- small cleanup for xtensa registers dumping (-d cpu);
- add support for debugging linux-user process with xtensa-linux-gdb
(as opposed to xtensa-elf-gdb), which can only access unprivileged
registers;
- enable MTTCG for target/xtensa;
- cleanup in linux-user/mmap area making sure that it works correctly
with limited 30-bit-wide user address space;
- import xtensa-specific definitions from the linux kernel,
conditionalize user-only/softmmu-only code and add handlers for
signals, exceptions, process/thread creation and core registers dumping.
-----BEGIN PGP SIGNATURE-----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=UCny
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/xtensa/tags/20180316-xtensa' into staging
target/xtensa linux-user support.
- small cleanup for xtensa registers dumping (-d cpu);
- add support for debugging linux-user process with xtensa-linux-gdb
(as opposed to xtensa-elf-gdb), which can only access unprivileged
registers;
- enable MTTCG for target/xtensa;
- cleanup in linux-user/mmap area making sure that it works correctly
with limited 30-bit-wide user address space;
- import xtensa-specific definitions from the linux kernel,
conditionalize user-only/softmmu-only code and add handlers for
signals, exceptions, process/thread creation and core registers dumping.
# gpg: Signature made Fri 16 Mar 2018 16:46:19 GMT
# gpg: using RSA key 51F9CC91F83FA044
# gpg: Good signature from "Max Filippov <filippov@cadence.com>"
# gpg: aka "Max Filippov <max.filippov@cogentembedded.com>"
# gpg: aka "Max Filippov <jcmvbkbc@gmail.com>"
# Primary key fingerprint: 2B67 854B 98E5 327D CDEB 17D8 51F9 CC91 F83F A044
* remotes/xtensa/tags/20180316-xtensa:
MAINTAINERS: fix W: address for xtensa
qemu-binfmt-conf.sh: add qemu-xtensa
target/xtensa: add linux-user support
linux-user: drop unused target_msync function
linux-user: fix target_mprotect/target_munmap error return values
linux-user: fix assertion in shmdt
linux-user: fix mmap/munmap/mprotect/mremap/shmat
target/xtensa: support MTTCG
target/xtensa: use correct number of registers in gdbstub
target/xtensa: mark register windows in the dump
target/xtensa: dump correct physical registers
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
# Conflicts:
# linux-user/syscall.c
Import list of syscalls from the kernel source. Conditionalize code/data
that is only used with softmmu. Implement exception handlers. Implement
signal hander (only the core registers for now, no coprocessors or TIE).
Cc: Riku Voipio <riku.voipio@iki.fi>
Cc: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Now we've dropped unicore32, all of the architectures we support
for linux-user implement the signal handling routines. The
dummy "just print a message" versions are unimplemented, so we
can drop them entirely.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20180308144733.25615-3-peter.maydell@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
We dropped the unicore32-linux-user target in commit 5e2b40f727
in 2016. Nobody has made any attempt to fix the issues that
caused us to drop it, so remove the associated code.
(The system emulation parts of unicore32 remain.)
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20180308144733.25615-2-peter.maydell@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Depending on the currently selected size of the SVE vector registers,
we can either store the data within the "standard" allocation, or we
may beedn to allocate additional space with an EXTRA record.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20180303143823.27055-6-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The EXTRA record allows for additional space to be allocated
beyon what is currently reserved. Add code to emit and read
this record type.
Nothing uses extra space yet.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20180303143823.27055-5-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This changes the qemu signal frame layout to be more like the kernel's,
in that the various records are dynamically allocated rather than fixed
in place by a structure.
For now, all of the allocation is out of uc.tuc_mcontext.__reserved,
so the allocation is actually trivial. That will change with SVE support.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20180303143823.27055-4-richard.henderson@linaro.org
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Split out helpers from target_setup_frame and target_restore_sigframe
for dealing with general registers, fpsimd registers, and the end record.
When we add support for sve registers, the relative positions of
these will change.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20180303143823.27055-3-richard.henderson@linaro.org
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Implementation of linux user emulation for RISC-V.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Sagar Karandikar <sagark@eecs.berkeley.edu>
Signed-off-by: Michael Clark <mjc@sifive.com>
The SPARC code in linux-user/signal.c defines a set of
MC_* constants. On some SPARC hosts these are also defined
by sys/ucontext.h, resulting in build failures:
linux-user/signal.c:2786:0: error: "MC_NGREG" redefined [-Werror]
#define MC_NGREG 19
In file included from /usr/include/signal.h:302:0,
from include/qemu/osdep.h:86,
from linux-user/signal.c:19:
/usr/include/sparc64-linux-gnu/sys/ucontext.h:59:0: note: this is the location of the previous definition
# define MC_NGREG __MC_NGREG
Rename all these constants to SPARC_MC_* to avoid the clash.
Cc: qemu-stable@nongnu.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1517318239-15764-1-git-send-email-peter.maydell@linaro.org
Helpers that return a pointer into env->vfp.regs so that we isolate
the logic of how to index the regs array for different cpu modes.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20180119045438.28582-7-richard.henderson@linaro.org
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Since for aarch64 the signal trampoline is synthesized directly into the
signal frame we need to make sure the instructions end up little-endian.
Otherwise the wrong endianness will cause a SIGILL upon return from the
signal handler on big-endian targets.
Signed-off-by: Michael Weiser <michael.weiser@gmx.de>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20171220212308.12614-4-michael.weiser@gmx.de
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Use cpu_m68k_get_ccr()/cpu_m68k_set_ccr() to setup and restore correctly
the value of SR in the context structure. Fix target_rt_setup_ucontext().
Fixes: 3219de458c ("linux-user: correctly manage SR in ucontext")
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20180104012913.30763-4-laurent@vivier.eu>
Thanks to Laszlo Ersek for spotting the double semicolon in target/i386/kvm.c
I have trivially grepped the tree for ';;' in C files.
Suggested-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Ladi Prosek <lprosek@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reading and writing to an sa_restorer member that isn't supposed to
exist corrupts user memory. Introduce TARGET_ARCH_HAS_SA_RESTORER,
similar to the kernel's __ARCH_HAS_SA_RESTORER.
Reported-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
glibc used to have:
typedef struct ucontext { ... } ucontext_t;
glibc now has:
typedef struct ucontext_t { ... } ucontext_t;
(See https://sourceware.org/bugzilla/show_bug.cgi?id=21457
for detail and rationale for the glibc change)
However, QEMU used "struct ucontext" in declarations. This is a
private name and compatibility cannot be guaranteed. Switch to
only using the standardized type name.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Message-id: 20170628204452.41230-1-raj.khem@gmail.com
Cc: Kamil Rytarowski <kamil@netbsd.org>
Cc: Riku Voipio <riku.voipio@iki.fi>
Cc: Laurent Vivier <laurent@vivier.eu>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[PMM: Rewrote commit message, based mostly on the one from
Nathaniel McCallum]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
If a signal is delivered during the execution of a delay slot,
or a gUSA region, clear those bits from the environment so that
the signal handler does not start in that same state.
Cleaning the bits on signal return is paranoid good sense.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Message-Id: <20170718200255.31647-10-rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
We translate gUSA regions atomically in a parallel context.
But in a serial context a gUSA region may be interrupted.
In that case, restart the region as the kernel would.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Message-Id: <20170718200255.31647-9-rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Shadow registers are part of the openrisc spec along with sr[cid], as
part of the fast context switching feature. When exceptions occur,
instead of having to save registers to the stack if enabled the CID will
increment and a new set of registers will be available.
This patch only implements shadow registers which can be used as extra
scratch registers via the mfspr and mtspr if required. This is
implemented in a way where it would be easy to add on the fast context
switching, currently cid is hardcoded to 0.
This is need for openrisc linux smp kernels to boot correctly.
Signed-off-by: Stafford Horne <shorne@gmail.com>
do_rt_sigreturn uses an uninitialised local variable instead of fetching
the old signal mask directly from the signal frame when restoring the mask,
so the signal mask is undefined after do_rt_sigreturn. As the signal
frame data is in target-endian order, target_to_host_sigset instead of
target_to_host_sigset_internal is required.
do_sigreturn is correct in using target_to_host_sigset_internal, because
get_user already did the endianness conversion.
Signed-off-by: Michael Karcher <karcher@physik.fu-berlin.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20170225110517.2832-3-laurent@vivier.eu>
Use cpu_m68k_get_ccr()/cpu_m68k_set_ccr() to setup and restore correctly
the value of SR in the ucontext structure
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20170225110517.2832-2-laurent@vivier.eu>
Note that x86_64 has only _rt signal handlers. This implementation
attempts to share code with the x86_32 implementation.
CC: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Allan Wirth <awirth@akamai.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20170226165345.8757-1-bobby.prani@gmail.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
A segfault is noticed when an emulated program uses any of ucontext
regs fields. Risu detected this issue in the following operation when
handling a signal:
ucontext_t *uc = (ucontext_t*)uc;
uc->uc_mcontext.regs->nip += 4;
but this works fine:
uc->uc_mcontext.gp_regs[PT_NIP] += 4;
This patch set regs to a valid location as well as other sigcontext
fields.
Signed-off-by: Jose Ricardo Ziviani <joserz@linux.vnet.ibm.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <1485900317-3256-1-git-send-email-joserz@linux.vnet.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Add missing bits for qemu-user required for emulating Altera Nios2
userspace binaries.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Chris Wulff <crwulff@gmail.com>
Cc: Jeff Da Silva <jdasilva@altera.com>
Cc: Ley Foon Tan <lftan@altera.com>
Cc: Sandra Loosemore <sandra@codesourcery.com>
Cc: Yves Vandervennet <yvanderv@altera.com>
Reviewed-by: Alexander Graf <agraf@suse.de>
Message-Id: <20170118220146.489-4-marex@denx.de>
Signed-off-by: Richard Henderson <rth@twiddle.net>
There were a number of bugs in the implementation:
- The structure alignment was wrong for 64-bit.
- Also 64-bit only does RT signals.
- On 64-bit, we need to put a pointer to the (aligned) vector registers
in the frame and use it for restoring
- We had endian bugs when saving/restoring vector registers
- My recent fixes for exception NIP broke sigreturn in user mode
causing us to resume one instruction too far.
- Add VSR second halves
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
struct target_ucontext_v2 is not at the begining of the signal frame,
therefore do_sigaltstack was being passed bogus arguments.
As the offset depends on the type of signal frame fixed by passing in the
beginning of the context from do_sigreturn_v2 and do_rt_sigreturn_v2.
Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Timothy Edward Baldwin <T.E.Baldwin99@members.leeds.ac.uk>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Now that we have a force_sig() with the semantics we need,
we can implement force_sigsegv() to call it rather than
open-coding the call to queue_signal().
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
If the sigreturn syscall fails to read memory then this causes a
SIGSEGV, but this is not necessarily a fatal signal -- the guest
process can catch it.
We don't implement this correctly because the behaviour of QEMU's
force_sig() function has drifted away from the kernel function of the
same name -- ours now does "always do a guest core dump and abort
execution", whereas the kernel version simply forces the guest to
take a signal, which may or may not eventually cause a core dump.
Rename our force_sig() to dump_core_and_abort(), and provide a
force_sig() which acts more like the kernel version as the sigreturn
implementations expect it to. Since force_sig() now returns, we must
update all the callsites to return -TARGET_QEMU_ESIGRETURN so that
the main loop doesn't change the guest registers before the signal
handler is invoked.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
The 32-bit ARM signal frame setup code was just bailing out
on error returns from lock_user_struct calls, without
generating the SIGSEGV that should happen here. Wire up
error return codes to call force_sigsegv().
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
A failed write to memory trying to set up the signal frame
should trigger a SIGSEGV, but this need not be fatal: the
guest has a chance to catch it. Implement this via a force_sigsegv()
function with the same behaviour as the kernel function of that
name: make sure that we don't try to re-take a failed SIGSEGV,
and force a synchronous signal.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Instead of assuming in queue_signal() that all callers are passing
a siginfo structure which uses the _sifields._sigfault part of the
union (and thus a si_type of QEMU_SI_FAULT), make callers pass
the si_type they require in as an argument.
[RV adjusted to apply]
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
In process_pending_signals() we restart the scan of possible
pending signals after calling handle_pending_signal() in
case some other signal has been generated. This rescan
should also include a check for a new synchronous signal
since those are in fact the only kind of new signal that
the signal frame setup process might produce.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Native strace reports when the process being traced takes a signal:
--- SIGSEGV {si_signo=SIGSEGV, si_code=SI_KERNEL, si_addr=0} ---
Report something similar when QEMU is doing its internal strace of
the guest process and is about to deliver it a signal.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
This patch is the result of coccinelle script
scripts/coccinelle/typecast.cocci
CC: Riku Voipio <riku.voipio@iki.fi>
CC: Alexander Graf <agraf@suse.de>
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Commit 655ed67c2a which switched synchronous signals to
benig recorded in ts->sync_signal rather than in a queue
with every other signal had a bug: we failed to clear
the flag indicating that a synchronous signal was pending
when we delivered it. This meant that we would take the signal
again and again every time the guest made a syscall.
(This is a bug introduced in my refactoring of Timothy Baldwin's
original code.)
Fix this by passing in the struct emulated_sigtable* to
handle_pending_signal(), so that we clear the pending flag
in the ts->sync_signal struct when handling a synchronous signal.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
The kernel and libc have different ideas about what a sigset_t
is -- for the kernel it is only _NSIG / 8 bytes in size (usually
8 bytes), but for libc it is much larger, 128 bytes. In most
situations the difference doesn't matter, because if you pass a
pointer to a libc sigset_t to the kernel it just acts on the first
8 bytes of it, but for the ucontext_t* argument to a signal handler
it trips us up. The kernel allocates this ucontext_t on the stack
according to its idea of the sigset_t type, but the type of the
ucontext_t defined by the libc headers uses the libc type, and
so do the manipulator functions like sigfillset(). This means that
(1) sizeof(uc->uc_sigmask) is much larger than the actual
space used on the stack
(2) sigfillset(&uc->uc_sigmask) will write garbage 0xff bytes
off the end of the structure, which can trash data that
was on the stack before the signal handler was invoked,
and may result in a crash after the handler returns
To avoid this, we use a memset() of the correct size to fill
the signal mask rather than using the libc function.
This fixes a problem where we would crash at least some of the
time on an i386 host when a signal was taken.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
host_to_target_siginfo() is implemented by a combination of
host_to_target_siginfo_noswap() followed by tswap_siginfo().
The first of these two functions assumes that the target_siginfo_t
it is writing to is correctly aligned, but the pointer passed
into host_to_target_siginfo() is directly from the guest and
might be misaligned. Use a local variable to avoid this problem.
(tswap_siginfo() does now correctly handle a misaligned destination.)
We have to add a memset() to host_to_target_siginfo_noswap()
to avoid some false positive "may be used uninitialized" warnings
from gcc about subfields of the _sifields union if it chooses to
inline both tswap_siginfo() and host_to_target_siginfo_noswap()
into host_to_target_siginfo().
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Peter Maydell <riku.voipio@linaro.org>
Use Coccinelle script to replace 'ret = E; return ret' with
'return E'. The script will do the substitution only when the
function return type and variable type are the same.
Manual fixups:
* audio/audio.c: coding style of "read (...)" and "write (...)"
* block/qcow2-cluster.c: wrap line to make it shorter
* block/qcow2-refcount.c: change indentation of wrapped line
* target-tricore/op_helper.c: fix coding style of
"remainder|quotient"
* target-mips/dsp_helper.c: reverted changes because I don't
want to argue about checkpatch.pl
* ui/qemu-pixman.c: fix line indentation
* block/rbd.c: restore blank line between declarations and
statements
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1465855078-19435-4-git-send-email-ehabkost@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Unused Coccinelle rule name dropped along with a redundant comment;
whitespace touched up in block/qcow2-cluster.c; stale commit message
paragraph deleted]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
The siginfo_t struct includes a union. The correct way to identify
which fields of the union are relevant is complicated, because we
have to use a combination of the si_code and si_signo to figure out
which of the union's members are valid. (Within the host kernel it
is always possible to tell, but the kernel carefully avoids giving
userspace the high 16 bits of si_code, so we don't have the
information to do this the easy way...) We therefore make our best
guess, bearing in mind that a guest can spoof most of the si_codes
via rt_sigqueueinfo() if it likes. Once we have made our guess, we
record it in the top 16 bits of the si_code, so that tswap_siginfo()
later can use it. tswap_siginfo() then strips these top bits out
before writing si_code to the guest (sign-extending the lower bits).
This fixes a bug where fields were sometimes wrong; in particular
the LTP kill10 test went into an infinite loop because its signal
handler got a si_pid value of 0 rather than the pid of the sending
process.
As part of this change, we switch to using __put_user() in the
tswap_siginfo code which writes out the byteswapped values to
the target memory, in case the target memory pointer is not
sufficiently aligned for the host CPU's requirements.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Block signals while emulating sigaction. This is a non-interruptible
syscall, and using block_signals() avoids races where the host
signal handler is invoked and tries to examine the signal handler
data structures while we are updating them.
Signed-off-by: Timothy Edward Baldwin <T.E.Baldwin99@members.leeds.ac.uk>
Message-id: 1441497448-32489-29-git-send-email-T.E.Baldwin99@members.leeds.ac.uk
[PMM: expanded commit message]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
If a synchronous signal and an asynchronous signal arrive near simultaneously,
and the signal number of the asynchronous signal is lower than that of the
synchronous signal the the handler for the asynchronous would be called first,
and then the handler for the synchronous signal would be called within or
after the first handler with an incorrect context.
This is fixed by queuing synchronous signals separately. Note that this does
risk delaying a asynchronous signal until the synchronous signal handler
returns rather than handling the signal on another thread, but this seems
unlikely to cause problems for real guest programs and is unavoidable unless
we could guarantee to roll back and reexecute whatever guest instruction
caused the synchronous signal (which would be a bit odd if we've already
logged its execution, for instance, and would require careful analysis of
all guest CPUs to check it was possible in all cases).
Signed-off-by: Timothy Edward Baldwin <T.E.Baldwin99@members.leeds.ac.uk>
Message-id: 1441497448-32489-24-git-send-email-T.E.Baldwin99@members.leeds.ac.uk
[PMM: added a comment]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
As host signals are now blocked whenever guest signals are blocked, the
queue of realtime signals is now in Linux. The QEMU queue is now
redundant and can be removed. (We already did not queue non-RT signals, and
none of the calls to queue_signal() except the one in host_signal_handler()
pass an RT signal number.)
Signed-off-by: Timothy Edward Baldwin <T.E.Baldwin99@members.leeds.ac.uk>
Message-id: 1441497448-32489-23-git-send-email-T.E.Baldwin99@members.leeds.ac.uk
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
[PMM: minor commit message tweak]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Both queue_signal() and process_pending_signals() did check for default
actions of signals, this is redundant and also causes fatal and stopping
signals to incorrectly cause guest system calls to be interrupted.
The code in queue_signal() is removed.
Signed-off-by: Timothy Edward Baldwin <T.E.Baldwin99@members.leeds.ac.uk>
Message-id: 1441497448-32489-21-git-send-email-T.E.Baldwin99@members.leeds.ac.uk
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
If multiple host signals are received in quick succession they would
be queued in TaskState then delivered to the guest in spite of
signals being supposed to be blocked by the guest signal handler's
sa_mask. Fix this by decoupling the guest signal mask from the
host signal mask, so we can have protected sections where all
host signals are blocked. In particular we block signals from
when host_signal_handler() queues a signal from the guest until
process_pending_signals() has unqueued it. We also block signals
while we are manipulating the guest signal mask in emulation of
sigprocmask and similar syscalls.
Blocking host signals also ensures the correct behaviour with respect
to multiple threads and the overrun count of timer related signals.
Alas blocking and queuing in qemu is still needed because of virtual
processor exceptions, SIGSEGV and SIGBUS.
Blocking signals inside process_pending_signals() protects against
concurrency problems that would otherwise happen if host_signal_handler()
ran and accessed the signal data structures while process_pending_signals()
was manipulating them.
Since we now track the guest signal mask separately from that
of the host, the sigsuspend system calls must track the signal
mask passed to them, because when we process signals as we leave
the sigsuspend the guest signal mask in force is that passed to
sigsuspend.
Signed-off-by: Timothy Edward Baldwin <T.E.Baldwin99@members.leeds.ac.uk>
Message-id: 1441497448-32489-19-git-send-email-T.E.Baldwin99@members.leeds.ac.uk
[PMM: make signal_pending a simple flag rather than a word with two flag bits;
ensure we don't call block_signals() twice in sigreturn codepaths;
document and assert() the guarantee that using do_sigprocmask() to
get the current mask never fails; use the qemu atomics.h functions
rather than raw volatile variable access; add extra commentary and
documentation; block SIGSEGV/SIGBUS in block_signals() and in
process_pending_signals() because they can't occur synchronously here;
check the right do_sigprocmask() call for errors in ssetmask syscall;
expand commit message; fixed sigsuspend() hanging]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
All the architecture specific handlers for sigreturn include calls
to do_sigprocmask(SIGSETMASK, &set, NULL) to set the signal mask
from the uc_sigmask in the context being restored. Factor these
out into calls to a set_sigmask() function. The next patch will
want to add code which is not run when setting the signal mask
via do_sigreturn, and this change allows us to separate the two
cases.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
