/isa/decoder.isa:
Change accesses to IPRs to go through the misc regs. readIpr() and setIpr() are now changed to calls to readMiscRegWithEffect() and setMiscRegWithEffect().
arch/alpha/isa/fp.isa:
Change accesses to IPRs and Fpcr to go through the misc regs.
arch/alpha/isa/main.isa:
Add support for all misc regs being accessed through readMiscReg() and setMiscReg(). Instead of readUniq and readFpcr, they are replaced by calls with Uniq_DepTag and Fpcr_DepTag passed in as the register index.
arch/alpha/isa_traits.hh:
Change the MiscRegFile to a class that handles all accesses to MiscRegs, which in Alpha include the FPCR, Uniq, Lock Addr, Lock Flag, and IPRs.
Two flavors of accesses are supported: normal register reads/writes, and reads/writes with effect. The latter are basically the original read/write IPR functions, while the former are normal reads/writes.
The lock flag and lock addr registers are added to the dependence tags in order to support being accessed through the misc regs.
arch/alpha/stacktrace.cc:
cpu/simple/cpu.cc:
dev/sinic.cc:
Change accesses to the IPRs to go through the XC.
arch/alpha/vtophys.cc:
Change access to the IPR to go through the XC.
arch/isa_parser.py:
Change generation of code for control registers to use the readMiscReg and setMiscReg functions.
base/remote_gdb.cc:
Change accesses to the IPR to go through the XC.
cpu/exec_context.hh:
Use the miscRegs to access the lock addr, lock flag, and other misc registers.
cpu/o3/alpha_cpu.hh:
cpu/simple/cpu.hh:
Support interface for reading and writing misc registers, which replaces readUniq, readFpcr, readIpr, and their set functions.
cpu/o3/alpha_cpu_impl.hh:
Change accesses to the IPRs to go through the miscRegs.
For now comment out some of the accesses to the misc regs until the proxy exec context is completed.
cpu/o3/alpha_dyn_inst.hh:
Change accesses to misc regs to use readMiscReg and setMiscReg.
cpu/o3/alpha_dyn_inst_impl.hh:
Remove old misc reg accessors.
cpu/o3/cpu.cc:
Comment out old misc reg accesses until the proxy exec context is completed.
cpu/o3/cpu.hh:
Change accesses to the misc regs.
cpu/o3/regfile.hh:
Remove old access methods for the misc regs, replace them with readMiscReg and setMiscReg. They are dummy functions for now until the proxy exec context is completed.
kern/kernel_stats.cc:
kern/system_events.cc:
Have accesses to the IPRs go through the XC.
kern/tru64/tru64.hh:
Have accesses to the misc regs use the new access methods.
--HG--
extra : convert_revision :
e32e0a3fe99522e17294bbe106ff5591cb1a9d76
Gabe Black [Mon, 27 Feb 2006 11:05:10 +0000 (06:05 -0500)]
Renamed arch files to remove alpha prefix, and changed alpha_memory.hh and cc to a more accurate tlb.hh and cc
--HG--
rename : arch/alpha/alpha_linux_process.cc => arch/alpha/linux_process.cc
rename : arch/alpha/alpha_linux_process.hh => arch/alpha/linux_process.hh
rename : arch/alpha/alpha_memory.cc => arch/alpha/tlb.cc
rename : arch/alpha/alpha_memory.hh => arch/alpha/tlb.hh
rename : arch/alpha/alpha_tru64_process.cc => arch/alpha/tru64_process.cc
rename : arch/alpha/alpha_tru64_process.hh => arch/alpha/tru64_process.hh
extra : convert_revision :
9fe6863cc37347d2a6bd716c529b0a4a50b36ea7
Gabe Black [Mon, 27 Feb 2006 10:35:43 +0000 (05:35 -0500)]
Changed targetarch to just arch.
SConscript:
Changed all of the "targetarch" files to be "arch" files
arch/SConscript:
Added all of the remaining targetarch files to isa_switch_hdrs
--HG--
extra : convert_revision :
a37d18349e27cf92dce12814f21944daa7fe9480
Gabe Black [Mon, 27 Feb 2006 09:05:02 +0000 (04:05 -0500)]
Fixed up some include paths.
kern/kernel_stats.cc:
Replaced targetarch with arch/alpha, since osfpal.hh is not a targetarched file.
--HG--
extra : convert_revision :
24a25f27e5912cb949568cf5fd0ab856771ba0d1
Gabe Black [Mon, 27 Feb 2006 09:02:45 +0000 (04:02 -0500)]
Added isMachineCheckFault and isAlignmentFault virtual functions to the fault base class, and replaced the isA templated function with them where appropriate.
arch/alpha/ev5.cc:
cpu/simple/cpu.cc:
Changed from the isA templated function to isMachineCheckFault and isAlignmentFault
sim/faults.hh:
Added isMachineCheckFault and isAlignmentFault virtual functions to the fault base class.
--HG--
extra : convert_revision :
3bf3a4369bc24a039648ee4f2a9c1663362ff2e2
Gabe Black [Mon, 27 Feb 2006 09:00:24 +0000 (04:00 -0500)]
Put the Alpha faults into the AlphaISA namespace
--HG--
extra : convert_revision :
d9586bb8ffaab74673aa0b3d17bb676b949ff212
Gabe Black [Mon, 27 Feb 2006 08:57:15 +0000 (03:57 -0500)]
MachineCheckFaults and AlignmentFaults are now generated by the ISA, rather than being created directly.
arch/alpha/alpha_memory.cc:
cpu/base_dyn_inst.cc:
dev/alpha_console.cc:
dev/pcidev.hh:
dev/sinic.cc:
MachineCheckFaults are now generated by the ISA, rather than being created directly.
--HG--
extra : convert_revision :
34a7da41639e93be21ed70dac681b27480008d19
Gabe Black [Mon, 27 Feb 2006 06:38:47 +0000 (01:38 -0500)]
Changed targetarch to arch for isa_traits.hh include
--HG--
extra : convert_revision :
e7ae43d812140ec99b782394e54903153f8d0476
Gabe Black [Mon, 27 Feb 2006 06:32:49 +0000 (01:32 -0500)]
Where architecture independent sources included arch/alpha/xxx.hh, they were changed to include targetarch/xxx.hh
cpu/base_dyn_inst.cc:
cpu/o3/bpred_unit.hh:
cpu/o3/comm.hh:
cpu/o3/cpu.hh:
cpu/o3/regfile.hh:
cpu/ozone/cpu_impl.hh:
cpu/ozone/ea_list.cc:
cpu/ozone/ea_list.hh:
kern/kernel_stats.cc:
Changed arch/alpha to targetarch
sim/process.cc:
Changed arch/alpha to targetarch, and removed gaurding ifdef
--HG--
extra : convert_revision :
3c29e6baeb1cd900f7b5e11144a5d547a6c7c5ab
Nathan Binkert [Mon, 27 Feb 2006 04:06:21 +0000 (23:06 -0500)]
fix some minor stats stuff
util/stats/stats.py:
fix up a few of the stats.
--HG--
extra : convert_revision :
61c600cd146900c8cc0cf60a9036f2e999a5e244
Nathan Binkert [Mon, 27 Feb 2006 01:31:08 +0000 (20:31 -0500)]
add some support for random access of data in packet fifos
dev/pktfifo.cc:
add support for copying arbitrary data out of a
packet fifo
dev/pktfifo.hh:
add support for copying arbitrary data out of a
packet fifo.
Add functions to determine where in the fifo a
particular packet is
--HG--
extra : convert_revision :
f8ddc994ce8577f29af0de3fa418a01e4e2cb0f1
Nathan Binkert [Sun, 26 Feb 2006 15:44:01 +0000 (10:44 -0500)]
Allow graph_group to not be selected so we can have a
normal ungrouped barchart
--HG--
extra : convert_revision :
7d55440c9bb060607eddbb72448a3413944bb6ba
Nathan Binkert [Sun, 26 Feb 2006 06:00:15 +0000 (01:00 -0500)]
better function categorization
util/stats/categories.py:
bit more stuff for categorizing functions
--HG--
extra : convert_revision :
03617246a9254a580684dce82836517d1efdfc5b
Nathan Binkert [Sun, 26 Feb 2006 05:57:37 +0000 (00:57 -0500)]
fix small python bug in database processing code
util/stats/db.py:
fix usage of hasattr
--HG--
extra : convert_revision :
b384e1efeda76921c565f9f391694c27273edcec
Nathan Binkert [Sun, 26 Feb 2006 05:35:10 +0000 (00:35 -0500)]
code cleanup
util/stats/barchart.py:
clean up some of lisa's messy code
remove trailing whitespace while I'm at it.
--HG--
extra : convert_revision :
f2fe6777fb4b458fa1d5b5b743f6274014c229ad
Nathan Binkert [Sun, 26 Feb 2006 05:19:02 +0000 (00:19 -0500)]
forgot to add a chart option
util/stats/chart.py:
add a bool config option for determining
if the legend is inside or outside the figure
--HG--
extra : convert_revision :
e862d1832a0cc3c1837758cc247bc77c0a02ec12
Nathan Binkert [Sun, 26 Feb 2006 05:11:54 +0000 (00:11 -0500)]
add error bars and more options for legend placement
util/stats/barchart.py:
Add support for error bars
util/stats/barchart.py:
add support to choose between a legend inside or
outside the figure.
--HG--
extra : convert_revision :
14273e385c106bf27a2013991f9f34ca6551b96c
Nathan Binkert [Sun, 26 Feb 2006 04:48:13 +0000 (23:48 -0500)]
better colors for barcharts
util/stats/barchart.py:
If there are fewer than 5 colors, pick from a subset of
5 so there is more consistency in colors between graphs
--HG--
extra : convert_revision :
6cf64c2f8ed81e714e24a3ebe5a7a60ca168b231
Steve Reinhardt [Sun, 26 Feb 2006 03:57:46 +0000 (22:57 -0500)]
Make sure cpu/static_inst_exec_sigs.hh get rebuilt when
CPU_MODELS parameter changes.
arch/SConscript:
Fix typo in comment.
cpu/SConscript:
Convert exec signature generator to Action so we can add dependency
on CPU_MODELS environment var.
Print nicer string while we're at it.
Also add some comments.
--HG--
extra : convert_revision :
bcb38a7941943cf071dac34cdbb2ece5456b8620
Nathan Binkert [Sun, 26 Feb 2006 03:03:47 +0000 (22:03 -0500)]
only build libelf.a, forget about the other
libelf junk.
--HG--
extra : convert_revision :
964473c0ff1fc2f8fd9fbb8a1533eb3730b61fac
Nathan Binkert [Sun, 26 Feb 2006 03:01:05 +0000 (22:01 -0500)]
Since the delayed write stuff is gone, get rid of regWrite
and merge it with writeBar0
--HG--
extra : convert_revision :
354642e0d528b6a5a7f2cdf0264d93e738b2d4eb
Gabe Black [Fri, 24 Feb 2006 23:45:46 +0000 (18:45 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
--HG--
extra : convert_revision :
d3e160f6e938af5d2118883f8d7c91fa29a0ccaa
Gabe Black [Fri, 24 Feb 2006 23:45:28 +0000 (18:45 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
SConscript:
arch/alpha/ev5.cc:
dev/alpha_console.cc:
Hand merged
--HG--
extra : convert_revision :
318a671e6803400d3ed086a90e70d6790e4f6b19
Gabe Black [Fri, 24 Feb 2006 23:33:57 +0000 (18:33 -0500)]
Merge zizzer.eecs.umich.edu:/z/m5/Bitkeeper/m5
into zizzer.eecs.umich.edu:/z/m5/Bitkeeper/multiarch
--HG--
extra : convert_revision :
c167d2bec89120258124bf34ada3adf23e3e8188
Lisa Hsu [Fri, 24 Feb 2006 23:08:55 +0000 (18:08 -0500)]
Merge zizzer:/bk/m5
into zed.eecs.umich.edu:/z/hsul/work/m5/clean
--HG--
extra : convert_revision :
34314698d4248a078c7b43125b2d048280ff576d
Lisa Hsu [Fri, 24 Feb 2006 23:08:14 +0000 (18:08 -0500)]
1) make it pretty for large clusters
2) make subticks vertical so they can be longer
3) make inner and outer axes farther apart to make room for subtick's vertical labels
--HG--
extra : convert_revision :
91a1aab3f1078921edd53428e6712744210c9f1b
Steve Reinhardt [Fri, 24 Feb 2006 13:52:38 +0000 (08:52 -0500)]
Merge zizzer.eecs.umich.edu:/z/m5/Bitkeeper/m5
into zizzer.eecs.umich.edu:/z/stever/bk/multiarch
arch/isa_parser.py:
SCCS merged
--HG--
extra : convert_revision :
080cca7616b37db3bf18976b63b3dbcb47d8b918
Gabe Black [Fri, 24 Feb 2006 08:51:21 +0000 (03:51 -0500)]
Removed a stray ::.
--HG--
extra : convert_revision :
f6114b78e30e8cba5af6276042b0f043d8773739
Gabe Black [Fri, 24 Feb 2006 06:51:45 +0000 (01:51 -0500)]
Changed Fault from a FaultBase * to a RefCountingPtr, added "new"s where appropriate, and took away the constant examples of each fault which where for comparing to a fault to determine its type.
arch/alpha/alpha_memory.cc:
arch/alpha/isa/decoder.isa:
Added news where faults are created.
arch/alpha/ev5.cc:
Changed places where a fault was compared to a fault type to use isA rather than ==
arch/alpha/faults.cc:
arch/alpha/faults.hh:
Changed Fault to be a RefCountingPtr
arch/alpha/isa/fp.isa:
Added a new where a FloatEnableFault was created.
arch/alpha/isa/unimp.isa:
arch/alpha/isa/unknown.isa:
Added a new where an UnimplementedFault is created.
base/refcnt.hh:
Added include of stddef.h for the NULL macro
cpu/base_dyn_inst.cc:
Added a new where an UnimplementedOpcodeFault is created.
cpu/o3/alpha_cpu_impl.hh:
Changed places where a fault was compared to a fault type to use isA rather than ==. Also changed fault->name to fault->name()
cpu/o3/regfile.hh:
Added new where UnimplementedOpcodeFaults are created.
cpu/simple/cpu.cc:
Changed places where a fault was compared to a fault type to use isA rather than ==. Also added a new where an Interrupt fault is created.
dev/alpha_console.cc:
Added news where MachineCheckFaults are created.
dev/pcidev.hh:
Added news where MachineCheckFaults are generated.
dev/sinic.cc:
Changed places where a fault was compared to a fault type to use isA rather than ==. Added news where MachineCheckFaults are created. Fixed a problem where m5.fast had unused variables.
kern/kernel_stats.cc:
Commented out where _faults is initialized. This statistic will probably be moved elsewhere in the future.
kern/kernel_stats.hh:
Commented out the declaration of _fault. when fault() is called, the fault increments its own stat.
sim/faults.cc:
sim/faults.hh:
Changed Fault from a FaultBase * to a RefCountingPtr.
--HG--
extra : convert_revision :
b40ccfc42482d5a115e111dd897fa378d23c6c7d
Gabe Black [Fri, 24 Feb 2006 00:32:38 +0000 (19:32 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
--HG--
extra : convert_revision :
77dc4dda9c72c871364e112ae8b10669b8d8fc86
Korey Sewell [Thu, 23 Feb 2006 23:46:12 +0000 (18:46 -0500)]
name changes ...
SConscript:
change to alpha_memory.hh
--HG--
rename : arch/alpha/memory.cc => arch/alpha/alpha_memory.cc
rename : arch/alpha/memory.hh => arch/alpha/alpha_memory.hh
extra : convert_revision :
62b1a41de22701160f04cb7a78242746cfcde819
Steve Reinhardt [Thu, 23 Feb 2006 22:00:29 +0000 (17:00 -0500)]
Enable building only selected CPU models via new scons
CPU_MODELS parameter. For example:
scons CPU_MODELS="SimpleCPU,FullCPU" ALPHA_SE/m5.debug
Unfortunately the option is not sticky due to a scons
bug with saving & restoring ListOption parameters.
SConscript:
Separate out cpu-model-specific files so they can be conditionally
included based on value of new CPU_MODELS parameter.
Most of these are now handled in cpu/SConscript, except for FullCPU
which is still in this file.
arch/SConscript:
The set of CPU-model-specific execute files must now be
determined from the CPU_MODELS parameter, via the new
cpu_models.py file.
Also pass the list of configured CPU models to isa_parser.py.
arch/isa_parser.py:
Move CpuModel definition and objects out to a
separate file so they can be shared with scons.
Global list of CPU models to generate code for is now
controlled by command-line parameters (so we can do
only a subset of the available ones).
build/SConstruct:
Define new CPU_MODELS ListOption.
cpu/static_inst.hh:
Rename static_inst_impl.hh to static_inst_exec_sigs.hh.
--HG--
extra : convert_revision :
163df32a76d4c05900490b2bce4c7962a5e3f614
Ali Saidi [Thu, 23 Feb 2006 20:08:08 +0000 (15:08 -0500)]
ev5.cc:
SCCS merged
arch/alpha/ev5.cc:
SCCS merged
--HG--
extra : convert_revision :
9d70c1d461dab0ec016fd0616d74a49942aac659
Ali Saidi [Thu, 23 Feb 2006 20:06:06 +0000 (15:06 -0500)]
Merge zizzer:/bk/m5
into zeep.eecs.umich.edu:/z/saidi/work/m5.head
cpu/simple/cpu.cc:
remove initCPU from constructor
dev/alpha_console.cc:
we are panicing, so no need to return a fault
--HG--
extra : convert_revision :
72389ea0c96e91a55f35b884200325224bfb6ed9
Ali Saidi [Thu, 23 Feb 2006 19:50:16 +0000 (14:50 -0500)]
Get rid of the xc from the alphaAccess/alphaConsole backdoor device.
Now allocate an array of stacks indexed by cpu number which specify
cpu stacks and are initialized by cpu 0. Othe cpus spin waiting for
their stacks before continuing. This change *REQUIRES* a the new
console code to operate correctly.
arch/alpha/ev5.cc:
Add cpuId to initCPU/initIPR functions
cpu/o3/cpu.cc:
cpu/simple/cpu.cc:
cpu/simple/cpu.hh:
Move the cpu initilization into an init() function since it now needs
the CPU id which isn't known at construction
dev/alpha_access.h:
dev/alpha_console.cc:
dev/alpha_console.hh:
instead of the bootstrap variables, add space for 64 cpu stacks in the
alpha access structure.
sim/system.cc:
start all cpus immediately rather than just the first one
--HG--
extra : convert_revision :
28c218af49d885a0f203ada419f16f25d5a3f37b
Steve Reinhardt [Thu, 23 Feb 2006 19:31:15 +0000 (14:31 -0500)]
Create a Builder object for .isa files in arch/SConscript.
Start using SCons File objects to avoid fixed paths in
subordinate SConscripts.
SConscript:
Push isa_parser stuff (including .isa scanner) down into
arch/SConscript.
arch/SConscript:
Create a Builder object for .isa files, including existing scanner.
Return file objects generated by isa-specific SConscript
back up to parent.
arch/alpha/SConscript:
arch/mips/SConscript:
arch/sparc/SConscript:
Convert sources to scons File objects, so file names can be specified
relative to the current directory.
Invoke new builder for isa description, and get generated sources from
there (instead of listing them explicitly).
arch/isa_parser.py:
Get rid of third argument ("include_path").
It was a pain to generate this from scons, and it turned out
it's not needed anyway, since the only included file
(decoder.hh) will be in the same directory as the sources.
--HG--
extra : convert_revision :
36861bcef36763f229704d8cb7a642b4486a3581
Kevin Lim [Thu, 23 Feb 2006 18:27:23 +0000 (13:27 -0500)]
Cast enum to int to fix compile error from regression.
--HG--
extra : convert_revision :
2d998ec7393e1c08c7552f7e586160d579eab2b1
Steve Reinhardt [Thu, 23 Feb 2006 13:17:09 +0000 (08:17 -0500)]
Merge zizzer.eecs.umich.edu:/z/m5/Bitkeeper/m5
into zizzer.eecs.umich.edu:/z/stever/bk/m5
--HG--
extra : convert_revision :
38367d6d1fa594b84b64e5930588904719a40c08
Steve Reinhardt [Thu, 23 Feb 2006 13:16:59 +0000 (08:16 -0500)]
Add pipe() syscall to Alpha Linux emulation.
arch/alpha/alpha_linux_process.cc:
Add pipeFunc.
--HG--
extra : convert_revision :
c094d2dff993d5e60bc43b7cd4b9586c15c634a3
Gabe Black [Thu, 23 Feb 2006 09:11:09 +0000 (04:11 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
--HG--
extra : convert_revision :
d92c611475aef2b911f76e2c21be12096e73048a
Gabe Black [Thu, 23 Feb 2006 09:08:55 +0000 (04:08 -0500)]
Merge zizzer.eecs.umich.edu:/z/m5/Bitkeeper/m5
into zizzer.eecs.umich.edu:/z/m5/Bitkeeper/multiarch
--HG--
extra : convert_revision :
850077a56aead260aa4bbd3df60b672a931d57ed
Nathan Binkert [Thu, 23 Feb 2006 05:21:35 +0000 (00:21 -0500)]
don't add an empty suboption description
--HG--
extra : convert_revision :
594744c3d438aed08a23db376959930071b2c368
Nathan Binkert [Thu, 23 Feb 2006 05:20:32 +0000 (00:20 -0500)]
make it possible to add filters for job names so that
parts of the full crossproduct of jobs can be ignored.
--HG--
extra : convert_revision :
c44b3daea0cf4b487b1d99eae92da16573b15930
Nathan Binkert [Thu, 23 Feb 2006 05:00:01 +0000 (00:00 -0500)]
fix stat name
--HG--
extra : convert_revision :
4c0f8c1b6ba1e77dc302b8ece3c8822b9b4a05ee
Steve Reinhardt [Thu, 23 Feb 2006 03:22:06 +0000 (22:22 -0500)]
Auto-generate arch/foo.hh "switch headers" in scons.
SConscript:
Include new arch/SConscript file.
arch/isa_specific.hh:
Get rid of unnecessary ISA_INCLUDE() macro and other
things that were used only for that purpose.
build/SConstruct:
Move list of ISAs to env var ALL_ISA_LIST.
--HG--
extra : convert_revision :
612c7ee4279d57209662be88dc577d80fdbd692c
Steve Reinhardt [Thu, 23 Feb 2006 02:11:45 +0000 (21:11 -0500)]
Clean excess comments out of SConscripts.
SConscript:
arch/alpha/SConscript:
Clean out excess comments.
--HG--
extra : convert_revision :
7aae68d36f9fce5f236d117d803b5e3cd4a3769d
Gabe Black [Thu, 23 Feb 2006 01:52:03 +0000 (20:52 -0500)]
Cleaned up the mapping of isa_parser.py inputs to outputs.
--HG--
extra : convert_revision :
a8431a67001b2916eb8d0548f1f34e1c948bb356
Korey Sewell [Wed, 22 Feb 2006 09:08:08 +0000 (04:08 -0500)]
make sure alpha still compiles , rename files back to original naming ...
Now that we have decoder.do, add new files so we can start compiling other files
needed for MIPS syscall emulation mode
arch/mips/linux_process.cc:
arch/mips/linux_process.hh:
New MIPS-specific file
--HG--
rename : arch/alpha/linux_process.cc => arch/alpha/alpha_linux_process.cc
rename : arch/alpha/linux_process.hh => arch/alpha/alpha_linux_process.hh
rename : arch/alpha/tru64_process.cc => arch/alpha/alpha_tru64_process.cc
rename : arch/alpha/tru64_process.hh => arch/alpha/alpha_tru64_process.hh
extra : convert_revision :
2bfc27e8772523cbeb95f40684f9a32fe5554f87
Korey Sewell [Wed, 22 Feb 2006 08:44:21 +0000 (03:44 -0500)]
Still builds "../MIPS_SE/arch/mips/decoder.do with no errors!
-----
uncomment out detailed model ... just commented to supress some compile errors
arch/isa_parser.py:
uncomment out detailed model ... just commented to supress some compile errors
--HG--
extra : convert_revision :
e884b9bd47794409f74043ad1aca6dadd1323185
Korey Sewell [Wed, 22 Feb 2006 08:33:35 +0000 (03:33 -0500)]
MIPS Compiles scons/MIPS_SE/arch/mips/decoder.do!!!!!!
arch/mips/faults.hh:
remove nonsense
arch/mips/isa/base.isa:
define R31
arch/mips/isa/bitfields.isa:
forgotten bitfields
arch/mips/isa/decoder.isa:
INT64 -> int64_t
arch/mips/isa/formats.isa:
fix comments
arch/mips/isa/formats/branch.isa:
Branch -> BranchLikely
RB -> RT
arch/mips/isa/formats/fp.isa:
Make FP ops generates
arch/mips/isa/formats/mem.isa:
RA,RB -> RS,RT
arch/mips/isa/formats/noop.isa:
Rc -> Rd
arch/mips/isa/formats/util.isa:
forgot brace and semicolon
arch/mips/isa/includes.isa:
remove unnecessary files
arch/mips/isa_traits.hh:
spacing
cpu/static_inst.hh:
add cond_delay_slot flag
--HG--
extra : convert_revision :
3bc7353b437f9a764e85cc462bed86c9d654eb37
Korey Sewell [Wed, 22 Feb 2006 03:06:18 +0000 (22:06 -0500)]
Merge zizzer:/bk/multiarch
into zazzer.eecs.umich.edu:/z/ksewell/research/m5-sim/multiarch
--HG--
rename : arch/alpha/alpha_memory.cc => arch/alpha/memory.cc
rename : arch/alpha/alpha_memory.hh => arch/alpha/memory.hh
extra : convert_revision :
c641ba3c1009829b7276279b2dca441be1da5b30
Korey Sewell [Wed, 22 Feb 2006 03:02:05 +0000 (22:02 -0500)]
Renaming alpha files and changing some MIPS stuff to be more like Alpha version
SConscript:
changed the alpha_memory.hh to memory.hh in SConscript
arch/isa_parser.py:
temporarily comment out o3 model
arch/mips/isa/base.isa:
arch/mips/isa_traits.cc:
arch/mips/isa_traits.hh:
Fix Up Base Class to mirror how Alpha generates StaticInsts
arch/mips/faults.cc:
MIPS fault.cc file
arch/mips/faults.hh:
MIPS fault.hh file
--HG--
rename : arch/alpha/alpha_linux_process.cc => arch/alpha/linux_process.cc
rename : arch/alpha/alpha_linux_process.hh => arch/alpha/linux_process.hh
rename : arch/alpha/alpha_memory.cc => arch/alpha/memory.cc
rename : arch/alpha/alpha_memory.hh => arch/alpha/memory.hh
rename : arch/alpha/alpha_tru64_process.cc => arch/alpha/tru64_process.cc
rename : arch/alpha/alpha_tru64_process.hh => arch/alpha/tru64_process.hh
extra : convert_revision :
f92d6e765ca96a8b952aef79ed119fa29464563b
Gabe Black [Wed, 22 Feb 2006 01:10:40 +0000 (20:10 -0500)]
Changed Fault * to Fault, which is a typedef to FaultBase *, which is the old Fault class renamed.
--HG--
extra : convert_revision :
5b2f457401f8ff94fe39fe071288eb117814b7bb
Gabe Black [Tue, 21 Feb 2006 08:38:21 +0000 (03:38 -0500)]
Made Addr a global type
--HG--
extra : convert_revision :
869bd9fa5d8591115ac9b4a7401eb2490986b835
Gabe Black [Tue, 21 Feb 2006 04:55:25 +0000 (23:55 -0500)]
Merge zizzer.eecs.umich.edu:/z/m5/Bitkeeper/m5
into zizzer.eecs.umich.edu:/z/m5/Bitkeeper/multiarch
--HG--
extra : convert_revision :
da72b3593037c2a67a56c799e292853b8aece907
Gabe Black [Tue, 21 Feb 2006 04:54:38 +0000 (23:54 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
--HG--
extra : convert_revision :
9d386beecc6d13625ff19ca72cbc3628dcd59d3c
Gabe Black [Tue, 21 Feb 2006 04:53:14 +0000 (23:53 -0500)]
Finished the implementing the change of the ISA from a class to a namespace
dev/sinic.cc:
When DPRINTF disappears, reg32 becomes an unused variable. With -Werror, this causes the compile to fail.
--HG--
extra : convert_revision :
c003c714228491e060155070d192521c53d9e929
Nathan Binkert [Tue, 21 Feb 2006 04:41:50 +0000 (23:41 -0500)]
Get rid of the code that delays PIO write accesses
until the cache access occurs. The fundamental problem
is that a subsequent read that occurs functionally will
get a functionally incorrect result that can break
driver code.
dev/ns_gige.cc:
dev/ns_gige.hh:
dev/sinic.cc:
dev/sinic.hh:
get rid of pio_delay write and the associated code to move
the write to the cache access function
dev/sinicreg.hh:
no more write delays
python/m5/objects/Ethernet.py:
get rid of pio_delay write
--HG--
extra : convert_revision :
1dcb51b8f4514e717bc334a782dfdf06d29ae69d
Korey Sewell [Mon, 20 Feb 2006 19:48:10 +0000 (14:48 -0500)]
make MIPS specific
--HG--
extra : convert_revision :
c019fad60fbf1a316bc6201b8ce8acf5a9875989
Korey Sewell [Mon, 20 Feb 2006 19:30:23 +0000 (14:30 -0500)]
load/store instruction format ... now generates load/store code
and breaks it into a separate EA and MemAccess templated
from how the Alpha ARch. was coded to do the same thing.
arch/mips/isa/bitfields.isa:
comment change
arch/mips/isa/decoder.isa:
re-structuring of load/store instruction definitions
arch/mips/isa/formats/mem.isa:
Define LoadMemory & Store Memory formats
Use style of formatting & base class similar to what was used for ALPHA
arch/mips/isa/formats/util.isa:
Insert LoadStoreBase function here from alpha/arch/isa/mem.isa
arch/mips/isa/operands.isa:
change shw->sh and uhw->uh
--HG--
extra : convert_revision :
5d85f15f4a600dd4c473a3b4a170ba39cf07fc8a
Korey Sewell [Mon, 20 Feb 2006 06:49:16 +0000 (01:49 -0500)]
Support for All Jump Instructions ...
Redo format for Branches and Jumps ( Must update NNPC not NPC )
Now all branches and jumps look like they auto-generate correctly from isa_parser.py!!!
arch/mips/isa/decoder.isa:
Support for All Jump Instructions ..
arch/mips/isa/formats/branch.isa:
Redo format for Branches and Jumps ( Must update NNPC not NPC )
arch/mips/isa/formats/util.isa:
define clear_exe_inst_hazards for later use
--HG--
extra : convert_revision :
63618ed12ee6ed94c47d29619cc1cab2cbaf5cda
Gabe Black [Sun, 19 Feb 2006 09:00:05 +0000 (04:00 -0500)]
Reapplied changes which were undone by a pull
arch/alpha/faults.hh:
kern/linux/linux.hh:
Added typedef for Addr
kern/tru64/tru64.hh:
Fixed up namespaces
--HG--
extra : convert_revision :
bf968e615bc0acc96abeb0eec0872f5b02b5a065
Gabe Black [Sun, 19 Feb 2006 08:20:05 +0000 (03:20 -0500)]
Merge zizzer.eecs.umich.edu:/z/m5/Bitkeeper/m5
into zizzer.eecs.umich.edu:/z/m5/Bitkeeper/multiarch
arch/alpha/faults.hh:
ur
Using cleaned up fault class dei
\7ff
\7ffinitions
--HG--
extra : convert_revision :
a600950d539be2be73358f072aa5426456bf3d2d
Gabe Black [Sun, 19 Feb 2006 08:04:44 +0000 (03:04 -0500)]
Remade some changes which were undone
cpu/base.hh:
cpu/static_inst.hh:
Changed include of targetarch/isa_traits.hh back to arch/isa_traits.hh
cpu/exec_context.hh:
Changed Fault back to Fault *
--HG--
extra : convert_revision :
410f2e2472f8aa5bf92619a5defdf85f689a5597
Gabe Black [Sun, 19 Feb 2006 07:34:52 +0000 (02:34 -0500)]
Merge gblack@m5.eecs.umich.edu:/bk/multiarch
into ewok.(none):/home/gblack/m5/multiarch
--HG--
extra : convert_revision :
090b30a7f70294e1aeb13ba0bc15da4061bdf348