Commit graph

9648 commits

Author SHA1 Message Date
Andreas Hansson
b3fc8839c4 mem: Make packet bus-related time accounting relative
This patch changes the bus-related time accounting done in the packet
to be relative. Besides making it easier to align the cache timing to
cache clock cycles, it also makes it possible to create a Last-Level
Cache (LLC) directly to a memory controller without a bus inbetween.

The bus is unique in that it does not ever make the packets wait to
reflect the time spent forwarding them. Instead, the cache is
currently responsible for making the packets wait. Thus, the bus
annotates the packets with the time needed for the first word to
appear, and also the last word. The cache then delays the packets in
its queues before passing them on. It is worth noting that every
object attached to a bus (devices, memories, bridges, etc) should be
doing this if we opt for keeping this way of accounting for the bus
timing.
2013-02-19 05:56:06 -05:00
Andreas Hansson
362160c8ae mem: Add deferred packet class to prefetcher
This patch removes the time field from the packet as it was only used
by the preftecher. Similar to the packet queue, the prefetcher now
wraps the packet in a deferred packet, which also has a tick
representing the absolute time when the packet should be sent.
2013-02-19 05:56:06 -05:00
Andreas Hansson
7cd49b24d2 sim: Make clock private and access using clockPeriod()
This patch makes the clock member private to the ClockedObject and
forces all children to access it using clockPeriod(). This makes it
impossible to inadvertently change the clock, and also makes it easier
to transition to a situation where the clock is derived from e.g. a
clock domain, or through a multiplier.
2013-02-19 05:56:06 -05:00
Andreas Hansson
5c7ebee434 x86: Move APIC clock divider to Python
This patch moves the 16x APIC clock divider to the Python code to
avoid the post-instantiation modifications to the clock. The x86 APIC
was the only object setting the clock after creation time and this
required some custom functionality and configuration. With this patch,
the clock multiplier is moved to the Python code and the objects are
instantiated with the appropriate clock.
2013-02-19 05:56:06 -05:00
Sascha Bischoff
86a4d09269 mem: Fix SenderState related cache deadlock
This patch fixes a potential deadlock in the caches. This deadlock
could occur when more than one cache is used in a system, and
pkt->senderState is modified in between the two caches. This happened
as the caches relied on the senderState remaining unchanged, and used
it for instantaneous upstream communication with other caches.

This issue has been addressed by iterating over the linked list of
senderStates until we are either able to cast to a MSHR* or
senderState is NULL. If the cast is successful, we know that the
packet has previously passed through another cache, and therefore
update the downstreamPending flag accordingly. Otherwise, we do
nothing.
2013-02-19 05:56:06 -05:00
Andreas Hansson
0622f30961 mem: Add predecessor to SenderState base class
This patch adds a predecessor field to the SenderState base class to
make the process of linking them up more uniform, and enable a
traversal of the stack without knowing the specific type of the
subclasses.

There are a number of simplifications done as part of changing the
SenderState, particularly in the RubyTest.
2013-02-19 05:56:05 -05:00
Andreas Hansson
f69d431ede base: Fix a bug in the address interleaving
This patch fixes a minor (but important) typo in the matching of an
address to an interleaved range.
2013-02-19 05:56:05 -05:00
Andreas Hansson
9947923c60 mem: Ensure trace captures packet fields before forwarding
This patch fixes a bug in the CommMonitor caused by the packet being
modified before it is captured in the trace. By recording the fields
before passing the packet on, and then putting these values in the
trace we ensure that even if the packet is modified the trace captures
what the CommMonitor saw.
2013-02-19 05:56:05 -05:00
Anthony Gutierrez
21aa950318 options: add command line option for dtb file 2013-02-15 18:48:59 -05:00
Anthony Gutierrez
f7107fb795 loader: add a flattened device tree blob (dtb) object
this adds a dtb_object so the loader can load in the dtb
file for linux/android ARM kernels.
2013-02-15 18:48:59 -05:00
Anthony Gutierrez
ba983f9387 ext lib: add libfdt to enable flattened device tree support
this patch adds libfdt, a library necessary for supporting
flattened device tree support in current and future versions of
the linux/android kernel for ARM.
2013-02-15 18:48:59 -05:00
Ali Saidi
bd31a5dc18 stats: update regressions for o3 changes in renaming and translation. 2013-02-15 17:40:14 -05:00
Mrinmoy Ghosh
8cef39fb67 arm: fix a page table walker issue where a page could be translated multiple times
If multiple memory operations to the same page are miss the TLB they are
all inserted into the page table queue and before this change could result
in multiple uncessesary walks as well as duplicate enteries being inserted
into the TLB.
2013-02-15 17:40:10 -05:00
Andreas Sandberg
3af59ab386 cpu: Document exec trace flags 2013-02-15 17:40:10 -05:00
Andreas Sandberg
08467a88a6 dev: Use the correct return type for disk offsets
Replace the use of off_t in the various DiskImage related classes with
std::streampos. off_t is a signed 32 bit integer on most 32-bit
systems, whereas std::streampos is normally a 64 bit integer on most
modern systems. Furthermore, std::streampos is the type used by
tellg() and seekg() in the standard library, so it should have been
used in the first place. This patch makes it possible to use disk
images larger than 2 GiB on 32 bit systems with a modern C++ standard
library.
2013-02-15 17:40:10 -05:00
Geoffrey Blake
ca96e7bff1 cpu: Avoid duplicate entries in tracking structures for writes to misc regs
setMiscReg currently makes a new entry for each write to a misc reg without
checking for duplicates, this can cause a triggering of the assert if an
instruction get replayed and writes to the same misc regs multiple times.
This fix prevents duplicate entries and instead updates the value.
2013-02-15 17:40:10 -05:00
Geoffrey Blake
8e79c68936 cpu: Fix rename mis-handling serializing instructions when resource constrained
The rename can mis-handle serializing instructions (i.e. strex) if it gets
into a resource constrained situation and the serializing instruction has
to be placed on the skid buffer to handle blocking.  In this situation the
instruction informs the pipeline it is serializing and logs that the next
instruction must be serialized, but since we are blocking the pipeline
defers this action to place the serializing instruction and
incoming instructions into the skid buffer. When resuming from blocking,
rename will pull the serializing instruction from the skid buffer and
the current logic will see this as the "next" instruction that has to
be serialized and because of flags set on the serializing instruction,
it passes through the pipeline stage as normal and resets rename to
non-serializing.  This causes instructions to follow the serializing inst
incorrectly and eventually leads to an error in the pipeline. To fix this
rename should check first if it has to block before checking for serializing
instructions.
2013-02-15 17:40:10 -05:00
Chris Emmons
27630e9cad ARM: Postpones creation of framebuffer output file until it is actually used.
This delay prevents a potential conflict with the HDLCD if both are in the same
system even if only one is enabled.
2013-02-15 17:40:10 -05:00
Andreas Hansson
f6550b3d20 mem: Tighten up cache constness and scoping
This patch merely adopts a more strict use of const for the cache
member functions and variables, and also moves a large portion of the
member functions from public to protected.
2013-02-15 17:40:10 -05:00
Sascha Bischoff
2f3b322280 base: Add warn() and inform() to m5.utils for use from python
This patch adds two fuctions to m5.util, warn and inform, which mirror those
found in the C++ side of gem5. These are added in addition to the already
existing m5.util.panic and m5.util.fatal which already mirror the C++
functionality. This ensures that warning and information messages generated
by python are in the same format as those generated by C++.

Occurrences of
    print "Warning: %s..." % name
have been replaced with
    warn("%s...", name)
2013-02-15 17:40:10 -05:00
Matt Horsnell
e88e7d88b9 o3: fix tick used for renaming and issue with range selection
Fixes the tick used from rename:
- previously this gathered the tick on leaving rename which was always 1 less
  than the dispatch. This conflated the decode ticks when back pressure built
  in the pipeline.
- now picks up tick on entry.

Added --store_completions flag:
- will additionally display the store completion tail in the viewer.
- this highlights periods when large numbers of stores are outstanding (>16 LSQ
  blocking)

Allows selection by tick range (previously this caused an infinite loop)
2013-02-15 17:40:09 -05:00
Andreas Sandberg
6459908069 arm: Don't export private GIC methods 2012-10-25 14:08:29 +01:00
Andreas Sandberg
81be8b9d15 arm: Create a GIC base class and make the PL390 derive from it
This patch moves the GIC interface to a separate base class and makes
all interrupt devices use that base class instead of a pointer to the
PL390 implementation. This allows us to have multiple GIC
implementations. Future implementations will allow in-kernel GIC
implementations when using hardware virtualization.

--HG--
rename : src/dev/arm/gic.cc => src/dev/arm/gic_pl390.cc
rename : src/dev/arm/gic.hh => src/dev/arm/gic_pl390.hh
2012-10-25 14:05:24 +01:00
Andreas Sandberg
b904bd5437 sim: Add a system-global option to bypass caches
Virtualized CPUs and the fastmem mode of the atomic CPU require direct
access to physical memory. We currently require caches to be disabled
when using them to prevent chaos. This is not ideal when switching
between hardware virutalized CPUs and other CPU models as it would
require a configuration change on each switch. This changeset
introduces a new version of the atomic memory mode,
'atomic_noncaching', where memory accesses are inserted into the
memory system as atomic accesses, but bypass caches.

To make memory mode tests cleaner, the following methods are added to
the System class:

 * isAtomicMode() -- True if the memory mode is 'atomic' or 'direct'.
 * isTimingMode() -- True if the memory mode is 'timing'.
 * bypassCaches() -- True if caches should be bypassed.

The old getMemoryMode() and setMemoryMode() methods should never be
used from the C++ world anymore.
2013-02-15 17:40:09 -05:00
Andreas Sandberg
1eec115c31 cpu: Refactor memory system checks
CPUs need to test that the memory system is in the right mode in two
places, when the CPU is initialized (unless it's switched out) and on
a drainResume(). This led to some code duplication in the CPU
models. This changeset introduces the verifyMemoryMode() method which
is called by BaseCPU::init() if the CPU isn't switched out. The
individual CPU models are responsible for calling this method when
resuming from a drain as this code is CPU model specific.
2013-02-15 17:40:08 -05:00
Andreas Sandberg
1c7aa665bf config: Remove O3 dependencies
The default cache configuration script currently import the O3_ARM_v7a
model configuration, which depends on the O3 CPU. This breaks if gem5
has been compiled without O3 support. This changeset removes the
dependency by only importing the model if it is requested by the
user. As a bonus, it actually removes some code duplication in the
configuration scripts.
2013-02-15 17:40:08 -05:00
Andreas Sandberg
e5dca84c3f config: Move CPU handover logic to m5.switchCpus()
CPU switching consists of the following steps:
 1. Drain the system
 2. Switch out old CPUs (cpu.switchOut())
 3. Change the system timing mode to the mode the new CPUs require
 4. Flush caches if switching to hardware virtualization
 5. Inform new CPUs of the handover (cpu.takeOverFrom())
 6. Resume the system

m5.switchCpus() previously only did step 2 & 5. Since information
about the new processors' memory system requirements is now exposed,
do all of the steps above.

This patch adds automatic memory system switching and flush (if
needed) to switchCpus(). Additionally, it adds optional draining to
switchCpus(). This has the following implications:

* changeToTiming and changeToAtomic are no longer needed, so they have
  been removed.

* changeMemoryMode is only used internally, so it is has been renamed
  to be private.

* switchCpus requires a reference to the system containing the CPUs as
  its first parameter.

WARNING: This changeset breaks compatibility with existing
configuration scripts since it changes the signature of
m5.switchCpus().
2013-02-15 17:40:08 -05:00
Andreas Sandberg
e9f66dceac config: Cleanup CPU configuration
The CPUs supported by the configuration scripts used to be
hard-coded. This was not ideal for several reasons. For example, the
configuration scripts depend on all CPU models even though only a
subset might have been compiled.

This changeset adds a new module to the configuration scripts that
automatically discovers the available CPU models from the compiled
SimObjects. As a nice bonus, the use of introspection allows us to
automatically generate a list of available CPU models suitable for
printing. This list is augmented with the Python doc string from the
underlying class if available.
2013-02-15 17:40:08 -05:00
Andreas Sandberg
7f1263f144 cpu: Make checker CPUs inherit from CheckerCPU in the Python hierarchy
Checker CPUs currently don't inherit from the CheckerCPU in the Python
object hierarchy. This has two consequences:
 * It makes CPU model discovery from the Python world somewhat
   complicated as there is no way of testing if a CPU is a checker.
 * Parameters are duplicated in the checker configuration
   specification.

This changeset makes all checker CPUs inherit from the base checker
CPU class.
2013-02-15 17:40:08 -05:00
Andreas Sandberg
7cd1fd4324 cpu: Add CPU metadata om the Python classes
The configuration scripts currently hard-code the requirements of each
CPU. This is clearly not optimal as it makes writing new configuration
scripts painful and adding new CPU models requires existing scripts to
be updated. This patch adds the following class methods to the base
CPU and all relevant CPUs:

 * memory_mode -- Return a string describing the current memory mode
                  (invalid/atomic/timing).

 * require_caches -- Does the CPU model require caches?

 * support_take_over -- Does the CPU support CPU handover?
2013-02-15 17:40:08 -05:00
Ali Saidi
db5c478e70 arm: fix some fp comparisons that worked by accident.
The explict tests in the follwing fp comparison operations were
incorrect as they checked for only signaling NaNs and not quite-NaNs
as well. When compiled with gcc, the comparison generates a fp exception
that causes the FE_INVALID flag to be set and we check for it, so even
though the check was incorrect, the correct exception was set. With clang
this behavior seems to not occur. The checks are updated to test for nans and
the behavior is now correct with both clang and gcc.
2013-02-15 17:40:08 -05:00
Ali Saidi
4412046041 cpu: include set in o3/commit_impl.
While the majority of compilers seemed to pickup set from else where,
one version of gcc 4.7 complains, so explictly add it.
2013-02-15 17:40:08 -05:00
Ali Saidi
68495a0748 ARM: Fix an issue with clang generating wrong code.
Clang generated executables would enter the if condition when it wasn't
supposted to, resulting in the wrong simulated behavior.
Implementing the operation this way is a bit faster anyway.
2013-02-15 17:40:08 -05:00
Ali Saidi
7ae06a3b3b cpu: fix case with o3 cpu blocking and unblocking decode in cycle
Fix a case in the O3 CPU where the decode stage blocks and unblocks in a
single cycle sending both signals to fetch which causes an assert or worse.
The previous check could never work before since the status was set to Blocked
before a test for the status being Unblocking was executed.
2013-02-15 17:40:08 -05:00
Ali Saidi
b84bd3028c cpu: Fix a livelock in the o3 cpu.
Check if an instruction just enabled interrupts and we've previously had an
interrupt pending that was not handled because interrupts were subsequently
disabled before the pipeline reached a place to handle the interrupt. In that
case squash now to make sure the interrupt is handled.
2013-02-15 17:40:07 -05:00
Andreas Sandberg
d4eca0591d base: Add support for newer versions of IPython
IPython is used for the interactive gem5 shell if it exists. IPython
made API changes in version 0.11. This patch adds support for IPython
version 0.11 and above.

--HG--
extra : rebase_source : 5388d0919adb58d97f49a1a637db48cba61283a3
2013-02-10 13:23:58 +01:00
Andreas Hansson
7c6bc52bf5 Ruby: Fix compilation errors on gcc 4.7 and clang 3.2
This patch fixes a few (recently added) errors that prevented gem5 from
compiling on more recent versions of gcc and clang.
2013-02-14 12:24:51 -05:00
Nilay Vaish
1962e9262d regressions: update stats due to changes to ruby 2013-02-10 21:43:23 -06:00
Nilay Vaish
71c27e6370 ruby: MI protocol: add a missing transition
The transition for state MII and event Store was found missing during testing.
The transition is being added. The controller will not stall the Store request
in state MII
2013-02-10 21:43:18 -06:00
Nilay Vaish
cb7782f78d ruby: enable multiple clock domains
This patch allows ruby to have multiple clock domains. As I understand
with this patch, controllers can have different frequencies. The entire
network needs to run at a single frequency.

The idea is that with in an object, time is treated in terms of cycles.
But the messages that are passed from one entity to another should contain
the time in Ticks. As of now, this is only true for the message buffers,
but not for the links in the network. As I understand the code, all the
entities in different networks (simple, garnet-fixed, garnet-flexible) should
be clocked at the same frequency.

Another problem is that the directory controller has to operate at the same
frequency as the ruby system. This is because the memory controller does
not make use of the Message Buffer, and instead implements a buffer of its
own. So, it has no idea of the frequency at which the directory controller
is operating and uses ruby system's frequency for scheduling events.
2013-02-10 21:43:17 -06:00
Nilay Vaish
253e8edf13 ruby: replace Time with Cycles (final patch in the series)
This patch is as of now the final patch in the series of patches that replace
Time with Cycles.This patch further replaces Time with Cycles in Sequencer,
Profiler, different protocols and related entities.

Though Time has not been completely removed, the places where it is in use
seem benign as of now.
2013-02-10 21:43:10 -06:00
Nilay Vaish
f6e3ab7bd4 ruby: replace Time with Cycles in garnet fixed and flexible 2013-02-10 21:43:09 -06:00
Nilay Vaish
9d6d6c6718 ruby: replace Time with Tick in replacement policy classes 2013-02-10 21:43:08 -06:00
Nilay Vaish
221d39284e ruby: convert block size, memory size to unsigned 2013-02-10 21:43:07 -06:00
Nilay Vaish
5e33045a2a ruby: replace Time with Cycles in MessageBuffer 2013-02-10 21:26:26 -06:00
Nilay Vaish
b742081cc1 ruby: replace Time with Cycles in Memory Controller 2013-02-10 21:26:25 -06:00
Nilay Vaish
89f86dbd28 ruby: Replace Time with Cycles in SequencerMessage 2013-02-10 21:26:25 -06:00
Nilay Vaish
7862478eef ruby: replace Time with Cycles in Message class
Concomitant changes are being committed as well, including the io operator<<
for the Cycles class.
2013-02-10 21:26:24 -06:00
Nilay Vaish
d3aebe1f91 ruby: replaces Time with Cycles in many places
The patch started of with replacing Time with Cycles in the Consumer class.
But to get ruby to compile, the rest of the changes had to be carried out.
Subsequent patches will further this process, till we completely replace
Time with Cycles.
2013-02-10 21:26:24 -06:00
Nilay Vaish
affd77ea77 base: add some mathematical operators to Cycles class 2013-02-10 21:26:23 -06:00