Commit graph

10576 commits

Author SHA1 Message Date
Marco Elver
9649395f85 cpu, o3: Ignored invalidate causing same-address load reordering
In case the memory subsystem sends a combined response with invalidate
(e.g.  ReadRespWithInvalidate), we cannot ignore the invalidate part
of the response.

If we were to ignore the invalidate part, under certain circumstances
this effectively leads to reordering of loads to the same address
which is not permitted under any memory consistency model implemented
in gem5.

Consider the case where a later load's address is computed before an
earlier load in program order, and is therefore sent to the memory
subsystem first. At some point the earlier load's address is computed
and in doing so correctly marks the later load as a
possibleLoadViolation. In the meantime some other node writes and
sends invalidations to all other nodes. The invalidation races with
the later load's ReadResp, and arrives before ReadResp and is
deferred.  Upon receipt of the ReadResp, the response is changed to
ReadRespWithInvalidate, and sent to the CPU. If we ignore the
invalidate part of the packet, we let the later load read the old
value of the address.  Eventually the earlier load's ReadResp arrives,
but with new data. As there was no invalidate snoop (sunk into the
ReadRespWithInvalidate), and if we did not process the invalidate of
the ReadRespWithInvalidate, we obtain a load reordering.

A similar scenario can be constructed where the earlier load's address
is computed after ReadRespWithInvalidate arrives for the younger
load. In this case hitExternalSnoop needs to be set to true on the
ReadRespWithInvalidate, so that upon knowing the address of the
earlier load, checkViolations will cause the later load to be
squashed.

Finally we must account for the case where both loads are sent to the
memory subsystem (reordered), a snoop invalidate arrives and correctly
sets the later loads fault to ReExec. However, before the CPU
processes the fault, the later load's ReadResp arrives and the
writeback discards the outstanding fault. We must add a check to
ensure that we do not skip any unprocessed faults.
2014-12-02 06:08:03 -05:00
Andreas Hansson
74bbe20141 cpu: Always mask the snoop address when performing lock check
Ensure the snoop address check is always using a cache-block aligned
address. This patch updates Alpha and Mips to match the other ISAs.
2014-12-02 06:08:00 -05:00
Stephan Diestelhorst
810349a8a7 cpu: Move packet deallocation to recvTimingResp in the O3 CPU
Move the packet deallocations in the O3 CPU so that the completeDataAccess
deals only with the LSQ specific parts and the generic recvTimingResp frees the
packet in all other cases.
2014-12-02 06:07:58 -05:00
Andreas Hansson
5c84157c29 mem: Relax packet src/dest check and shift onus to crossbar
This patch allows objects to get the src/dest of a packet even if it
is not set to a valid port id. This simplifies (ab)using the bridge as
a buffer and latency adapter in situations where the neighbouring
MemObjects are not crossbars.

The checks that were done in the packet are now shifted to the
crossbar where the fields are used to index into the port
arrays. Thus, the carrier of the information is not burdened with
checking, and the crossbar can check not only that the destination is
set, but also that the port index is within limits.
2014-12-02 06:07:56 -05:00
Andreas Hansson
ea5ccc7041 mem: Clean up packet data allocation
This patch attempts to make the rules for data allocation in the
packet explicit, understandable, and easy to verify. The constructor
that copies a packet is extended with an additional flag "alloc_data"
to enable the call site to explicitly say whether the newly created
packet is short-lived (a zero-time snoop), or has an unknown life-time
and therefore should allocate its own data (or copy a static pointer
in the case of static data).

The tricky case is the static data. In essence this is a
copy-avoidance scheme where the original source of the request (DMA,
CPU etc) does not ask the memory system to return data as part of the
packet, but instead provides a pointer, and then the memory system
carries this pointer around, and copies the appropriate data to the
location itself. Thus any derived packet actually never copies any
data. As the original source does not copy any data from the response
packet when arriving back at the source, we must maintain the copy of
the original pointer to not break the system. We might want to revisit
this one day and pay the price for a few extra memcpy invocations.

All in all this patch should make it easier to grok what is going on
in the memory system and how data is actually copied (or not).
2014-12-02 06:07:54 -05:00
Andreas Hansson
f012166bb6 mem: Cleanup Packet::checkFunctional and hasData usage
This patch cleans up the use of hasData and checkFunctional in the
packet. The hasData function is unfortunately suggesting that it
checks if the packet has a valid data pointer, when it does in fact
only check if the specific packet type is specified to have a data
payload. The confusion led to a bug in checkFunctional. The latter
function is also tidied up to avoid name overloading.
2014-12-02 06:07:52 -05:00
Andreas Hansson
a2ee51f631 mem: Make the requests carried by packets const
This adds a basic level of sanity checking to the packet by ensuring
that a request is not modified once the packet is created. The only
issue that had to be worked around is the relaying of
software-prefetches in the cache. The specific situation is now solved
by first copying the request, and then creating a new packet
accordingly.
2014-12-02 06:07:50 -05:00
Andreas Hansson
fa60d5cf27 mem: Make Request getters const
This patch tidies up the Request class, making all getters const. The
odd one out is incAccessDepth which is called by the memory system as
packets carry the request around. This is also const to enable the
packet to hold on to a const Request.
2014-12-02 06:07:48 -05:00
Andreas Hansson
3d6ec81e66 mem: Add checks and explanation for assertMemInhibit usage 2014-12-02 06:07:46 -05:00
Andreas Hansson
41846cb61b mem: Assume all dynamic packet data is array allocated
This patch simplifies how we deal with dynamically allocated data in
the packet, always assuming that it is array allocated, and hence
should be array deallocated (delete[] as opposed to delete). The only
uses of dataDynamic was in the Ruby testers.

The ARRAY_DATA flag in the packet is removed accordingly. No
defragmentation of the flags is done at this point, leaving a gap in
the bit masks.

As the last part the patch, it renames dataDynamicArray to dataDynamic.
2014-12-02 06:07:43 -05:00
Andreas Hansson
5df96cb690 mem: Remove redundant Packet::allocate calls
This patch cleans up the packet memory allocation confusion. The data
is always allocated at the requesting side, when a packet is created
(or copied), and there is never a need for any device to allocate any
space if it is merely responding to a paket. This behaviour is in line
with how SystemC and TLM works as well, thus increasing
interoperability, and matching established conventions.

The redundant calls to Packet::allocate are removed, and the checks in
the function are tightened up to make sure data is only ever allocated
once. There are still some oddities in the packet copy constructor
where we copy the data pointer if it is static (without ownership),
and allocate new space if the data is dynamic (with ownership). The
latter is being worked on further in a follow-on patch.
2014-12-02 06:07:41 -05:00
Andreas Hansson
0706a25203 mem: Use const pointers for port proxy write functions
This patch changes the various write functions in the port proxies
to use const pointers for all sources (similar to how memcpy works).

The one unfortunate aspect is the need for a const_cast in the packet,
to avoid having to juggle a const and a non-const data pointer. This
design decision can always be re-evaluated at a later stage.
2014-12-02 06:07:38 -05:00
Andreas Hansson
9779ba2e37 mem: Add const getters for write packet data
This patch takes a first step in tightening up how we use the data
pointer in write packets. A const getter is added for the pointer
itself (getConstPtr), and a number of member functions are also made
const accordingly. In a range of places throughout the memory system
the new member is used.

The patch also removes the unused isReadWrite function.
2014-12-02 06:07:36 -05:00
Andreas Hansson
25bfc24999 mem: Remove null-check bypassing in Packet::getPtr
This patch removes the parameter that enables bypassing the null check
in the Packet::getPtr method. A number of call sites assume the value
to be non-null.

The one odd case is the RubyTester, which issues zero-sized
prefetches(!), and despite being reads they had no valid data
pointer. This is now fixed, but the size oddity remains (unless anyone
object or has any good suggestions).

Finally, in the Ruby Sequencer, appropriate checks are made for flush
packets as they have no valid data pointer.
2014-12-02 06:07:34 -05:00
Omar Naji
0e63d2cd62 mem: Add a GDDR5 DRAM config
This patch adds a first cut GDDR5 config to accommodate the users
combining gem5 and GPUSim. The config is based on a SK Hynix
datasheet, and the Nvidia GTX580 specification. Someone from the
GPUSim user-camp should tweak the default page-policy and static
frontend and backend latencies.
2014-12-02 06:07:32 -05:00
Andreas Hansson
b0aa5a326d stats: Bump stats after static analysis fixes
Fixing up the uninitialised values changes two of the x86 Linux boot
regressions slightly.
2014-11-24 09:03:39 -05:00
Andreas Hansson
d66b14ca61 misc: Another round of static analysis fixups
Mostly addressing uninitialised members.
2014-11-24 09:03:38 -05:00
Alexandru Dutu
1f539f13c3 mem: Page Table map api modification
This patch adds uncacheable/cacheable and read-only/read-write attributes to
the map method of PageTableBase. It also modifies the constructor of TlbEntry
structs for all architectures to consider the new attributes.
2014-11-23 18:01:09 -08:00
Alexandru Dutu
c11bcb8119 mem: Multi Level Page Table bug fix
The multi level page table was giving false positives for already mapped
translations. This patch fixes the bogus behavior.
2014-11-23 18:01:09 -08:00
Alexandru Dutu
e4859fae5b mem: Page Table long lines
Trimmed down all the lines greater than 78 characters.
2014-11-23 18:01:09 -08:00
Alexandru Dutu
a19cf6943b config, kvm: Enabling KvmCPU in SE mode
This patch modifies se.py such that it can now use kvm cpu model.
2014-11-23 18:01:08 -08:00
Alexandru Dutu
f743bdcb69 x86: Segment initialization to support KvmCPU in SE
This patch sets up low and high privilege code and data segments and places them
in the following order: cs low, ds low, ds, cs, in the GDT. Additionally, a
syscall and page fault handler for KvmCPU in SE mode are defined. The order of
the segment selectors in GDT is required in this manner for interrupt handling
to work properly. Segment initialization is done for all the thread
contexts.
2014-11-23 18:01:08 -08:00
Alexandru Dutu
adbaa4dfde kvm, x86: Adding support for SE mode execution
This patch adds methods in KvmCPU model to handle KVM exits caused by syscall
instructions and page faults. These types of exits will be encountered if
KvmCPU is run in SE mode.
2014-11-23 18:01:08 -08:00
Alexandru Dutu
335514dfdc cpuid, x86: Enabling more features in CPUid
Adding more features in the CPUid with the purpose of supporting running the
KvmCPU in SE mode.
2014-11-23 18:01:08 -08:00
Steve Reinhardt
252a463b6b Backed out prior changeset f9fb64a72259
Back out use of importlib to avoid implicitly creating
dependency on Python 2.7.
2014-11-23 18:00:47 -08:00
Gabe Black
12243a3835 config: ruby: Get rid of an "eval" and an "exec" operating on generated code.
We can get the same result using importlib.
2014-11-23 05:55:26 -08:00
Gabe Black
2d2a5aa410 x86: Update stats for the new Linux delay port. 2014-11-21 17:22:19 -08:00
Gabe Black
8bbfb1b39d x86: pc: Put a stub IO device at port 0xed which the kernel can use for delays.
There was already a stub device at 0x80, the port traditionally used for an IO
delay. 0x80 is also the port used for POST codes sent by firmware, and that
may have prompted adding this port as a second option.
2014-11-21 17:22:02 -08:00
Nilay Vaish
708e80d9bb configs: small fix to ruby portion of fs.py and se.py
In fs.py the io port controller was being attached to the iobus multiple
times.  This should be done only once.  In se.py, the the option use_map
was being set which no longer exists.
2014-11-18 19:17:29 -06:00
Gabe Black
b5fd6050a2 dev: Use fixed size member variables to describe fixed size PL111 registers. 2014-11-18 02:38:23 -08:00
Gabe Black
a08cfd797b vnc: Add a conversion function for bgr888. 2014-11-17 01:45:42 -08:00
Gabe Black
aceeecb192 x86: Fix setting segment bases in real mode.
The data size used for actually writing the base value for the segment was the
default size, but really it should set the entire value without any possible
truncation.
2014-11-17 01:00:53 -08:00
Gabe Black
f8603fa120 x86: Fix some bugs in the real mode far jmp instruction.
The far pointer should be shifted right to get the selector value, not left.
Also, when calculating the width of the offset, the wrong register was used in
one spot.
2014-11-17 00:20:01 -08:00
Gabe Black
7739c24fbe x86: APIC: Only set deliveryStatus if our IPI is going somewhere.
Otherwise the IPI which isn't sent will never arrive, and the deliveryStatus
bit will never be cleared.
2014-11-17 00:19:07 -08:00
Gabe Black
79e7ca307e x86: APIC: Fix the getRegArrayBit function.
The getRegArrayBit function extracts a bit from a series of registers which
are treated as a single large bit array. A previous change had modified the
logic which figured out which bit to extract from ">> 5" to "% 5" which seems
wrong, especially when other, similar functions were changed to use "% 32".
2014-11-17 00:17:06 -08:00
Gabe Black
994c44035d x86: Update the stats for the x86 FS o3 boot test. 2014-11-17 00:16:36 -08:00
Gabe Black
d228db1143 x86: Fix the CPUID Long Mode Address Size function.
The value in EAX has an 8 bit field for the linear address size and one for
the physical address size when calling that function. A recent change
implemented it but returned 0xff for both of those fields. That implies that
linear and physical addresses are 255 bits wide which is wrong. When using the
KVM CPU model this causes an error, presumably because some of those bits are
actually reserved, or the CPU or kernel realizes 255 bits is a bad value.

This change makes those values 48.
2014-11-16 23:12:42 -08:00
Andrew Bardsley
27b7b9e561 config: Fix checkpoint restore in C++ config example
This patch fixes the checkpoint restore option in the example of C++
configuration (util/cxx_config).

The fix introduces a call to config_manager->startup() (which calls startup
on all SimObjects managed by that manager) to replicate the loop of
SimObject::startup calls in src/python/m5/simulate.py::simulate guarded by
need_startup.  As util/cxx_config/main.cc is a C++ analogue of
src/python/mt/simulate.py, it should make a similar set of calls.
2014-11-14 03:54:02 -05:00
Andreas Hansson
481eb6ae80 arm: Fixes based on UBSan and static analysis
Another churn to clean up undefined behaviour, mostly ARM, but some
parts also touching the generic part of the code base.

Most of the fixes are simply ensuring that proper intialisation. One
of the more subtle changes is the return type of the sign-extension,
which is changed to uint64_t. This is to avoid shifting negative
values (undefined behaviour) in the ISA code.
2014-11-14 03:53:51 -05:00
Andreas Hansson
9ffe0e7ba6 mem: Clarify unit of DRAM controller buffer size 2014-11-14 03:53:48 -05:00
Andreas Hansson
4583a5114a stats: Bump regressions to match latest changes
Updates after timezone hick-up and sorting of dictionary items in the
SimObject.
2014-11-12 09:05:25 -05:00
Mitch Hayenga
9d6d8e02aa mem: Delete unused variable in Garnet NetworkLink
With recent changes OSX clang compilation fails due to an unused variable.
2014-11-12 09:05:23 -05:00
Ali Saidi
b6f32253dd arm: Fix timing wakeup with LLSC 2014-11-12 09:05:22 -05:00
Andreas Hansson
7d05895120 sim: Sort SimObject descendants and ports
This patch fixes a number of occurences where the sorting order of the
objects was implementation defined.
2014-11-12 09:05:21 -05:00
Andreas Hansson
cc336ecb5e base: Revert 9277177eccff and use getenv/setenv for UTC time
This patch reverts changeset 9277177eccff which does not do what it
was intended to do. In essence, we go back to implementing mkutctime
much like the non-standard timegm extension.
2014-11-12 09:05:20 -05:00
Nilay Vaish
02b4605da0 stats: changes to x86 o3 fs and sparc fs regression tests. 2014-11-11 14:17:10 -06:00
Marc Orr
bf80734b2c x86 isa: This patch attempts an implementation at mwait.
Mwait works as follows:
1. A cpu monitors an address of interest (monitor instruction)
2. A cpu calls mwait - this loads the cache line into that cpu's cache.
3. The cpu goes to sleep.
4. When another processor requests write permission for the line, it is
   evicted from the sleeping cpu's cache. This eviction is forwarded to the
   sleeping cpu, which then wakes up.

Committed by: Nilay Vaish <nilay@cs.wisc.edu>
2014-11-06 05:42:22 -06:00
Marc Orr
3947f88d0f tests: A test program for the new mwait implementation.
This is a simple test program for the new mwait implemenation. It is uses
m5threads to create to threads of execution in syscall emulation mode that
interact using the mwait instruction.

Committed by: Nilay Vaish <nilay@cs.wisc.edu>
2014-11-06 05:42:21 -06:00
Andrew Lukefahr
bd32d55a2c cpu: Minor Draining Bug
Fixes a bug where Minor drains in the midst of committing a
conditional store.

While committing a conditional store, lastCommitWasEndOfMacroop is true
(from the previous instruction) as we still haven't finished the conditional
store. If a drain occurs before the cache response, Minor would check just
lastCommitWasEndOfMacroop, which was true, and set drainState=DrainHaltFetch,
which increases the streamSeqNum.  This caused the conditional store to be
squashed when the memory responded and it completed.  However, to the memory
the store succeeded, while to the instruction sequence it never occurred.

In the case of an LLSC, the instruction sequence will replay the squashed
STREX, which will fail as the cache is no longer in LLSC.  Then the
instruction sequence will loop back to a LDREX, which receives the updated
(incorrect) value.

Committed by: Nilay Vaish <nilay@cs.wisc.edu>
2014-11-06 05:42:21 -06:00
Nilay Vaish
a75e27b4a6 stats: updates due to changes to ruby 2014-11-06 05:42:21 -06:00