While the sim code doesn't utilize HAVE_LONG_LONG itself, other code
(like libiberty) seem to, so check for it in the top-level for all
ports to leverage.
Move the stub logic to the device files themselves. This makes the
configure & build logic more static which will make it easier to move
to the top-level build, and matches what we did with the common/ hw
tree already.
This also decouples the logic from the two -- in the past, you needed
both sem & shm in order to enable the device models, but now each one
is tied to its own independent knob. Practically speaking, this will
probably not make a difference, but it simplifies the build a bit.
Instead of executing code to see if SysV semaphores & shared memory
are available, switch to just a compile-time test. The system used
to compile might not match the system used to run the code wrt the
current kernel & OS settings, but the library APIs should. So move
the failures from compile-time to runtime so the program is more
portable, and works correctly even when cross-compiling.
Compile tests can use earlier defines, so hoist the HAVE_UNION_SEMUN
define to before the semaphore check, and use it in the test so that
we can merge the 2 versions into one.
This also defines HAVE_UNION_SEMUN even when ac_cv_sysv_sem is not
set, but that's OK as this define is only about a type existing, not
about whether the overall code is usable.
The first arg is the cache var name, and this one was typoed relative
to what the call actually set. We also don't need the manual call to
AC_MSG_RESULT as the AC_CACHE_CHECK takes care of it for us.
The common igen code provides the same misc APIs as the ppc version,
so delete the ppc code and pull in the common one. There is one
minor difference: the ppc code has a unique dumpf function. The
common code switched to lf_printf for the same functionality, but
since that requires changes throughout the igen codebase, delay that
cleanup for now so we can merge the rest.
We want to avoid conflicts with the common igen enums. This should
get migrated over to the common parsing logic, but for now, switch
the name to avoid redefinition.
Now that both igen implementations are in the top-level, we can unify
the filter_filename implementation between them since they're the same
(literally the same code).
The lbasename function from libiberty provides the same API as this
custom function. The common/ code already made the switch, so make
the same change to the ppc code to avoid target duplication.
This simplifies the build a bit (especially for deps in port subdirs),
and avoids recursive make. This in turn speeds up the build, and lets
us reuse existing build-time vs host-time logic from Makefile.am.
This header is only used by the igen tool, and none of the igen code
depends on the configure-time checks. Delete the logic to simplify
to prepare for moving it to the local.mk code.
Switch this from a build-time generation to a static include. This
makes the build rules a bit simpler, especially as we move them to
Automake from hand-written makefiles.
This code sets up the cc variable based on the comparison of other
registers, but it does so incrementally with bit operations, and it
never initializes the cc variable. Initialize it to 0 which the
cmpba insn is already doing.
The PRU architecture documentation [1] was used for the initial GNU
simulator implementation. But recently [2] TI confirmed the carry
behaviour was wrongly documented. In reality, the PRU carry behaves
like the carry in ARM processors.
This patch fixes simulator to align with latest recommendations from TI.
The new carry.s test was also validated to pass on real hardware -
a BeaglePlay board [3]. That test is a bit long because TI still
has not released official updates for the PRU documents. And I wanted
to ensure simulator handles all edge cases exactly as the real hardware
does.
[1] https://www.ti.com/lit/pdf/spruij2
[2] https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1244359/sk-am64b-am64x-pru-assembler-how-works-this-bloody-carry
[3] https://www.beagleboard.org/boards/beagleplay
Signed-off-by: Dimitar Dimitrov <dimitar@dinux.eu>
Since newlib allows people to opt-in to extra errno names, pull them
into our table too. The values don't conflict with each other -- the
newlib names & values are distinct from newlib's Linux compatibility.
Since this function never returns, mark it as such to fix some unused
variable warnings in error code paths.
For example, cris triggers:
sim/cris/semcrisv10f-switch.c:3558:11: error:
variable 'tmp_newval' is used uninitialized whenever 'if' condition is false [-Werror,-Wsometimes-uninitialized]
Even though it has an "else" path that calls this error function.
Since these insns don't do anything and are effectively ignored,
return early to avoid doing any common processing at the end as
that requires initializing variables like "res" with something.
The plds Dz,MACL insn stores the Dz bit into MACL. The current code
was storing the "res" variable into Dz and then into MACL, but not
setting "res" to anything. Delete that logic and make it match the
existing plds Dz,MACH insn.
These decoders declare a lot of common variables for use by substeps,
and then shadows a few because of how the opc generator is implemented.
Easiest way around it is to rename the per-substep vars as needed as
anything more would require substantial changes to the opc logic.
The mloop files include CGEN generated switch files which have some
nested assignments that expand into repeated shadowed variables.
Fixing this looks fairly non-trivial as it appears to be interplay
between the common CGEN code and how this particular set of cris
insns are defined. Disable the warning instead.
In file included from sim/cris/mloop.in:286:
sim/cris/semcrisv10f-switch.c: In function ‘crisv10f_engine_run_full’:
sim/cris/semcrisv10f-switch.c:12383:8: error: declaration of ‘opval’ shadows a previous local [-Werror=shadow=local]
12383 | SI opval = tmp_addr;
| ^~~~~
sim/cris/semcrisv10f-switch.c:12371:9: note: shadowed declaration is here
12371 | USI opval = ({ SI tmp_addr;
| ^~~~~
And the code looks like:
USI opval = ({
...
{
SI opval = tmp_addr;
...
}
...
});
Since the CGEN code treats "opval" as an internal variable that the cpu
definitions don't have direct access to, the likelihood of this being a
real bug is low, so leave it be. The warning is suppressed for more code
that is hand written (e.g. the mloop logic), but disabling for the entire
file is the easiest way to suppress while keeping it on everywhere else in
the sim.
