Commit graph

4287 commits

Author SHA1 Message Date
Kevin Lim
4e0ec56868 Update stats for changes.
--HG--
extra : convert_revision : a24c4cd7e2fcd732f5da5679f0c0fbf205f22815
2007-03-25 01:05:48 -04:00
Kevin Lim
5c044cf1f6 Update for new trace data behavior.
--HG--
extra : convert_revision : c3df20c5187614febc4cc9f4d4c68bfecfba1ea7
2007-03-24 23:47:14 -05:00
Kevin Lim
4bad33ce9d Merge ktlim@zizzer:/bk/newmem
into  zamp.eecs.umich.edu:/z/ktlim2/clean/tmp/clean2

--HG--
extra : convert_revision : f3d193dd1e0b82c496d8224f014123b7cb028c02
2007-03-24 14:00:16 -04:00
Gabe Black
5e61e07613 Added a SPARC_SE simple atomic regression for the mcf benchmark.
--HG--
extra : convert_revision : 2284e41c03659db8fc8f284f7d9b587a3708fadf
2007-03-24 02:14:24 -05:00
Gabe Black
e7bbd85ae6 Merge zizzer.eecs.umich.edu:/bk/newmem
into  zower.eecs.umich.edu:/home/gblack/m5/newmem-o3-spec

--HG--
extra : convert_revision : 6b1c8025d29f3e8f90906805dd51a5d523d56004
2007-03-23 21:47:03 -04:00
Kevin Lim
047f77102b Merge ktlim@zizzer:/bk/newmem
into  zamp.eecs.umich.edu:/z/ktlim2/clean/tmp/clean2

src/cpu/base_dyn_inst.hh:
    Hand merge.  Line is no longer needed because it's handled in the ISA.

--HG--
extra : convert_revision : 0be4067aa38759a5631c6940f0167d48fde2b680
2007-03-23 13:20:19 -04:00
Kevin Lim
2330adfa28 Make hardware loads/stores serializing; they need to avoid certain out-of-order interactions in the 21264.
--HG--
extra : convert_revision : d83940af7d0e8efe891d574ac42c6d70d179e2b1
2007-03-23 13:14:05 -04:00
Kevin Lim
941d3168d0 Updates for commit.
1. Move interrupt handling to a separate function to clean up main commit() function a bit.  Also gate the function call off properly based on whether or not there are outstanding interrupts, and the system is not in PAL mode.
2. Better handling of updating instruction's status bits.  Instructions are not marked "atCommit" until other stages view it (pushed off to IEW/IQ), and they have been properly handled (faults).
3. Don't consider the ROB "empty" for the purpose of other stages until the ROB is empty, all stores have written back, and there was no store commits this cycle.  The last is necessary in case a store committed, in which case it would look like all stores have written back but in actuality have not.

src/cpu/o3/commit.hh:
    Slightly modify how interrupts are handled.  Also include some extra bools to keep track of state properly.
src/cpu/o3/commit_impl.hh:
    Slightly modify how interrupts are handled.  Also include some extra bools to keep track of state.

    General correctness updates, most specifically for when commit broadcasts to other stages that the ROB is empty.

--HG--
extra : convert_revision : 682ec6ccf4ee6ed0c8a030ceaba1c90a3619d102
2007-03-23 13:13:10 -04:00
Kevin Lim
78de00091b 3 memory system fixes:
1. Update packet's flags properly when a snoop happens
2. Don't allow accesses to read a block's data if the block has outstanding MSHRs.  This avoids a RAW hazard in MP systems that the memory system was not detecting properly earlier (a write required a block to upgrade, and while the upgrade was outstanding, a read came along and read old data).
3. Update MSHR's request upon a response being handled.  If the MSHR has more targets than it can respond to in one cycle, then its request must be properly updated to the new head of the targets list.

src/mem/bus.cc:
    Update packet's flags properly upon snoop.
src/mem/cache/cache_impl.hh:
    Be sure to not allow accesses to a block with outstanding MSHRs.
src/mem/cache/miss/miss_queue.cc:
    Update MSHR's request upon a response being handled.

--HG--
extra : convert_revision : 76a9abc610ca3f1904f075ad21637148a41982d6
2007-03-23 13:09:37 -04:00
Kevin Lim
e21878c3f2 Handle status bits a little better, as well as non-speculative instructions.
src/cpu/o3/iew_impl.hh:
    Allow for slightly more flexible handling of non-speculative instructions.  They can be other classes now, such as loads or stores.

    Also be sure to clear the state associated with squashes that are not used.  i.e. if a squash due to a memory ordering violation happens on the same cycle as an older branch squashing, clear the state associated with the memory ordering violation.

    Lastly don't consider uncached loads to officially be "at commit" until IEW receives the signal back from commit about the load.
src/cpu/o3/inst_queue_impl.hh:
    Don't consider non-speculative instructions to be "at commit" until the IQ has received a signal from commit about the instruction.  This prevents non-speculative instructions from being issued too early.
src/cpu/o3/mem_dep_unit_impl.hh:
    Clear instruction's ability to issue if it's replayed.

--HG--
extra : convert_revision : d69dae878a30821222885485f4dee87170d56eb3
2007-03-23 11:40:53 -04:00
Kevin Lim
31e78b0b92 Two fixes:
1. Requests are handled more properly now.  They assume the memory system takes control of the request upon sending out an access.
2. load-load ordering is maintained.

src/cpu/base_dyn_inst.hh:
    Update how requests are handled.  The BaseDynInst should not be able to hold a pointer to the request because the request becomes owned by the memory system once it is sent out.

    Also include some functions to allow certain status bits to be cleared.
src/cpu/base_dyn_inst_impl.hh:
    Update how requests are handled.  The BaseDynInst should not be able to hold a pointer to the request because the request becomes owned by the memory system once it is sent out.
src/cpu/o3/fetch_impl.hh:
    General correctness fixes.  retryPkt is not necessarily always set, so handle it properly.  Also consider the cache unblocked only when recvRetry is called.
src/cpu/o3/lsq_unit.hh:
    Handle requests a little more correctly.  Now that the requests aren't pointed to by the DynInst, be sure to delete the request if it's not being used by the memory system.

    Also be sure to not store-load forward from an uncacheable store.
src/cpu/o3/lsq_unit_impl.hh:
    Check to make sure load-load ordering was maintained.

    Also handle requests a little more correctly.

--HG--
extra : convert_revision : e86bead2886d02443cf77bf7a7a1492845e1690f
2007-03-23 11:33:08 -04:00
Kevin Lim
abb07d9da3 Set progress_interval in terms of CPU cycles.
--HG--
extra : convert_revision : 76b0918276cb613eb314ab1479b5ffdb31f31dee
2007-03-23 11:26:30 -04:00
Kevin Lim
55a45d3644 A couple of minor fixes.
1. Set CPU ID in all modes for the O3 CPU.
2. Use nextCycle() function to prevent phase drift in O3 CPU.
3. Remove assertion in rename map that is no longer true.

src/cpu/o3/alpha/cpu_builder.cc:
    Allow for CPU id in all modes, not just full system.  Also include a parameter that was left out by accident.
src/cpu/o3/alpha/cpu_impl.hh:
    Set the CPU ID properly.
src/cpu/o3/cpu.cc:
src/cpu/o3/cpu.hh:
    Use nextCycle() function so that the CPU does not get out of phase when starting up from quiesces.
src/cpu/o3/rename_map.cc:
    Remove assertion that is no longer true.
tests/configs/o3-timing.py:
    Set CPU's id to 0.

--HG--
extra : convert_revision : 2b69c19adfce2adcc2d1939e89d702bd6674d5d5
2007-03-23 11:22:43 -04:00
Ali Saidi
2c47413a7a Merge zizzer:/bk/newmem
into  zeep.pool:/z/saidi/work/m5.newmem

--HG--
extra : convert_revision : fd590d464359d8ae859998a9f446b960781a5e33
2007-03-22 18:39:51 -04:00
Ali Saidi
12b7ebcbca finish up the coding of the Intel Gb NIC... Many Many bugs to squash
src/dev/i8254xGBe.cc:
src/dev/i8254xGBe.hh:
src/dev/i8254xGBe_defs.hh:
    finish coding the Intel Gb NIC device
src/dev/io_device.hh:
    we really don't want to be able to pass a null buffer to dma read, at least not the way we have things setup now... it won't work at all

--HG--
extra : convert_revision : 6739497232317ec407cfa7a96de4575a9a6cfc46
2007-03-22 18:39:41 -04:00
Gabe Black
e3d3ab4513 Add structure based bitfield syntax to the isa_parser. This is primarily useful for x86.
--HG--
extra : convert_revision : dfe6df160d00adec1830d9b88520ba20834d1209
2007-03-22 04:10:57 +00:00
Gabe Black
39182808bc Merge zizzer.eecs.umich.edu:/bk/newmem
into  ahchoo.blinky.homelinux.org:/home/gblack/m5/newmem-x86

--HG--
extra : convert_revision : 79c337f18d635acc176f0ca8d6e71fbc429cb258
2007-03-22 04:10:56 +00:00
Gabe Black
10adec5b53 Merge zizzer.eecs.umich.edu:/bk/newmem
into  zower.eecs.umich.edu:/home/gblack/m5/newmem-statetrace

--HG--
extra : convert_revision : da83ee4c3251fa95faf66e8c6d3412158ffe6887
2007-03-22 00:10:55 -04:00
Gabe Black
39c4ea3473 Fix mcf benchmark object so it gets the arguments it expects.
--HG--
extra : convert_revision : 47087be1f89699e9f8e0dc023abbf593bc0f6618
2007-03-22 00:10:47 -04:00
Gabe Black
276f6d794d Add a junk operand. With no operands, the parser breaks.
--HG--
extra : convert_revision : 7410fd3681ed3d9b1293d982ed5f3553a6c75f3f
2007-03-21 21:09:24 +00:00
Gabe Black
bbffaa8ee0 Start implementing groups of instructions which do the same thing on different sets of inputs.
--HG--
extra : convert_revision : 6a5be61831588f801965dd4e80cb52f28911c320
2007-03-21 21:07:43 +00:00
Gabe Black
1707aa750f put the int register count in intregs.hh
--HG--
extra : convert_revision : c48c13d9c4606c8cb7c60d18cd0f4dac9103a501
2007-03-21 21:04:54 +00:00
Gabe Black
bf8b2e12ea Add a s SPARC_SE gzip regression
--HG--
extra : convert_revision : 5d75c9ad6e4599ce9b1afc176eaafab03bed43f0
2007-03-21 14:31:48 -05:00
Gabe Black
0a80d06dea Break out the one and two byte opcodes into different files. Also change what bits decode is done on to reflect where clumps of instructions are.
--HG--
extra : convert_revision : 8768676eac25e6a4f0dc50ce2dc576bdcdd6e025
2007-03-21 19:19:53 +00:00
Gabe Black
3efec59fc5 Missed a const
--HG--
rename : src/arch/x86/isa/decoder.isa => src/arch/x86/isa/decoder/decoder.isa
extra : convert_revision : a60e7495da6fe99fa2375a3f801f2962c3e41adb
2007-03-21 19:15:40 +00:00
Gabe Black
390b868c87 created SPARC_SE vortex regression.
--HG--
extra : convert_revision : cdb60bcfbea4b79491fb035cc7ecd08cbaa2100b
2007-03-21 01:22:22 -05:00
Gabe Black
63e2d3dcbf Merge zizzer.eecs.umich.edu:/bk/newmem
into  zower.eecs.umich.edu:/home/gblack/m5/newmem-statetrace

--HG--
extra : convert_revision : 41214c71e7fa11d47395975a141793337d020463
2007-03-21 01:18:55 -04:00
Gabe Black
2b4c02b829 The m5 side of statetrace. This is fairly ugly, but I don't want to lose it.
--HG--
extra : convert_revision : 171b41418567c1f41f43363a46fa9aeaa58ae606
2007-03-21 01:18:47 -04:00
Gabe Black
475d36ee93 Ignore "time" and "times" syscalls.
--HG--
extra : convert_revision : 3ff55e35877c0fd74823ce5e52ed16c38da92068
2007-03-20 23:53:52 -04:00
Gabe Black
076fd9a707 Fixed up some types and const placement, and added signed bitfields that sign extend themselves.
--HG--
extra : convert_revision : 84bda8fc14f9a6f7dc7982c9aeb15bf688457706
2007-03-20 11:40:30 +00:00
Gabe Black
e7b015cee1 Added syntax for structure oriented extMachInsts.
--HG--
extra : convert_revision : 4a30c58019ad8e3dd8dffb4c4c08eb6914e5c5be
2007-03-20 06:08:52 +00:00
Gabe Black
4df9411e59 Merge zizzer.eecs.umich.edu:/bk/newmem
into  ahchoo.blinky.homelinux.org:/home/gblack/m5/newmem-x86

--HG--
extra : convert_revision : a6ae6ce183aaa4d1a9434f9ddd64cc30878d2147
2007-03-19 17:13:23 +00:00
Gabe Black
be3550dfe0 Ditched read or write only subbitfields for now since they were proving difficult to implement. Allow component Bitfields to be instantiated without templates, clean up the implementation a little, and adjust the comments to match.
--HG--
extra : convert_revision : b9b8aea285a95eeabbb4c0233c1bce49b8c773b8
2007-03-19 17:12:50 +00:00
Gabe Black
277ded3de7 For the _BitfieldRO and _BitfieldWO classes, make sure the undesired operator is redefined as private.
--HG--
extra : convert_revision : b18d8e49547c2712ab255d048850c5231313b80a
2007-03-19 14:28:19 +00:00
Gabe Black
898652fe04 Formatting fixes.
--HG--
extra : convert_revision : 276d0667daa4626288b56af3b4b17a3f9052e81a
2007-03-19 14:22:28 +00:00
Gabe Black
43dea39dd4 Lots and lots of comments.
--HG--
extra : convert_revision : 0de510464e2e002775ccd79e2922ccb0055845e5
2007-03-19 14:20:27 +00:00
Gabe Black
ad9ab66175 Reworked the BitUnion stuff a bit. There is moderately better isolation of the backend parts, although there are still macros.
--HG--
extra : convert_revision : e9692c5e697c96061ef70cf78ef532c99dbbd672
2007-03-19 13:28:36 +00:00
Gabe Black
7c0825ccf9 Compile fixes for SPARC_FS.
src/arch/alpha/predecoder.hh:
src/arch/sparc/predecoder.hh:
    Put in a missing include
src/cpu/exetrace.cc:
    Convert the legion lockstep stuff from makeExtMI to the predecoder object.

--HG--
extra : convert_revision : 91bad4466f8db1447fff8608fa46a5f236dc3a89
2007-03-18 23:09:51 -04:00
Gabe Black
a1f92af0fb The syntax used for twin stores was confusing the parser so it's now broken down farther.
--HG--
extra : convert_revision : d36bef2d15bc013b3c6199901f57855dfb9dab76
2007-03-17 21:23:03 -04:00
Gabe Black
594c415a0d Created BitUnion type which lets you define nested bitfields for an integer in a portable way.
--HG--
extra : convert_revision : 56a9d06b6b7274a493dae4b290c5f9b42e59f20d
2007-03-17 23:33:00 +00:00
Gabe Black
b54fa0edda Merge zizzer.eecs.umich.edu:/bk/newmem
into  ahchoo.blinky.homelinux.org:/home/gblack/m5/newmem-x86

--HG--
extra : convert_revision : 3246c58850586b880641638bedb1f0025d0ef491
2007-03-16 10:57:52 +00:00
Gabe Black
725ee42ba7 Fix ALPHA_FS compile. The MachInst -> StaticInstPtr constructor is no longer a conversion constructor because it caused ambiguous conversions when setting the pointer to NULL.
--HG--
extra : convert_revision : ce9ecfc03a47642d105f2378208bbe923d6b765b
2007-03-16 10:57:34 +00:00
Gabe Black
3ccaee976a Make the SPARC branch instructions use ExtMachInsts in their constructors. This isn't necessary since they don't use the extended fields, but it's more consistent and more correct.
--HG--
extra : convert_revision : afd4f408122ad5e497012eb9744d6bce66a1de37
2007-03-16 10:55:50 +00:00
Ali Saidi
50475e0e2a Merge zizzer:/bk/newmem
into  pb15.local:/Users/ali/work/m5.newmem

--HG--
extra : convert_revision : 0aa2fadf8978ae30ebe4fde03c525e6b9115117a
2007-03-15 23:21:52 -04:00
Ali Saidi
a18baae026 fix a bug gabe found
--HG--
extra : convert_revision : 12e8f09f6c59a16b2d2eb78bdd1345fc38c4de40
2007-03-15 23:21:01 -04:00
Gabe Black
9ad3f1e479 Refactor things a little.
--HG--
extra : convert_revision : 8167455ffc05130d4afcc68466879c7c439bee57
2007-03-15 19:16:39 +00:00
Gabe Black
f4eee4fb81 File with the predecoder in it.
src/arch/x86/predecoder.cc:
    File for the x86 predecoder process function.

--HG--
extra : convert_revision : f7b53c38ff152cb2677d641074218ffd8434457b
2007-03-15 19:16:38 +00:00
Gabe Black
ae9bed4f8f Split the x86 "process" predecoder method into it's own file.
--HG--
extra : convert_revision : 88185e592df2a7527d36efcce7376fb05f469cbc
2007-03-15 19:16:37 +00:00
Gabe Black
05c71c6194 Merge zizzer.eecs.umich.edu:/bk/newmem
into  ahchoo.blinky.homelinux.org:/home/gblack/m5/newmem-x86

--HG--
extra : convert_revision : cf2cc07f18b877f980e2d1fc83916f7849d9c7d9
2007-03-15 19:16:36 +00:00
Ali Saidi
c6e1dc61c2 Merge zizzer:/bk/newmem
into  zeep.pool:/z/saidi/work/m5.newmem

--HG--
extra : convert_revision : 6a75fa02391c4c65063c5412a568705bb1dd892b
2007-03-15 15:16:35 -04:00