Some compilers won't build ARM due to how guest ABI
has been implemented.
The error is: "left shift count >= width of type"
[-Werror=shift-count-overflow]
The error is triggered when there is a left shift > the variable size
(in bits); this leads to undefined behaviour.
This is a compile time vs run time problem; the code is technically
fine, but the compiler is not able to understand this.
For example in aapcs64:
struct Argument<Aapcs64, Integer, typename std::enable_if<
std::is_integral<Integer>::value>::type> : public Aapcs64ArgumentBase
{
[...]
if (sizeof(Integer) == 16 && state.ngrn + 1 <= state.MAX_GRN) {
Integer low = tc->readIntReg(state.ngrn++);
Integer high = tc->readIntReg(state.ngrn++);
high = high << 64;
return high | low;
}
}
Even if the sizeof operator will be evaluated at compile time,
the block will be executed at runtime: the block will still be part of
the code if Integer = uint32_t.
The compiler will then throw an error because we are left shifting an
uint32_t by 64 bits.
Error arising on:
Compiler: gcc/5.4.0
Distro: Ubuntu 16.04 LTS
Change-Id: Iaafe030b7262c5fb162afe7118ae592a1a759a58
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26990
Reviewed-by: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
When handling a system call, external code would call Process::syscall
which would extract the syscall number, that would call the base
class' doSyscall method, that would call into the subclass' getDesc
to get the appropriate descriptor, and then doSyscall would check
that a syscall was found and call into it.
Instead, we can just make the SyscallDescTable optionally check for
missing syscalls (in case we want to check multiple tables), and
make syscall look up the appropriate descriptor and call it. The base
implementation of syscall would then do the only bit of doSyscall that
is no longer being handled, incrementing the numSyscalls stat.
Change-Id: If102c156830ed2997d177dc6937cc85dddadf3f9
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/24119
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Gem5 Cloud Project GCB service account <345032938727@cloudbuild.gserviceaccount.com>
Maintainer: Gabe Black <gabeblack@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Also add the syscall number into the SyscallDesc class.
The common table structure is basically just a map that extracts its
key value from the SyscallDesc class using a new num() accessor. By
using a map instead of an array (like RISCV was already doing), it's
easy to support gaps of arbitrary size and non-zero offsets of groups
of system calls without lots of filler or additional logic. This
simplified the ARM system call tables in particular which had a lot
of filler entries.
Also, both the 32 and 64 bit ARM syscall tables had entries for a
syscall at 123456 which was the "Angel SWI system call". This value
is actually the immediate constant passed to the SWI system call
instruction and is not interpreted as the system call number in linux.
This constant can be intercepted by hardware or a simulator to, for
instance, implement ARM semihosting.
Also, that constant in combination with the SWI instruction is only
used for semihosting in 32 bit ARM mode, not in 64 bit mode or in
thumb.
Since checking for that system call number was very likely a mistake
from misinterpreting how the semihosting calls work, this change
drops those checks.
Change-Id: I9b2a902d7326791449cf0e1b98e932dcadba54f7
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/24117
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
With:
https://gem5-review.googlesource.com/c/public/gem5/+/26466
The ArmSystem reset address (_resetAddr) is always forced by the
workload:
_resetAddr = workload->entry
So there is no possibility to manually specify a reset address.
This was not the case before:
The resetAddr was forced only if auto_reset_addr was true or if there
was an associated bootloader to the kernel image. In that case even if
auto_reset_addr was false, the reset address was determined by the
bootloader entry.
This was also not ideal (but it was working)
This patch is cleaning all of this:
If you want to have automatic detection (recommended), you would need to
set auto_reset_addr (now turned to true by default). This will allow to
keep most fs script untouched. If you don't want to use automatic
detection, set auto_reset_addr to False and provide your own reset
address.
Change-Id: I5d7a55fd9060b9973c7d5b5542bd199950e1073e
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26723
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Gabe Black <gabeblack@google.com>
The only functional difference between them was that the SE one might
have optionally fixed up missing translations for demand paging.
This lets us get rid of some code recreating the proxy ports in
setProcessPtr since the SE translating port no longer keeps a copy of
the process object pointer.
Change-Id: Id97df1874f1de138ffd4f2dbb5846dda79d9e4ac
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26550
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Matthew Poremba <matthew.poremba@amd.com>
Maintainer: Gabe Black <gabeblack@google.com>
This function is no longer used anywhere in gem5.
Small helper functions which had been put alongside vtophys on ARM and
RISCV were also moved into src/arch/arm/remote_gdb.cc and
src/arch/power/pagetable.hh, the only places they were used.
Change-Id: Iba72f6c4b797a35a785a5bb781d602c943541fa7
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26234
Reviewed-by: Jason Lowe-Power <power.jg@gmail.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
* dict methods dict.keys(), dict.items() and dict.values()
return "views" instead of lists
* The dict.iterkeys(), dict.iteritems() and dict.itervalues()
methods are no longer supported.
* map() and filter() return iterators.
* range() now behaves like xrange() used to behave, except it works with
values of arbitrary size. The latter no longer exists.
* zip() now returns an iterator.
Change-Id: Id480018239db88d7f5d60588c93719056de4a0c0
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26248
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
In case a workload is run with no bootloader we still want to be able
to provide the simulation with the correct arch version.
Without this patch every baremetal simulation will default to AArch64, which
is the default value of ArmFsWorkload.
Change-Id: I0f766167d8983cafc1fd30d054862339eb21f73f
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26606
Reviewed-by: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
This flag makes constructing a generic Request object for ARM
impossible, since generic consumers won't know to set that flag to one.
As a principle, Request flags should change the behavior of a request
away from whatever the default/typical behavior is in the current
context, so flags set to zero means no special behavior.
Change-Id: Id606dc0bf42210218e1745585327671a98a8dba4
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26546
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
This is a reduced model of the FVP Base platforms Power Controller.
As of now it allows the following functions from software:
- Checking for core presence
- Reporting the power state of a core / cluster
- Explicitly powering off a core on WFI
- Explicitly powering off cores in a CPU cluster on WFI
- Explicitly powering on a core through writes to PPONR register
Change-Id: Ia1d4d3ae8e4bfb2d23b2c6077396a4d8500e2e30
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26463
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
This implementation has been tested a tiny bit by intercepting a call
which passed an argument of this type to a function.
struct Test
{
int32_t a;
float *b;
};
The gem5 intercept printed out the value of a, the value of b, and the
value of the float it pointed to.
I was able to get things to work by commenting out the panic in
fixFuncEventAddr and making it return its argument unmodified, and by
calling addFuncEvent instead of addKernelFuncEvent which injects the
kernel symbol table. I substitured the Process's debugSymbolTable which
had the right symbols.
Note that this implementation is not completely correct. First of all,
I used a dummy type in place of the Short Vector type which is just
a byte array with the appropriate alignment forced on it. It sounds
like this type would be something the compiler would need an intrinsic
and architecture specific type for to behave correctly, and so in
gem5 we'd have to define our own type for ARM which could feed in here.
Also, strictly speaking, it sounds like HVA and HFA category of types,
the Homogeneous Short-Vector Aggregates and Homogeneous Floating-point
Aggregates, are supposed to apply to any type which is an aggregate of
all the same type (short vector for one, floating point for the other)
with 4 or fewer members.
In this implementation, I capture any *array* of 4 or fewer elements of
the appropriate type as an HVA or HFA, but I believe these structures
would also count and are not included in my implementation.
struct {
float a;
float b;
float c;
};
struct {
ShortVector a;
ShortVector b;
};
This only matters if those sorts of structures are passed by value as
top level arguments to a function, ie they are not included in some
larger structure.
Also, rule B.6 talks about what to do with an "aignment adjusted type",
and I have no idea what that's supposed to be. Those may not be handled
correctly either.
Change-Id: I5a599a03d38075d7c0a06988c05e7fb5423c68c0
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/23751
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Gabe Black <gabeblack@google.com>
Information about what kernel to load and how to load it was built
into the System object and its subclasses. That overloaded the System
object and made it responsible for too many things, and also was
somewhat awkward when working with SE mode which doesn't have a kernel.
This change extracts the kernel and information related to it from the
System object and puts into into a OsKernel or Workload object.
Currently the idea of a "Workload" to run and a kernel are a bit
muddled, an unfortunate carry-over from the original code. It's also an
implication of trying not to make too sweeping of a change, and to
minimize the number of times configs need to change, ie avoiding
creating a "kernel" parameter which would shortly thereafter be
renamed to "workload".
In future changes, the ideas of a kernel and a workload will be
disentangled, and workloads will be expanded to include emulated
operating systems which shephard and contain Process-es for syscall
emulation.
This change was originally split into pieces to make reviewing it
easier. Those reviews are here:
https: //gem5-review.googlesource.com/c/public/gem5/+/22243
https: //gem5-review.googlesource.com/c/public/gem5/+/24144
https: //gem5-review.googlesource.com/c/public/gem5/+/24145
https: //gem5-review.googlesource.com/c/public/gem5/+/24146
https: //gem5-review.googlesource.com/c/public/gem5/+/24147
https: //gem5-review.googlesource.com/c/public/gem5/+/24286
Change-Id: Ia3d863db276a023b6a2c7ee7a656d8142ff75589
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26466
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
getArmSystem is the building block for a lot of ArmSystem getters
a client can use to check for a specific feature.
This method is called very often during simulation and it is basically
casting a System pointer into an ArmSystem pointer.
To do so, it is using dynamic casting to check if the system is really
an ArmSystem. This is very expensive and usually not needed.
The only chance arm code would use a non ArmSystem is when in SE mode.
But if that's the case, we can just replace the assertion with a
assert(FullSystem).
Testing Linux boot with this patch provides a speedup of nearly 2x!
(atomic mode).
This is partially related to:
JIRA: https://gem5.atlassian.net/browse/GEM5-337
Since the PAuth patch changed the purifyTagged helper (on the critical
path of simulation) to rely more heavilly on getArmSystem (via
ArmSystem:: static methods)
Change-Id: Idbf079548ffe03513b4fc58c76f0d69613952a50
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Ciro Santilli <ciro.santilli@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/25964
Tested-by: kokoro <noreply+kokoro@google.com>
The version of vtophys which didn't take a ThreadContext had only been
implemented on Alpha which has since been removed, so this version of
the function was completely unimplemented and never used.
This change also gets rid of the dbg_vtophys which was sometimes
implemented but also never used, and takes the opportunity to fix up
some style problems in some of the vtophys arch files.
Change-Id: Ie10f881f8ce08c7188e71805357cf3264be4c81a
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/26224
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
The new local access mechanism installs a callback in the request which
implements what the mmapped IPR was doing. That avoids having to have
stubs in ISAs that don't have mmapped IPRs, avoids having to encode
what to do to communicate from the TLB and the mmapped IPR functions,
and gets rid of another global ISA interface function and header files.
Jira Issue: https://gem5.atlassian.net/browse/GEM5-187
Change-Id: I772c2ae2ca3830a4486919ce9804560c0f2d596a
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/23188
Reviewed-by: Matthew Poremba <matthew.poremba@amd.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Currently the System class has a mechanism to wait for a GDB connection
for each CPU which has requested it through one of its parameters.
Unfortunately, not every thread context/CPU will be ready for GDB at
that point, particularly considering that in an FS simulation the
kernel won't have been read so there will be no symbols, none of the
registers or the entry point will have been set.
Also in the fast models, the CPUs haven't had a chance to initialize
themselves enough by that point to respond to the API calls which are
used to implement GDB support.
Change-Id: If27cb3e0259a1f67599ab0493695b2f8af640d8e
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/24963
Reviewed-by: Gabe Black <gabeblack@google.com>
Reviewed-by: Chun-Chen TK Hsu <chunchenhsu@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
When defining a static constexpr variable in C++11, it is still
required to have a separate definition someplace, something that can
be particularly problematic in template classes. C++17 fixes this
problem by adding inline variables which don't, but in the mean time
having a static constexpr value with no backing store will, if the
compiler decides to not fold away the storage location, cause linking
errors.
This happened to me when trying to build the debug build of ARM just
now.
By turning these expressions into static inline functions, then they
no longer need definitions elsewhere, still fold away to nothing, and
are compliant with C++11 which is currently the standard gem5 expects
to be using.
Change-Id: I647d7cf4a1e8de98251ee9ef116f007e08eac1f3
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/24964
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chun-Chen TK Hsu <chunchenhsu@google.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
gem5 does not historically distinguish between address spaces when
interacting with gdb, and gdb doesn't really give it any address space
information to work with. To ensure we catch whatever address space
we might be in by the time we get to the interesting address, we'll set
a breakpoint in all possible address spaces simultaneously with the
expectation that we'll hit one of them.
Change-Id: I9f4b93d04914db7a3c42be6236a523d35194afda
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/25268
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
Reviewed-by: Chun-Chen TK Hsu <chunchenhsu@google.com>
When the last event is removed from a breakpoint, then the breakpoint
itself is uninstalled from IRIS, and the list is deleted. Even though
the list has been traversed and so we don't lose track of any other
events that need to be processed, we also still need to check against
end() to see that we're done. If that now freed memory gets
overwritten, then we won't see the end and will wander right off the
end of the list into nonsense.
This change modifies the breakpoint info tracking structure to keep a
shared pointer to the event list. The pointer will still automatically
manage the list's memory so that it doesn't leak, and it won't get
deleted out from under us as we're iterating through it.
Change-Id: I5ad0f095d07f0a3a5cce9c10f03121827a674c33
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/24965
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chun-Chen TK Hsu <chunchenhsu@google.com>
