This change is a low level and pervasive reorganization of how PCs are managed
in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about,
the PC and the NPC, and the lsb of the PC signaled whether or not you were in
PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next
micropc, x86 and ARM introduced variable length instruction sets, and ARM
started to keep track of mode bits in the PC. Each CPU model handled PCs in
its own custom way that needed to be updated individually to handle the new
dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack,
the complexity could be hidden in the ISA at the ISA implementation's expense.
Areas like the branch predictor hadn't been updated to handle branch delay
slots or micropcs, and it turns out that had introduced a significant (10s of
percent) performance bug in SPARC and to a lesser extend MIPS. Rather than
perpetuate the problem by reworking O3 again to handle the PC features needed
by x86, this change was introduced to rework PC handling in a more modular,
transparent, and hopefully efficient way.
PC type:
Rather than having the superset of all possible elements of PC state declared
in each of the CPU models, each ISA defines its own PCState type which has
exactly the elements it needs. A cross product of canned PCState classes are
defined in the new "generic" ISA directory for ISAs with/without delay slots
and microcode. These are either typedef-ed or subclassed by each ISA. To read
or write this structure through a *Context, you use the new pcState() accessor
which reads or writes depending on whether it has an argument. If you just
want the address of the current or next instruction or the current micro PC,
you can get those through read-only accessors on either the PCState type or
the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the
move away from readPC. That name is ambiguous since it's not clear whether or
not it should be the actual address to fetch from, or if it should have extra
bits in it like the PAL mode bit. Each class is free to define its own
functions to get at whatever values it needs however it needs to to be used in
ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the
PC and into a separate field like ARM.
These types can be reset to a particular pc (where npc = pc +
sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as
appropriate), printed, serialized, and compared. There is a branching()
function which encapsulates code in the CPU models that checked if an
instruction branched or not. Exactly what that means in the context of branch
delay slots which can skip an instruction when not taken is ambiguous, and
ideally this function and its uses can be eliminated. PCStates also generally
know how to advance themselves in various ways depending on if they point at
an instruction, a microop, or the last microop of a macroop. More on that
later.
Ideally, accessing all the PCs at once when setting them will improve
performance of M5 even though more data needs to be moved around. This is
because often all the PCs need to be manipulated together, and by getting them
all at once you avoid multiple function calls. Also, the PCs of a particular
thread will have spatial locality in the cache. Previously they were grouped
by element in arrays which spread out accesses.
Advancing the PC:
The PCs were previously managed entirely by the CPU which had to know about PC
semantics, try to figure out which dimension to increment the PC in, what to
set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction
with the PC type itself. Because most of the information about how to
increment the PC (mainly what type of instruction it refers to) is contained
in the instruction object, a new advancePC virtual function was added to the
StaticInst class. Subclasses provide an implementation that moves around the
right element of the PC with a minimal amount of decision making. In ISAs like
Alpha, the instructions always simply assign NPC to PC without having to worry
about micropcs, nnpcs, etc. The added cost of a virtual function call should
be outweighed by not having to figure out as much about what to do with the
PCs and mucking around with the extra elements.
One drawback of making the StaticInsts advance the PC is that you have to
actually have one to advance the PC. This would, superficially, seem to
require decoding an instruction before fetch could advance. This is, as far as
I can tell, realistic. fetch would advance through memory addresses, not PCs,
perhaps predicting new memory addresses using existing ones. More
sophisticated decisions about control flow would be made later on, after the
instruction was decoded, and handed back to fetch. If branching needs to
happen, some amount of decoding needs to happen to see that it's a branch,
what the target is, etc. This could get a little more complicated if that gets
done by the predecoder, but I'm choosing to ignore that for now.
Variable length instructions:
To handle variable length instructions in x86 and ARM, the predecoder now
takes in the current PC by reference to the getExtMachInst function. It can
modify the PC however it needs to (by setting NPC to be the PC + instruction
length, for instance). This could be improved since the CPU doesn't know if
the PC was modified and always has to write it back.
ISA parser:
To support the new API, all PC related operand types were removed from the
parser and replaced with a PCState type. There are two warts on this
implementation. First, as with all the other operand types, the PCState still
has to have a valid operand type even though it doesn't use it. Second, using
syntax like PCS.npc(target) doesn't work for two reasons, this looks like the
syntax for operand type overriding, and the parser can't figure out if you're
reading or writing. Instructions that use the PCS operand (which I've
consistently called it) need to first read it into a local variable,
manipulate it, and then write it back out.
Return address stack:
The return address stack needed a little extra help because, in the presence
of branch delay slots, it has to merge together elements of the return PC and
the call PC. To handle that, a buildRetPC utility function was added. There
are basically only two versions in all the ISAs, but it didn't seem short
enough to put into the generic ISA directory. Also, the branch predictor code
in O3 and InOrder were adjusted so that they always store the PC of the actual
call instruction in the RAS, not the next PC. If the call instruction is a
microop, the next PC refers to the next microop in the same macroop which is
probably not desirable. The buildRetPC function advances the PC intelligently
to the next macroop (in an ISA specific way) so that that case works.
Change in stats:
There were no change in stats except in MIPS and SPARC in the O3 model. MIPS
runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could
likely be improved further by setting call/return instruction flags and taking
advantage of the RAS.
TODO:
Add != operators to the PCState classes, defined trivially to be !(a==b).
Smooth out places where PCs are split apart, passed around, and put back
together later. I think this might happen in SPARC's fault code. Add ISA
specific constructors that allow setting PC elements without calling a bunch
of accessors. Try to eliminate the need for the branching() function. Factor
out Alpha's PAL mode pc bit into a separate flag field, and eliminate places
where it's blindly masked out or tested in the PC.
In the process make add skipFuction() to handle isa specific function skipping
instead of ifdefs and other ugliness. For almost all ABIs, 64 bit arguments can
only start in even registers. Size is now passed to getArgument() so that 32
bit systems can make decisions about register selection for 64 bit arguments.
The number argument is now passed by reference because getArgument() will need
to change it based on the size of the argument and the current argument number.
For ARM, if the argument number is odd and a 64-bit register is requested the
number must first be incremented to because all 64 bit arguments are passed
in an even argument register. Then the number will be incremented again to
access both halves of the argument.
Also move the "Fault" reference counted pointer type into a separate file,
sim/fault.hh. It would be better to name this less similarly to sim/faults.hh
to reduce confusion, but fault.hh matches the name of the type. We could change
Fault to FaultPtr to match other pointer types, and then changing the name of
the file would make more sense.
When decoding a srs instruction, invalid mode encoding returns invalid instruction.
This can happen when garbage instructions are fetched from mispredicted path
This allows one two different OS requirements for the same ISA to be handled.
Some OSes are compiled for a virtual address and need to be loaded into physical
memory that starts at address 0, while other bare metal tools generate
images that start at address 0.
Replace direct call to unserialize() on each SimObject with a pair of
calls for better control over initialization in both ckpt and non-ckpt
cases.
If restoring from a checkpoint, loadState(ckpt) is called on each
SimObject. The default implementation simply calls unserialize() if
there is a corresponding checkpoint section, so we get backward
compatibility for existing objects. However, objects can override
loadState() to get other behaviors, e.g., doing other programmed
initializations after unserialize(), or complaining if no checkpoint
section is found. (Note that the default warning for a missing
checkpoint section is now gone.)
If not restoring from a checkpoint, we call the new initState() method
on each SimObject instead. This provides a hook for state
initializations that are only required when *not* restoring from a
checkpoint.
Given this new framework, do some cleanup of LiveProcess subclasses
and X86System, which were (in some cases) emulating initState()
behavior in startup via a local flag or (in other cases) erroneously
doing initializations in startup() that clobbered state loaded earlier
by unserialize().
When each load or store is sent to the LSQ, we check whether it will cross a
cache line boundary and, if so, split it in two. This creates two TLB
translations and two memory requests. Care has to be taken if the first
packet of a split load is sent but the second blocks the cache. Similarly,
for a store, if the first packet cannot be sent, we must store the second
one somewhere to retry later.
This modifies the LSQSenderState class to record both packets in a split
load or store.
Finally, a new const variable, HasUnalignedMemAcc, is added to each ISA
to indicate whether unaligned memory accesses are allowed. This is used
throughout the changed code so that compiler can optimise away code dealing
with split requests for ISAs that don't need them.
When accessing arguments for a syscall, the position of an argument depends on
the policies of the ISA, how much space preceding arguments took up, and the
"alignment" of the index for this particular argument into the number of
possible storate locations. This change adjusts getSyscallArg to take its
index parameter by reference instead of value and to adjust it to point to the
possible location of the next argument on the stack, basically just after the
current one. This way, the rules for the new argument can be applied locally
without knowing about other arguments since those have already been taken into
account implicitly.
All system calls have also been changed to reflect the new interface. In a
number of cases this made the implementation clearer since it encourages
arguments to be collected in one place in order and then used as necessary
later, as opposed to scattering them throughout the function or using them in
place in long expressions. It also discourages using getSyscallArg over and
over to retrieve the same value when a temporary would do the job.
Using a look up table changed the run time of the SPARC_FS solaris boot
regression from:
real 14m45.951s
user 13m57.528s
sys 0m3.452s
to:
real 12m19.777s
user 12m2.685s
sys 0m2.420s
This patch adds limited multithreading support in syscall-emulation
mode, by using the clone system call. The clone system call works
for Alpha, SPARC and x86, and multithreaded applications run
correctly in Alpha and SPARC.
Basically merge it in with Halted.
Also had to get rid of a few other functions that
called ThreadContext::deallocate(), including:
- InOrderCPU's setThreadRescheduleCondition.
- ThreadContext::exit(). This function was there to avoid terminating
simulation when one thread out of a multi-thread workload exits, but we
need to find a better (non-cpu-centric) way.
Bogus calls to ChunkGenerator with negative size were triggering
a new assertion that was added there.
Also did a little renaming and cleanup in the process.
the primary identifier for a hardware context should be contextId(). The
concept of threads within a CPU remains, in the form of threadId() because
sometimes you need to know which context within a cpu to manipulate.
SE. Process still keeps track of the tc's it owns, but registration occurs
with the System, this eases the way for system-wide context Ids based on
registration.
across the subclasses. generally make it so that member data is _cpuId and
accessor functions are cpuId(). The ID val comes from the python (default -1 if
none provided), and if it is -1, the index of cpuList will be given. this has
passed util/regress quick and se.py -n4 and fs.py -n4 as well as standard
switch.
Make them easier to express by only having the cxx_type parameter which
has the full namespace name, and drop the cxx_namespace thing.
Add support for multiple levels of namespace.
Even though we're not incorrect about operator precedence, let's add
some parens in some particularly confusing places to placate GCC 4.3
so that we don't have to turn the warning off. Agreed that this is a
bit of a pain for those users who get the order of operations correct,
but it is likely to prevent bugs in certain cases.
We should always refer to the specific ISA in that arch directory.
This is especially necessary if we're ever going to make it to the
point where we actually have heterogeneous systems.
This appears to work, but I don't want to commit it until it gets tested a lot more.
I haven't deleted the functionality in this patch that will come later, but one question
is how to enforce encourage objects that call getVirtPort() to not cache the virtual port
since if the CPU changes out from under them it will be worse than useless. Perhaps a null
function like delVirtPort() is still useful in that case.
Also some bug fixes in MIPS ISA uncovered by g++ warnings
(Python string compares don't work in C++!).
--HG--
extra : convert_revision : b347cc0108f23890e9b73b3ee96059f0cea96cf6
SimObjects not yet updated:
- Process and subclasses
- BaseCPU and subclasses
The SimObject(const std::string &name) constructor was removed. Subclasses
that still rely on that behavior must call the parent initializer as
: SimObject(makeParams(name))
--HG--
extra : convert_revision : d6faddde76e7c3361ebdbd0a7b372a40941c12ed
The page table now stores actual page table entries. It is still a templated
class here, but this will be corrected in the near future.
--HG--
extra : convert_revision : 804dcc6320414c2b3ab76a74a15295bd24e1d13d
In O3, a nop is used to carry faults down the pipeline that didn't originate
from an instruction. If the instruction doesn't do anything, that is just
returns NoFault, but doesn't have IsNop set, the NoFault will overwrite the
fault that's being sent down and nothing will happen.
--HG--
extra : convert_revision : 54d99002b550ca0e1cf14603f588dc1038e3e535
Each "strand" may need to have a private copy of this state, but I couldn't
find anywhere in the spec that said that after looking briefly.
This prevents writes to the thread context in o3 which was causing the
pipeline to be flushed and stopping any forward progress. The other ASI
accessible state will probably need to be accessed differently if/when we get
O3 full system up and running.
--HG--
extra : convert_revision : fa7fba812d7f76564ef4a23818e60f536710d557
Move the argument files to src/sim and add a utility.cc file with a function
getArguments() that returns the given argument in the architecture specific fashion.
getArguments() was getArg() is the architecture specific Argument class and has had
all magic numbers replaced with meaningful constants. Also add a function to the
Argument class for testing if an argument is NULL.
--HG--
rename : src/arch/alpha/arguments.cc => src/sim/arguments.cc
rename : src/arch/alpha/arguments.hh => src/sim/arguments.hh
extra : convert_revision : 8b93667bafaa03b52aadb64d669adfe835266b8e
The initial stack frame for x86 is now substantially more correct. The fixes made here can be back ported to SPARC and possible the other ISAs as well. The auxiliary vector types were moved to the LiveProcess base class because they are independent of ISA. Some of the types may only apply to Linux, though, so they may have to be moved.
--HG--
extra : convert_revision : 89ace35fcc8eb9586d2fee8eeccbc3686499ef24
Code was assuming that all argument registers followed in order from ArgumentReg0. There is now an ArgumentReg array which is indexed to find the right index. There is a constant, NumArgumentRegs, which can be used to protect against using an invalid ArgumentReg.
--HG--
extra : convert_revision : f448a3ca4d6adc3fc3323562870f70eec05a8a1f
creation and initialization now happens in python. Parameter objects
are generated and initialized by python. The .ini file is now solely for
debugging purposes and is not used in construction of the objects in any
way.
--HG--
extra : convert_revision : 7e722873e417cb3d696f2e34c35ff488b7bff4ed
into ahchoo.blinky.homelinux.org:/home/gblack/m5/newmem-o3-micro
src/cpu/o3/fetch_impl.hh:
hand merge
--HG--
extra : convert_revision : 3f71f3ac2035eec8b6f7bceb6906edb4dd09c045
into doughnut.hpl.hp.com:/home/gblack/newmem-o3-micro
src/cpu/base_dyn_inst_impl.hh:
src/cpu/o3/fetch_impl.hh:
Hand merge
--HG--
extra : convert_revision : 0c0692033ac30133672d8dfe1f1a27e9d9e95a3d
src/arch/x86/predecoder.cc:
Seperate the pc-pc and the pc of the incoming bytes, and get rid of the "moreBytes" which just takes a MachInst. Also make the "opSize" field describe the number of bytes and not the log of the number of bytes.
--HG--
extra : convert_revision : 3a5ec7053ec69c5cba738a475d8b7fd9e6e6ccc0
src/arch/x86/isa/macroop.isa:
Make microOp vs microop and macroOp vs macroop capitilization consistent. Also fill out the emulation environment handling a little more, and use an object to pass around output code.
src/arch/x86/isa/microops/base.isa:
Make microOp vs microop and macroOp vs macroop capitilization consistent. Also adjust python to C++ bool translation.
--HG--
extra : convert_revision : 6f4bacfa334c42732c845f9a7f211cbefc73f96f
Initialize a temporary variable for thread->readPC() at setupFetchRequest() to reduce function calls.
exec tracing isn't needed for m5.fast binaries
Moved MISCREG_GL, MISCREG_CWP, and MISCREG_TLB_DATA out of switch statement and use if blocks instead.
src/arch/sparc/miscregfile.cc:
Moved MISCREG_GL, MISCREG_CWP, and MISCREG_TLB_DATA out of switch statement and use if blocks instead.
src/cpu/simple/base.cc:
Assign traceData to be NULL at BaseSimpleCPU constructor.
Initialize a temporary variable for thread->readPC() at setupFetchRequest() to reduce function calls.
exec tracing isn't needed for m5.fast binaries
--HG--
extra : convert_revision : 5dc92fff05c9bde994f1e0f1bb40e11c44eb72c6
(which defines fenv) doesn't necessarily extend to c++ and it is a problem with solaris. If really
desired this could wrap the ieeefp interface found in bsd* as well, but I see no need at the moment.
src/arch/alpha/isa/fp.isa:
src/arch/sparc/isa/formats/basic.isa:
use m5_fesetround()/m5_fegetround() istead of fenv interface directly
src/arch/sparc/isa/includes.isa:
use base/fenv instead of fenv directly
src/base/SConscript:
add fenv to sconscript
src/base/fenv.hh:
src/base/random.cc:
m5 implementation to standerdize fenv across platforms.
--HG--
extra : convert_revision : 38d2629affd964dcd1a5ab0db4ac3cb21438e72c
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
into ahchoo.blinky.homelinux.org:/home/gblack/m5/newmem-x86
src/arch/mips/utility.hh:
src/arch/x86/SConscript:
Hand merge
--HG--
extra : convert_revision : 0ba457aab52bf6ffc9191fd1fe1006ca7704b5b0
Removed the getOpcode function from StaticInst which only made sense for Alpha.
Started implementing the x86 predecoder.
--HG--
extra : convert_revision : a13ea257c8943ef25e9bc573024a99abacf4a70d
src/arch/sparc/ua2005.cc:
fix interrupting when quisced. Since sticks correspond to instructions when not quisced we need to
check if were suspended and interrupt at the guess time
src/base/traceflags.py:
add trace flag for Iob
src/cpu/simple/base.cc:
Use Quisce instead of IPI trace flag
src/dev/sparc/iob.cc:
add some Dprintfs
--HG--
extra : convert_revision : 72e18fcc750ad1e4b2bb67b19b354eaffc6af6d5
src/arch/sparc/miscregfile.cc:
this code should be in readFSreg
src/arch/sparc/ua2005.cc:
move code froh miscregfile to ua2005.cc
--HG--
extra : convert_revision : fa450b04ad73ab6f6e25d66fa0368054263f09f9
automatic. The point is that now a subdirectory can be added
to the build process just by creating a SConscript file in it.
The process has two passes. On the first pass, all subdirs
of the root of the tree are searched for SConsopts files.
These files contain any command line options that ought to be
added for a particular subdirectory. On the second pass,
all subdirs of the src directory are searched for SConscript
files. These files describe how to build any given subdirectory.
I have added a Source() function. Any file (relative to the
directory in which the SConscript resides) passed to that
function is added to the build. Clean up everything to take
advantage of Source().
function is added to the list of files to be built.
--HG--
extra : convert_revision : 103f6b490d2eb224436688c89cdc015211c4fd30
src/arch/alpha/utility.hh:
src/arch/mips/utility.hh:
src/arch/sparc/utility.hh:
src/arch/x86/utility.hh:
add hook for system to startup the cpu or not... in the case of FS sparc, only the first cpu would get spunup.. the rest sit in an idle state until they get an ipi
src/arch/sparc/isa/decoder.isa:
handle writable bits of strandstatus register in miscregfile
src/arch/sparc/miscregfile.hh:
some constants for the strand status register
src/arch/sparc/ua2005.cc:
properly implement the strand status register
src/dev/sparc/iob.cc:
implement ipi generation properly
src/sim/system.cc:
call into the ISA to start the CPU (or not)
--HG--
extra : convert_revision : 0003b2032337d8a031a9fc044da726dbb2a9e36f
src/arch/sparc/asi.cc:
src/arch/sparc/asi.hh:
add CMT ASI registers
src/arch/sparc/tlb.cc:
Panic if any of the CMT registers are being accessed
--HG--
extra : convert_revision : b9a94281e2074a576ac21d042b756950d509e758
configs/common/FSConfig.py:
Use binaries we've compiled instead of the ones that come with Legion
src/arch/alpha/interrupts.hh:
get rid of post(int int_type) and add a get_vec function that gets the interrupt vector for an interrupt number
src/arch/sparc/asi.cc:
Add AsiIsInterrupt() to AsiIsMmu()
src/arch/sparc/faults.cc:
src/arch/sparc/faults.hh:
Add InterruptVector type
src/arch/sparc/interrupts.hh:
rework interrupts. They are no longer cleared when created... A I/O or ASI read/write needs to happen before they are cleared
src/arch/sparc/isa_traits.hh:
Add the "interrupt" trap types to isa traits
src/arch/sparc/miscregfile.cc:
add names for all the misc registers and possible post an interrupt when TL is changed.
src/arch/sparc/miscregfile.hh:
Add a helper function to post an interrupt when pil < some set softint
src/arch/sparc/regfile.cc:
src/arch/sparc/regfile.hh:
InterruptLevel shouldn't really live here, moved to interrupt.hh
src/arch/sparc/tlb.cc:
Add interrupt ASIs to TLB
src/arch/sparc/ua2005.cc:
Add checkSoftInt to check if a softint needs to be posted
Check that a tickCompare isn't scheduled before scheduling one
Post and clear interrupts on queue writes and what not
src/base/bitfield.hh:
Add an helper function to return the msb that is set
src/cpu/base.cc:
src/cpu/base.hh:
get rid of post_interrupt(type) since it's no longer needed.. Add a way to see what interrupts are pending
src/cpu/intr_control.cc:
src/cpu/intr_control.hh:
src/dev/alpha/tsunami_cchip.cc:
src/python/m5/objects/IntrControl.py:
Make IntrControl have a system pointer rather than using a cpu pointer to get one
src/dev/sparc/SConscript:
add iob to SConsscrip
tests/quick/10.linux-boot/ref/alpha/linux/tsunami-simple-atomic-dual/config.ini:
tests/quick/10.linux-boot/ref/alpha/linux/tsunami-simple-atomic-dual/config.out:
tests/quick/10.linux-boot/ref/alpha/linux/tsunami-simple-atomic/config.ini:
tests/quick/10.linux-boot/ref/alpha/linux/tsunami-simple-atomic/config.out:
tests/quick/10.linux-boot/ref/alpha/linux/tsunami-simple-timing-dual/config.ini:
tests/quick/10.linux-boot/ref/alpha/linux/tsunami-simple-timing-dual/config.out:
tests/quick/10.linux-boot/ref/alpha/linux/tsunami-simple-timing/config.ini:
tests/quick/10.linux-boot/ref/alpha/linux/tsunami-simple-timing/config.out:
tests/quick/80.netperf-stream/ref/alpha/linux/twosys-tsunami-simple-atomic/config.ini:
tests/quick/80.netperf-stream/ref/alpha/linux/twosys-tsunami-simple-atomic/config.out:
update config.ini/out for intrcntrl not having a cpu pointer anymore
--HG--
extra : convert_revision : 38614f6b9ffc8f3c93949a94ff04b7d2987168dd
