OSDN Git Service

qmiga/qemu.git
22 months agolinux-user: Combine do_futex and do_futex_time64
Richard Henderson [Mon, 29 Aug 2022 02:09:59 +0000 (19:09 -0700)]
linux-user: Combine do_futex and do_futex_time64

Pass a boolean to select between time32 and time64.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220829021006.67305-2-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Set ELF_BASE_PLATFORM for MIPS
Jiaxun Yang [Wed, 3 Aug 2022 10:30:09 +0000 (11:30 +0100)]
linux-user: Set ELF_BASE_PLATFORM for MIPS

Match most appropriate base platform string based on insn_flags.
Logic is aligned with aligned with set_isa() from
arch/mips/kernel/cpu-probe.c in Linux kernel.

Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220803103009.95972-3-jiaxun.yang@flygoat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Introduce stubs for ELF AT_BASE_PLATFORM
Jiaxun Yang [Wed, 3 Aug 2022 10:30:08 +0000 (11:30 +0100)]
linux-user: Introduce stubs for ELF AT_BASE_PLATFORM

AT_BASE_PLATFORM is a elf auxiliary vector pointing to a string
to pass some architecture information.
See getauxval(3) man-page.

Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220803103009.95972-2-jiaxun.yang@flygoat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/s390x: Save/restore fpc when handling a signal
Ilya Leoshkevich [Wed, 17 Aug 2022 12:39:02 +0000 (14:39 +0200)]
linux-user/s390x: Save/restore fpc when handling a signal

Linux kernel does this in fpregs_store() and fpregs_load(), so
qemu-user should do this as well.

Found by running valgrind's none/tests/s390x/test_sig.

Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220817123902.585623-1-iii@linux.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Don't assume 0 is not a valid host timer_t value
Peter Maydell [Mon, 25 Jul 2022 11:00:35 +0000 (12:00 +0100)]
linux-user: Don't assume 0 is not a valid host timer_t value

For handling guest POSIX timers, we currently use an array
g_posix_timers[], whose entries are a host timer_t value, or 0 for
"this slot is unused".  When the guest calls the timer_create syscall
we look through the array for a slot containing 0, and use that for
the new timer.

This scheme assumes that host timer_t values can never be zero.  This
is unfortunately not a valid assumption -- for some host libc
versions, timer_t values are simply indexes starting at 0.  When
using this kind of host libc, the effect is that the first and second
timers end up sharing a slot, and so when the guest tries to operate
on the first timer it changes the second timer instead.

Rework the timer allocation code, so that:
 * the 'slot in use' indication uses a separate array from the
   host timer_t array
 * we grab the free slot atomically, to avoid races when multiple
   threads call timer_create simultaneously
 * releasing an allocated slot is abstracted out into a new
   free_host_timer_slot() function called in the correct places

This fixes:
 * problems on hosts where timer_t 0 is valid
 * the FIXME in next_free_host_timer() about locking
 * bugs in the error paths in timer_create where we forgot to release
   the slot we grabbed, or forgot to free the host timer

Reported-by: Jon Alduan <jon.alduan@gmail.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20220725110035.1273441-1-peter.maydell@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: fix bug about missing signum convert of sigqueue
fanwenjie [Wed, 31 Aug 2022 03:55:25 +0000 (11:55 +0800)]
linux-user: fix bug about missing signum convert of sigqueue

Fixes: 66fb9763af ("basic signal handling")
Fixes: cf8b8bfc50 ("linux-user: add support for rt_tgsigqueueinfo() system call")
Signed-off-by: fanwenjie <fanwj@mail.ustc.edu.cn>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/hppa: Fix setup_sigcontext()
Helge Deller [Sat, 24 Sep 2022 11:45:01 +0000 (13:45 +0200)]
linux-user/hppa: Fix setup_sigcontext()

We don't emulate a preemptive kernel on this level, and the hppa architecture
doesn't allow context switches on the gateway page. So we always have to return
to sc_iaoq[] and not to gr[31].
This fixes the remaining random segfaults which still occured.

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20220924114501.21767-8-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/hppa: Allow PROT_GROWSUP and PROT_GROWSDOWN in mprotect()
Helge Deller [Sat, 24 Sep 2022 11:45:00 +0000 (13:45 +0200)]
linux-user/hppa: Allow PROT_GROWSUP and PROT_GROWSDOWN in mprotect()

The hppa platform uses an upwards-growing stack and required in Linux
kernels < 5.18 an executable stack for signal processing.  For that some
executables and libraries are marked to have an executable stack, for
which glibc uses the mprotect() syscall to mark the stack like this:
 mprotect(xfa000000,4096,PROT_EXEC|PROT_READ|PROT_WRITE|PROT_GROWSUP).

Currently qemu will return -TARGET_EINVAL for this syscall because of the
checks in validate_prot_to_pageflags(), which doesn't allow the
PROT_GROWSUP or PROT_GROWSDOWN flags and thus triggers this error in the
guest:
 error while loading shared libraries: libc.so.6: cannot enable executable stack as shared object requires: Invalid argument

Allow mprotect() to handle both flags and thus fix the guest.
The glibc tst-execstack testcase can be used to reproduce the issue.

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20220924114501.21767-7-deller@gmx.de>
[lvivier: s/elif TARGET_HPPA/elif defined(TARGET_HPPA)/]
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/hppa: Increase guest stack size to 80MB for hppa target
Helge Deller [Sat, 24 Sep 2022 11:44:59 +0000 (13:44 +0200)]
linux-user/hppa: Increase guest stack size to 80MB for hppa target

The hppa target requires a much bigger stack than many other targets,
and the Linux kernel allocates 80 MB by default for it.

This patch increases the guest stack for hppa to 80MB, and prevents
that this default stack size gets reduced by a lower stack limit on the
host.

Since the stack grows upwards on hppa, the stack_limit value marks the
upper boundary of the stack. Fix the output of /proc/self/maps (in the
guest) to show the [stack] marker on the correct memory area.

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20220924114501.21767-6-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/hppa: Drop stack guard page on hppa target
Helge Deller [Sat, 24 Sep 2022 11:44:58 +0000 (13:44 +0200)]
linux-user/hppa: Drop stack guard page on hppa target

The stack-overflow check when building the "grep" debian package fails
on the debian hppa target. Reason is, that the guard page at the top
of the stack (which is added by qemu) prevents the fault handler in the
grep program to properly detect the stack overflow.

The Linux kernel on a physical machine doesn't install a guard page
either, so drop it and as such fix the build of "grep".

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20220924114501.21767-5-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/hppa: Add signal trampoline for hppa target
Helge Deller [Sat, 24 Sep 2022 11:44:57 +0000 (13:44 +0200)]
linux-user/hppa: Add signal trampoline for hppa target

In Linux kernel v5.18 the vDSO for signal trampoline was added.
This code mimiks the bare minimum of this vDSO and thus avoids that the
parisc emulation needs executable stacks.

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20220924114501.21767-4-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Add proper strace format strings for getdents()/getdents64()
Helge Deller [Sat, 24 Sep 2022 11:44:56 +0000 (13:44 +0200)]
linux-user: Add proper strace format strings for getdents()/getdents64()

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20220924114501.21767-3-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Fix TARGET_PROT_SEM for XTENSA
Helge Deller [Sat, 24 Sep 2022 11:44:55 +0000 (13:44 +0200)]
linux-user: Fix TARGET_PROT_SEM for XTENSA

The xtensa platform has a value of 0x10 for PROT_SEM.

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220924114501.21767-2-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agotests/tcg/linux-test: Add linux-madvise test
Ilya Leoshkevich [Tue, 6 Sep 2022 00:08:39 +0000 (02:08 +0200)]
tests/tcg/linux-test: Add linux-madvise test

Add a test that checks madvise(MADV_DONTNEED) behavior with anonymous
and file mappings in order to prevent regressions.

Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220906000839.1672934-6-iii@linux.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Passthrough MADV_DONTNEED for certain file mappings
Ilya Leoshkevich [Tue, 6 Sep 2022 00:08:38 +0000 (02:08 +0200)]
linux-user: Passthrough MADV_DONTNEED for certain file mappings

This is a follow-up for commit 892a4f6a750a ("linux-user: Add partial
support for MADV_DONTNEED"), which added passthrough for anonymous
mappings. File mappings can be handled in a similar manner.

In order to do that, mark pages, for which mmap() was passed through,
with PAGE_PASSTHROUGH, and then allow madvise() passthrough for these
pages. Drop the explicit PAGE_ANON check, since anonymous mappings are
expected to have PAGE_PASSTHROUGH anyway.

Add PAGE_PASSTHROUGH to PAGE_STICKY in order to keep it on mprotect().

Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220725125043.43048-1-iii@linux.ibm.com>
Message-Id: <20220906000839.1672934-5-iii@linux.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Implement stracing madvise()
Ilya Leoshkevich [Tue, 6 Sep 2022 00:08:37 +0000 (02:08 +0200)]
linux-user: Implement stracing madvise()

The default implementation has several problems: the first argument is
not displayed as a pointer, making it harder to grep; the third
argument is not symbolized; and there are several extra unused
arguments.

Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220906000839.1672934-4-iii@linux.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Fix madvise(MADV_DONTNEED) on alpha
Ilya Leoshkevich [Tue, 6 Sep 2022 00:08:36 +0000 (02:08 +0200)]
linux-user: Fix madvise(MADV_DONTNEED) on alpha

MADV_DONTNEED has a different value on alpha, compared to all the other
architectures. Fix by using TARGET_MADV_DONTNEED instead of
MADV_DONTNEED.

Fixes: 892a4f6a750a ("linux-user: Add partial support for MADV_DONTNEED")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220906000839.1672934-3-iii@linux.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Provide MADV_* definitions
Ilya Leoshkevich [Tue, 6 Sep 2022 00:08:35 +0000 (02:08 +0200)]
linux-user: Provide MADV_* definitions

Provide MADV_* definitions using target_mman.h header, similar to what
kernel does. Most architectures use the same values, with the exception
of alpha and hppa.

Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220906000839.1672934-2-iii@linux.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Show timespec on strace for futex()
Helge Deller [Sun, 18 Sep 2022 19:45:53 +0000 (21:45 +0200)]
linux-user: Show timespec on strace for futex()

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220918194555.83535-11-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Add strace for clock_nanosleep()
Helge Deller [Sun, 18 Sep 2022 19:45:52 +0000 (21:45 +0200)]
linux-user: Add strace for clock_nanosleep()

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220918194555.83535-10-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/hppa: Set TASK_UNMAPPED_BASE to 0xfa000000 for hppa arch
Helge Deller [Sun, 18 Sep 2022 19:45:51 +0000 (21:45 +0200)]
linux-user/hppa: Set TASK_UNMAPPED_BASE to 0xfa000000 for hppa arch

On the parisc architecture the stack grows upwards.
Move the TASK_UNMAPPED_BASE to high memory area as it's done by the
kernel on physical machines.

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20220918194555.83535-9-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Fix strace of chmod() if mode == 0
Helge Deller [Sun, 18 Sep 2022 19:45:50 +0000 (21:45 +0200)]
linux-user: Fix strace of chmod() if mode == 0

If the mode parameter of chmod() is zero, this value isn't shown
when stracing a program:
    chmod("filename",)
This patch fixes it up to show the zero-value as well:
    chmod("filename",000)

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220918194555.83535-8-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/hppa: Dump IIR on register dump
Helge Deller [Sun, 18 Sep 2022 19:45:49 +0000 (21:45 +0200)]
linux-user/hppa: Dump IIR on register dump

Include the IIR register (which holds the opcode of the failing
instruction) when dumping the hppa registers.

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220918194555.83535-7-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user/hppa: Use EXCP_DUMP() to show enhanced debug info
Helge Deller [Sun, 18 Sep 2022 19:45:48 +0000 (21:45 +0200)]
linux-user/hppa: Use EXCP_DUMP() to show enhanced debug info

Enhance the hppa linux-user cpu_loop() to show more debugging info
on hard errors.

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220918194555.83535-6-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Log failing executable in EXCP_DUMP()
Helge Deller [Sun, 18 Sep 2022 19:45:47 +0000 (21:45 +0200)]
linux-user: Log failing executable in EXCP_DUMP()

Enhance the EXCP_DUMP() macro to print out the failing program too.
During debugging it's sometimes hard to track down the actual failing
program if you are e.g. building a whole debian package.

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220918194555.83535-5-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Add pidfd_open(), pidfd_send_signal() and pidfd_getfd() syscalls
Helge Deller [Sun, 18 Sep 2022 19:45:46 +0000 (21:45 +0200)]
linux-user: Add pidfd_open(), pidfd_send_signal() and pidfd_getfd() syscalls

I noticed those were missing when running the glib2.0 testsuite.
Add the syscalls including the strace output.

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220918194555.83535-4-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Add missing clock_gettime64() syscall strace
Helge Deller [Sun, 18 Sep 2022 19:45:45 +0000 (21:45 +0200)]
linux-user: Add missing clock_gettime64() syscall strace

Allow linux-user to strace the clock_gettime64() syscall.
This syscall is used a lot on 32-bit guest architectures which use newer
glibc versions.

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220918194555.83535-3-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: Add missing signals in strace output
Helge Deller [Sun, 18 Sep 2022 19:45:44 +0000 (21:45 +0200)]
linux-user: Add missing signals in strace output

Some of the guest signal numbers are currently not converted to
their representative names in the strace output, e.g. SIGVTALRM.

This patch introduces a smart way to generate and keep in sync the
host-to-guest and guest-to-host signal conversion tables for usage in
the qemu signal and strace code. This ensures that any signals
will now show up in both tables.

There is no functional change in this patch - with the exception that yet
missing signal names now show up in the strace code too.

Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220918194555.83535-2-deller@gmx.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: fix readlinkat handling with magic exe symlink
Jameson Nash [Mon, 8 Aug 2022 19:07:27 +0000 (15:07 -0400)]
linux-user: fix readlinkat handling with magic exe symlink

Exactly the same as f17f4989fa193fa8279474c5462289a3cfe69aea before was
for readlink. I suppose this was simply missed at the time.

Signed-off-by: Jameson Nash <vtjnash@gmail.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220808190727.875155-1-vtjnash@gmail.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agolinux-user: use 'max' instead of 'qemu32' / 'qemu64' by default
Daniel P. Berrangé [Fri, 23 Sep 2022 11:04:13 +0000 (12:04 +0100)]
linux-user: use 'max' instead of 'qemu32' / 'qemu64' by default

The 'qemu64' CPU model implements the least featureful x86_64 CPU that's
possible. Historically this hasn't been an issue since it was rare for
OS distros to build with a higher mandatory CPU baseline.

With RHEL-9, however, the entire distro is built for the x86_64-v2 ABI
baseline:

  https://developers.redhat.com/blog/2021/01/05/building-red-hat-enterprise-linux-9-for-the-x86-64-v2-microarchitecture-level

It is likely that other distros may take similar steps in the not too
distant future. For example, it has been suggested for Fedora on a
number of occasions.

This new baseline is not compatible with the qemu64 CPU model though.
While it is possible to pass a '-cpu xxx' flag to qemu-x86_64, the
usage of QEMU doesn't always allow for this. For example, the args
are typically controlled via binfmt rules that the user has no ability
to change. This impacts users who are trying to use podman on aarch64
platforms, to run containers with x86_64 content. There's no arg to
podman that can be used to change the qemu-x86_64 args, and a non-root
user of podman can not change binfmt rules without elevating privileges:

  https://github.com/containers/podman/issues/15456#issuecomment-1228210973

Changing to the 'max' CPU model gives 'qemu-x86_64' maximum
compatibility with binaries it is likely to encounter in the wild,
and not likely to have a significant downside for existing usage.

Most other architectures already use an 'any' CPU model, which is
often mapped to 'max' (or similar) already, rather than the oldest
possible CPU model.

For the sake of consistency the 'i386' architecture is also changed
from using 'qemu32' to 'max'.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220923110413.70593-1-berrange@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agoMerge tag 'edgar/xilinx-next-2022-09-21.for-upstream' of https://github.com/edgarigl...
Stefan Hajnoczi [Thu, 22 Sep 2022 17:24:27 +0000 (13:24 -0400)]
Merge tag 'edgar/xilinx-next-2022-09-21.for-upstream' of https://github.com/edgarigl/qemu into staging

Xilinx queue

# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCgAdFiEErET+3BT38evtv0FRKcWWeA9ryoMFAmMraeQACgkQKcWWeA9r
# yoOUUwgAz/GPBURkdzYcFbOa4VEBrCaJxCu2WvyBdj6kjEgw/ZmdXbDHfeKTr2Rl
# 8mo7UWugfj1KTMqo5Tmte6+3EeNcuB9qA/flKj7wfKwLX1S0V6EFNGJn8sdNe+1r
# c6fYm0bUBk9FKauUT4t3U1Zl9SGJ2Pwa8JvztVgBvvZG6G+J0bkVmw2mMXb3jaaV
# skzmBuvDT0G8JkxWnpEXwznrPSHSjivqawVlBfjC5JmCjZ16bz05okju+cckfzTJ
# 0VF7/27MWfVCpMtWjRSA6p+3gwLziz7xOhPe9excSVsDyAs4WdcuI+/42/MBBUIn
# rlg7pNIslI9r8eP5Jyk6LcSmBsPDMA==
# =GZsl
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 21 Sep 2022 15:45:40 EDT
# gpg:                using RSA key AC44FEDC14F7F1EBEDBF415129C596780F6BCA83
# gpg: Good signature from "Edgar E. Iglesias (Xilinx key) <edgar.iglesias@xilinx.com>" [unknown]
# gpg:                 aka "Edgar E. Iglesias <edgar.iglesias@gmail.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: AC44 FEDC 14F7 F1EB EDBF  4151 29C5 9678 0F6B CA83

* tag 'edgar/xilinx-next-2022-09-21.for-upstream' of https://github.com/edgarigl/qemu:
  hw/microblaze: pass random seed to fdt

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22 months agoMerge tag 'pull-hex-20220919' of https://github.com/quic/qemu into staging
Stefan Hajnoczi [Thu, 22 Sep 2022 17:21:50 +0000 (13:21 -0400)]
Merge tag 'pull-hex-20220919' of https://github.com/quic/qemu into staging

Hexagon update
    remove unused encodings
    add fmin/fmax tests for signed zero

# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCgAdFiEENjXHiM5iuR/UxZq0ewJE+xLeRCIFAmMou7IACgkQewJE+xLe
# RCIYbQgAgjFujecgbbCJfBPVMmpTXNOgk+Jt3w+jfg7/WJRZuhxAU3xB2qpismUH
# 5MntMlFHAGOjlPXfg6U5AZFSw3RhlanH/RChHpVKuL6peOXFImIfEqdyVXHXfCuu
# FlpQFGwJ3Rs50UJhd7lVdlx0I7lup4E4X77hFvFcZQP6aNrt6Ic1Zq5eXhEq9k2A
# NnXol1R416JRT/senujYVvcTpgYVHlQCS+4dJEzKUqvFlTdo7lnAbPdjO8MPrz7B
# 0NgPUGjGZJ70Dcqvd1n8HePIU1YyKTlHJNaWyTlAmw4MECyHyAJnd64jEMNECDb5
# 0BrpHcY1HCt1Rh4QratemTfJglAJlA==
# =UUyr
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 19 Sep 2022 14:57:54 EDT
# gpg:                using RSA key 3635C788CE62B91FD4C59AB47B0244FB12DE4422
# gpg: Good signature from "Taylor Simpson (Rock on) <tsimpson@quicinc.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 3635 C788 CE62 B91F D4C5  9AB4 7B02 44FB 12DE 4422

* tag 'pull-hex-20220919' of https://github.com/quic/qemu:
  Hexagon (tests/tcg/hexagon): add fmin/fmax tests for signed zero
  Hexagon (target/hexagon) remove unused encodings

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22 months agohw/microblaze: pass random seed to fdt
Jason A. Donenfeld [Thu, 8 Sep 2022 09:40:30 +0000 (11:40 +0200)]
hw/microblaze: pass random seed to fdt

If the FDT contains /chosen/rng-seed, then the Linux RNG will use it to
initialize early. Set this using the usual guest random number
generation function. This FDT node is part of the DT specification.

Reviewed-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
22 months agoMerge tag 'm68k-for-7.2-pull-request' of https://github.com/vivier/qemu-m68k into...
Stefan Hajnoczi [Wed, 21 Sep 2022 17:12:36 +0000 (13:12 -0400)]
Merge tag 'm68k-for-7.2-pull-request' of https://github.com/vivier/qemu-m68k into staging

m68k pull request 20220921

- several fixes for SR
- implement TAS
- feature cleanup

# -----BEGIN PGP SIGNATURE-----
#
# iQJGBAABCAAwFiEEzS913cjjpNwuT1Fz8ww4vT8vvjwFAmMrMx0SHGxhdXJlbnRA
# dml2aWVyLmV1AAoJEPMMOL0/L748UB0P/1w5w+ogpcWVp9uBPE9m6lTT8sTricWD
# oGMIEG0kgpS3xTp7pZ/WeCp38IShFBfBcz5aypvR5nS/1aclyJnzGsCqyWdBe9c4
# jJJY5r7JKP2aKcolGNilNyd20ldOiaxZGe6yFLJDWi2spFbRx3iiRJ4/2NXF5Xi8
# TlHKdlDGlNLFiFNtBXMMwqvL77qJ8LH/aE4cAr8JTOb1VszKXrFxqEqxoUucRirB
# u0LbM+DP3u2xXjTGLMLlMcKf9X2BXwuWBSAXslB8xWmRlX+B6fMudBFglTgbu0Cc
# bpoBBqY4s3QYPb21i89osYevJAJSdrtEzkKus3xAI08ACSffb9k9m/naVJJDSSNC
# HZeKVbAd7I0Xw2xzzO5yQB+7rdfgoL1miE1rs936WKHi0WWHZpdJqzl/3G3ZPhRz
# NmqczF9rRR9B9SabXx5lWlhK+Ys/W7PzY+R4gc6ose0wF4T70qmVF3EoioP1c5Y8
# 6OonMpRu6L5sW5KM3IUmkBo3KcnLezlxtebfaDyaKC9tB0qg4aM14ikL36nsLFbh
# 2nGExYSyMJ6U4tqpxyQxijMMSQG20vyVIup6cUsrSD+rGmbSuWZWJwsTmaAw2W6k
# 6HtDgtFk40ZB1WttYupQBa/LgjshGLl28jyLI9nNEdFYb4H1JAallEERF/tb6AUD
# WEiu8vcUMYEp
# =IAMt
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 21 Sep 2022 11:51:57 EDT
# gpg:                using RSA key CD2F75DDC8E3A4DC2E4F5173F30C38BD3F2FBE3C
# gpg:                issuer "laurent@vivier.eu"
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>" [full]
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>" [full]
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>" [full]
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* tag 'm68k-for-7.2-pull-request' of https://github.com/vivier/qemu-m68k:
  target/m68k: always call gen_exit_tb() after writes to SR
  target/m68k: rename M68K_FEATURE_M68000 to M68K_FEATURE_M68K
  target/m68k: Perform writback before modifying SR
  target/m68k: Fix MACSR to CCR
  target/m68k: Implement atomic test-and-set

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22 months agoMerge tag 'pull-ppc-20220920' of https://gitlab.com/danielhb/qemu into staging
Stefan Hajnoczi [Wed, 21 Sep 2022 17:11:57 +0000 (13:11 -0400)]
Merge tag 'pull-ppc-20220920' of https://gitlab.com/danielhb/qemu into staging

ppc patch queue for 2022-09-20:

This queue contains a implementation of PowerISA 3.1B hash insns, ppc
TCG insns cleanups and fixes, and miscellaneus fixes in the spapr and
pnv_phb models.

# -----BEGIN PGP SIGNATURE-----
#
# iHUEABYKAB0WIQQX6/+ZI9AYAK8oOBk82cqW3gMxZAUCYyoWlAAKCRA82cqW3gMx
# ZDYhAP0eQMeA4NS3hiw7WMcAVg0pei3ZJL9oEh1UE3+MfK7MhQEA0q8qExWnQJAA
# a0hfnFH9pLjI+v0f/FbFK6QJBpu/bg8=
# =qT+H
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 20 Sep 2022 15:37:56 EDT
# gpg:                using EDDSA key 17EBFF9923D01800AF2838193CD9CA96DE033164
# gpg: Good signature from "Daniel Henrique Barboza <danielhb413@gmail.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 17EB FF99 23D0 1800 AF28  3819 3CD9 CA96 DE03 3164

* tag 'pull-ppc-20220920' of https://gitlab.com/danielhb/qemu:
  hw/ppc/spapr: Fix code style problems reported by checkpatch
  hw/pci-host: pnv_phb{3, 4}: Fix heap out-of-bound access failure
  hw/ppc: spapr: Use qemu_vfree() to free spapr->htab
  target/ppc: Clear fpstatus flags on helpers missing it
  target/ppc: Zero second doubleword of VSR registers for FPR insns
  target/ppc: Set OV32 when OV is set
  target/ppc: Zero second doubleword for VSX madd instructions
  target/ppc: Set result to QNaN for DENBCD when VXCVI occurs
  target/ppc: Zero second doubleword in DFP instructions
  target/ppc: Remove unused xer_* macros
  target/ppc: Remove extra space from s128 field in ppc_vsr_t
  target/ppc: Merge fsqrt and fsqrts helpers
  target/ppc: Move fsqrts to decodetree
  target/ppc: Move fsqrt to decodetree
  target/ppc: Implement hashstp and hashchkp
  target/ppc: Implement hashst and hashchk
  target/ppc: Add HASHKEYR and HASHPKEYR SPRs

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22 months agoMerge tag 'pull-testing-next-200922-2' of https://github.com/stsquad/qemu into staging
Stefan Hajnoczi [Wed, 21 Sep 2022 17:10:51 +0000 (13:10 -0400)]
Merge tag 'pull-testing-next-200922-2' of https://github.com/stsquad/qemu into staging

Testing and CI changes:

  - reduce number of targets for cross_user_build
  - update avocado xlnx_versal test with new binaries
  - add explicit timeouts to a number of avocado TCG tests
  - reduce default timeout to 120s
  - update lcitool to support cross-amd64
  - flatten a number of docker cross containers
  - clean up stale qemu/debian10 dependencies
  - remove obsolete Fedora VM test
  - add configure workaround for meson --disable-pie bug
  - disable --static-pie for aarch64 gitlab runner
  - update aarch32/aarch64 jobs to 22.04
  - deprecate 32 bit big-endian MIPS as a host
  - remove FROM qemu/ support from docker.py
  - remove Debian base images now everything is flat

# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCgAdFiEEZoWumedRZ7yvyN81+9DbCVqeKkQFAmMp8Q8ACgkQ+9DbCVqe
# KkQmlwf/awT+jOmAW7TjlQnUTgHJ2hyOo7EViY/nmRkPOCT3ZG32pWFHBorHPX7s
# BeqZzpzCvhzaIfObnjIssx13C5QId5XjJGuTgMAnSsGhzTrp7VUJc1/bBfHcD9L2
# dJJduG+bfAkh95heBkry5EhFt2ZMui5yv9DjEH44hUUc9nwKtIQGts3H3fnVqzvv
# rzLZ7c2lhdLpAxHjmjSiiD8H59lJ+DpoziaobW4D7teGgecnyGVvJ9m1YH4Rc+kM
# gpLTOGMhADkQlysf5e5cvxXSJbP7YpXYrsr9X+DfEy5PMt2L3y4Yv0wiAz9ClYvm
# obD4wMQS5echYvb77qS1G8A0VMEPqA==
# =3oYu
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 20 Sep 2022 12:57:51 EDT
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* tag 'pull-testing-next-200922-2' of https://github.com/stsquad/qemu: (30 commits)
  tests/docker: remove the Debian base images
  tests/docker: remove FROM qemu/ support from docker.py
  tests/docker: update and flatten debian-toolchain
  tests/docker: update and flatten debian-hexagon-cross
  tests/docker: update and flatten debian-loongarch-cross
  tests/docker: update and flatten debian-amd64-cross
  tests/lcitool: bump to latest version
  tests/docker: update and flatten debian-all-test-cross
  tests/docker: flatten debian-riscv64-test-cross
  Deprecate 32 bit big-endian MIPS
  gitlab-ci: update aarch32/aarch64 custom runner jobs
  gitlab-ci/custom-runners: Disable -static-pie for ubuntu-20.04-aarch64
  configure: explicitly set cflags for --disable-pie
  tests/vm: Remove obsolete Fedora VM test
  tests/docker: remove amd64 qemu/debian10 dependency
  tests/docker: remove tricore qemu/debian10 dependency
  tests/docker: flatten debian-powerpc-test-cross
  tests/docker: update and flatten debian-sparc64-cross
  tests/docker: update and flatten debian-sh4-cross
  tests/docker: update and flatten debian-mips64-cross
  ...

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22 months agotarget/m68k: always call gen_exit_tb() after writes to SR
Mark Cave-Ayland [Sat, 17 Sep 2022 11:25:15 +0000 (12:25 +0100)]
target/m68k: always call gen_exit_tb() after writes to SR

Any write to SR can change the security state so always call gen_exit_tb() when
this occurs. In particular MacOS makes use of andiw/oriw in a few places to
handle the switch between user and supervisor mode.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220917112515.83905-5-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agotarget/m68k: rename M68K_FEATURE_M68000 to M68K_FEATURE_M68K
Mark Cave-Ayland [Sat, 17 Sep 2022 11:25:12 +0000 (12:25 +0100)]
target/m68k: rename M68K_FEATURE_M68000 to M68K_FEATURE_M68K

The M68K_FEATURE_M68000 feature is misleading in that its name suggests the feature
is defined just for Motorola 68000 CPUs, whilst in fact it is defined for all
Motorola 680X0 CPUs.

In order to avoid confusion with the other M68K_FEATURE_M680X0 constants which
define the features available for specific Motorola CPU models, rename
M68K_FEATURE_M68000 to M68K_FEATURE_M68K and add comments to clarify its usage.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220917112515.83905-2-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agotarget/m68k: Perform writback before modifying SR
Richard Henderson [Tue, 13 Sep 2022 14:28:18 +0000 (15:28 +0100)]
target/m68k: Perform writback before modifying SR

Writes to SR may change security state, which may involve
a swap of %ssp with %usp as reflected in %a7.  Finish the
writeback of %sp@+ before swapping stack pointers.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1206
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20220913142818.7802-3-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agotarget/m68k: Fix MACSR to CCR
Richard Henderson [Tue, 13 Sep 2022 14:28:17 +0000 (15:28 +0100)]
target/m68k: Fix MACSR to CCR

First, we were writing to the entire SR register, instead
of only the flags portion.  Second, we were not clearing C
as per the documentation (X was cleared via the 0xf mask).

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220913142818.7802-2-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agotarget/m68k: Implement atomic test-and-set
Richard Henderson [Mon, 29 Aug 2022 05:17:46 +0000 (22:17 -0700)]
target/m68k: Implement atomic test-and-set

This is slightly more complicated than cas,
because tas is allowed on data registers.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220829051746.227094-1-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
22 months agoMerge tag 'pull-request-2022-09-20' of https://gitlab.com/thuth/qemu into staging
Stefan Hajnoczi [Tue, 20 Sep 2022 20:24:07 +0000 (16:24 -0400)]
Merge tag 'pull-request-2022-09-20' of https://gitlab.com/thuth/qemu into staging

* Skip tests if the corresponding feature is missing
* Update NetBSD VM test to version 9.3
* Update the FreeBSD CI to version 13.1
* Some small fixes for the qtests
* Update wordings in the QEMU guest-agent

# -----BEGIN PGP SIGNATURE-----
#
# iQJFBAABCAAvFiEEJ7iIR+7gJQEY8+q5LtnXdP5wLbUFAmMpvqURHHRodXRoQHJl
# ZGhhdC5jb20ACgkQLtnXdP5wLbXtaQ//ap4P8Gdg7HyShflpHj3+Z+UC/THtcCAD
# F5Oj/E89wYJZJvWpjELLUf+218O+zbcKkv6z25AHo37JRSLQMzsD21RpDYcu8qnN
# Xk77KrOLFaVRRo6e3FAi5I+uWsm1h/9jJfC/sFrzfKGnWcW5pTzXUkbvWT3LZo/q
# d0wJCMn/+h0j6kqhjm8mhFCsGsSaUKFCMzm/6rsWxVsMsVnlTyYhg/IzhocKrkVD
# 9ZWz+jAv+h4F0CLl/SeRvZFq4LEGZoP3KHWrDAgEO9i1yC+tNpGDFyV/hJXwapcm
# jYGaPPgS2dLToMXf5uubfmyptircQKqdpC2Hl8UwgUS6OgwA2YRCxad8rrrdBezq
# 2Nb8KissVB8ySL1Asf00gxioJNwLKrY+4NIeDVxAQ56f1/9kd0w0ylrweR492V/p
# z3Yrs3uqqALR1jWfB7/Rh4EN2R0oe6bnlfb8bB6lzPKynXQicZLeoOnNPbkOQMw8
# HgjKVFT130ydXJKMJkgEQmi4JUGoaKqSKCmlSaHz9TAobwNiVBXaH0cyuX5QEuoX
# WgZosiF5nBzdbhkC2IPcQYHzIIkelfx9j9mKE0PNfUSkFgMiywcI5TfqabPWGeaA
# rxYPU4EZmbgCZPlBVdPFwog/BXap2EIo9+xtINM0COQDoy+Q/l5Z3keHhpjayBBS
# qWc6CRSmhLs=
# =sHK+
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 20 Sep 2022 09:22:45 EDT
# gpg:                using RSA key 27B88847EEE0250118F3EAB92ED9D774FE702DB5
# gpg:                issuer "thuth@redhat.com"
# gpg: Good signature from "Thomas Huth <th.huth@gmx.de>" [full]
# gpg:                 aka "Thomas Huth <thuth@redhat.com>" [full]
# gpg:                 aka "Thomas Huth <huth@tuxfamily.org>" [full]
# gpg:                 aka "Thomas Huth <th.huth@posteo.de>" [unknown]
# Primary key fingerprint: 27B8 8847 EEE0 2501 18F3  EAB9 2ED9 D774 FE70 2DB5

* tag 'pull-request-2022-09-20' of https://gitlab.com/thuth/qemu:
  qga: Replace 'blacklist' and 'whitelist' in the guest agent sources
  qga: Replace 'blacklist' command line and config file options by 'block-rpcs'
  gitlab-ci: Update the FreeBSD 13 job from 13.0 to 13.1
  tests: sb16 has both pc and q35 tests
  tests: Only run intel-hda-tests if machine type is compiled in
  bios-tables-test: Only run test for machine types compiled in
  bios-tables-test: Sort all x86_64 tests by machine type
  bios-tables-test: Make oem-fields tests be consistent
  meson-build: Enable CONFIG_REPLICATION only when replication is set
  tests: Fix error strings
  qtest/fuzz-lsi53c895a-test: set guest RAM to 2G
  tests/qtest: npcm7xx-emc-test: Skip checking MAC
  .gitlab-ci.d/windows.yml: Drop the sed processing in the 64-bit build
  tests/vm: update NetBSD to 9.3
  tests: mark io-command test as skipped if socat is missing

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22 months agoMerge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging
Stefan Hajnoczi [Tue, 20 Sep 2022 20:22:26 +0000 (16:22 -0400)]
Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging

* add help option for -audio and -audiodev
* another missing memory barrier for dirty pages
* target/i386: Raise #GP on unaligned m128 accesses
* coverity fixes + improvements to components
* add MMX and 3DNow! tests
* SSE4a fixes
* target/i386: TCG translation cleanups
* update qboot submodule

# -----BEGIN PGP SIGNATURE-----
#
# iQFIBAABCAAyFiEE8TM4V0tmI4mGbHaCv/vSX3jHroMFAmMocZcUHHBib256aW5p
# QHJlZGhhdC5jb20ACgkQv/vSX3jHroNV7Qf+NEoB8R0ug+ClMRe1Qqt8FXEd0eXE
# nT19q4rOWfmW4/L+wI6gpxhbxrxOuLwoZ8YvD8c6rQAdexMoHoeTvA1PAca4zZTo
# ISmW3bXsoHN2uGLPz4CKhjKBLCANtDkh3EWCwRFkLSRCLSRDhKPrG1Ue3fOgQ6GO
# riROcxbyYzvU/4uefSW+xG/Im9gftNF6occZZ59LrK7Xd8kwlb+E+EdsmzFw5f8O
# Q9irVQ8pX9ZM4BK2KiT16nZ0uSRwJqSJKbLI670nUEsj1jQCIgU3srgZHjAIvoir
# yivDs6oktgS/HkPD5CQoTX+fVDgEDM1TTF6P8r7uJopPXpzz+AHswfSJmg==
# =RVCS
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 19 Sep 2022 09:41:43 EDT
# gpg:                using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg:                issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* tag 'for-upstream' of https://gitlab.com/bonzini/qemu: (21 commits)
  qboot: update to latest submodule
  build: remove extra parentheses causing missing rebuilds
  target/i386: introduce insn_get_addr
  target/i386: REPZ and REPNZ are mutually exclusive
  target/i386: fix INSERTQ implementation
  target/i386: correctly mask SSE4a bit indices in register operands
  audio: add help option for -audio and -audiodev
  tests/tcg: remove old SSE tests
  tests/tcg: refine MMX support in SSE tests
  tests/tcg: i386: add MMX and 3DNow! tests
  tests/tcg: i386: fix typos in 3DNow! instructions
  tests: unit: add NULL-pointer check
  tests: test-qga: close socket on failure to connect
  tests: unit: simplify test-visitor-serialization list tests
  smbios: sanitize type from external type before checking have_fields_bitmap
  coverity: put NUBus under m68k component
  coverity: add new RISC-V component
  spapr_pci: fix leak in spapr_phb_vfio_get_loc_code
  kvm: fix memory leak on failure to read stats descriptors
  target/i386: Raise #GP on unaligned m128 accesses when required.
  ...

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22 months agoMerge tag 'pull-loongarch-20220920' of https://gitlab.com/gaosong/qemu into staging
Stefan Hajnoczi [Tue, 20 Sep 2022 18:17:02 +0000 (14:17 -0400)]
Merge tag 'pull-loongarch-20220920' of https://gitlab.com/gaosong/qemu into staging

v2: fix compile error.

# -----BEGIN PGP SIGNATURE-----
#
# iLMEAAEIAB0WIQS4/x2g0v3LLaCcbCxAov/yOSY+3wUCYymK3wAKCRBAov/yOSY+
# 31trA/oDy6CHaPvT9mwKOM/ScYcwCf+dEtFGkqpsjcPlEB7GLJ9moJxp1xlm+Htw
# KA0MlA4QrI6MWd9udEbnoB7aRScG/ugzWHDPGocg+qjqiZvmHpE9ShdxYeDVJaLk
# 6hN9dkom+ANWeZ7T4gmdQreK/lwjUAqkiPSjw1HhvSHr3kAeDA==
# =uUxe
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 20 Sep 2022 05:41:51 EDT
# gpg:                using RSA key B8FF1DA0D2FDCB2DA09C6C2C40A2FFF239263EDF
# gpg: Can't check signature: No public key

* tag 'pull-loongarch-20220920' of https://gitlab.com/gaosong/qemu:
  hw/loongarch: Improve acpi dsdt table
  hw/loongarch: Support memory hotplug
  hw/loongarch: Fix acpi ged irq number in dsdt table
  hw/loongarch: Add RAMFB to dynamic_sysbus_devices list
  hw/loongarch: Add hotplug handler for machine
  hw/loongarch: Add platform bus support
  hw/loongarch: Add interrupt information to FDT table
  hw/loongarch: Support fw_cfg dma function
  hw/loongarch: Remove vga device when loongarch init

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22 months agotests/docker: remove the Debian base images
Alex Bennée [Wed, 14 Sep 2022 15:59:50 +0000 (16:59 +0100)]
tests/docker: remove the Debian base images

We no longer use these in any of our images. Clean-up the remaining
comments and documentation that reference them and remove from the
build.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-31-alex.bennee@linaro.org>

22 months agotests/docker: remove FROM qemu/ support from docker.py
Alex Bennée [Wed, 14 Sep 2022 15:59:49 +0000 (16:59 +0100)]
tests/docker: remove FROM qemu/ support from docker.py

We want to migrate from docker.py to building our images directly with
docker/podman. Before we get there we need to make sure we don't
re-introduce our layered builds so bug out if we see FROM qemu/ in a
Dockerfile.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-30-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-toolchain
Alex Bennée [Wed, 14 Sep 2022 15:59:48 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-toolchain

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile as we do not need anything from the base image to
build the toolchain. This is used to build both the nios and
microblaze toolchains.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-29-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-hexagon-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:47 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-hexagon-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile as we do not some of the extraneous packages from
the base image to build the toolchain.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-28-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-loongarch-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:46 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-loongarch-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile. We really don't need the rest of the stuff from the
QEMU base image just to compile test images. In this case it is a
binary distribution of the toolchain anyway.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-27-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-amd64-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:45 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-amd64-cross

Now lcitool has support for building a x86_64 cross image we can use
it for this.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-26-alex.bennee@linaro.org>

22 months agotests/lcitool: bump to latest version
Alex Bennée [Wed, 14 Sep 2022 15:59:44 +0000 (16:59 +0100)]
tests/lcitool: bump to latest version

We need this to be able to cleanly build the x86 cross images. There
are a few minor updates triggered by lcitool-refresh including adding
"libslirp" to the freebsd vars and opensuse-leap which will help when
we finally drop the slirp submodule from QEMU.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-25-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-all-test-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:43 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-all-test-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile. We also need to ensure we install clang as it is
used for those builds as well.

It would be nice to port this to lcitool but for now this will do.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-24-alex.bennee@linaro.org>

22 months agotests/docker: flatten debian-riscv64-test-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:42 +0000 (16:59 +0100)]
tests/docker: flatten debian-riscv64-test-cross

Flatten into a single dockerfile and update to match the rest of the
test cross compile dockerfiles.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-23-alex.bennee@linaro.org>

22 months agoDeprecate 32 bit big-endian MIPS
Alex Bennée [Wed, 14 Sep 2022 15:59:41 +0000 (16:59 +0100)]
Deprecate 32 bit big-endian MIPS

It's becoming harder to maintain a cross-compiler to test this host
architecture as the old stable Debian 10 ("Buster") moved into LTS
which supports fewer architectures. For now:

  - mark it's deprecation in the docs
  - downgrade the containers to build TCG tests only
  - drop the cross builds from our CI

Users with an appropriate toolchain and user-space can still take
their chances building it.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Huacai Chen <chenhuacai@kernel.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-22-alex.bennee@linaro.org>

22 months agogitlab-ci: update aarch32/aarch64 custom runner jobs
Alex Bennée [Wed, 14 Sep 2022 15:59:40 +0000 (16:59 +0100)]
gitlab-ci: update aarch32/aarch64 custom runner jobs

The custom runner is now using 22.04 so we can drop our hacks to deal
with broken libssh and glusterfs. The provisioning scripts will be
updated in a separate commit.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-21-alex.bennee@linaro.org>

22 months agogitlab-ci/custom-runners: Disable -static-pie for ubuntu-20.04-aarch64
Richard Henderson [Wed, 14 Sep 2022 15:59:39 +0000 (16:59 +0100)]
gitlab-ci/custom-runners: Disable -static-pie for ubuntu-20.04-aarch64

The project has reached the magic size at which we see

/usr/aarch64-linux-gnu/lib/libc.a(init-first.o): in function `__libc_init_first':
(.text+0x10): relocation truncated to fit: R_AARCH64_LD64_GOTPAGE_LO15 against \
symbol `__environ' defined in .bss section in /usr/aarch64-linux-gnu/lib/libc.a(environ.o)
/usr/bin/ld: (.text+0x10): warning: too many GOT entries for -fpic, please recompile with -fPIC

The bug has been reported upstream, but in the meantime there is
nothing we can do except build a non-pie executable.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220823210329.1969895-1-richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-20-alex.bennee@linaro.org>

22 months agoconfigure: explicitly set cflags for --disable-pie
Alex Bennée [Wed, 14 Sep 2022 15:59:38 +0000 (16:59 +0100)]
configure: explicitly set cflags for --disable-pie

This is working around current limitation of Meson's handling of
--disable-pie.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-19-alex.bennee@linaro.org>

22 months agohw/ppc/spapr: Fix code style problems reported by checkpatch
Bernhard Beschow [Mon, 19 Sep 2022 23:17:15 +0000 (01:17 +0200)]
hw/ppc/spapr: Fix code style problems reported by checkpatch

Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Message-Id: <20220919231720.163121-5-shentey@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agohw/pci-host: pnv_phb{3, 4}: Fix heap out-of-bound access failure
Xuzhou Cheng [Tue, 20 Sep 2022 10:31:48 +0000 (18:31 +0800)]
hw/pci-host: pnv_phb{3, 4}: Fix heap out-of-bound access failure

pnv_phb3_root_bus_info and pnv_phb4_root_bus_info are missing the
instance_size initialization. This results in accessing out-of-bound
memory when setting 'chip-id' and 'phb-id', and eventually crashes
glib's malloc functionality with the following message:

  "qemu-system-ppc64: GLib: ../glib-2.72.3/glib/gmem.c:131: failed to allocate 3232 bytes"

This issue was noticed only when running qtests with QEMU Windows
32-bit executable. Windows 64-bit, Linux 32/64-bit do not expose
this bug though.

Fixes: 9ae1329ee2fe ("ppc/pnv: Add models for POWER8 PHB3 PCIe Host bridge")
Fixes: 4f9924c4d4cf ("ppc/pnv: Add models for POWER9 PHB4 PCIe Host bridge")
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Xuzhou Cheng <xuzhou.cheng@windriver.com>
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Message-Id: <20220920103159.1865256-29-bmeng.cn@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agohw/ppc: spapr: Use qemu_vfree() to free spapr->htab
Xuzhou Cheng [Tue, 20 Sep 2022 10:31:47 +0000 (18:31 +0800)]
hw/ppc: spapr: Use qemu_vfree() to free spapr->htab

spapr->htab is allocated by qemu_memalign(), hence we should use
qemu_vfree() to free it.

Fixes: c5f54f3e31bf ("pseries: Move hash page table allocation to reset time")
Fixes: b4db54132ffe ("target/ppc: Implement H_REGISTER_PROCESS_TABLE H_CALL"")
Signed-off-by: Xuzhou Cheng <xuzhou.cheng@windriver.com>
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20220920103159.1865256-28-bmeng.cn@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotests/vm: Remove obsolete Fedora VM test
Thomas Huth [Wed, 14 Sep 2022 15:59:37 +0000 (16:59 +0100)]
tests/vm: Remove obsolete Fedora VM test

It's still based on Fedora 30 - which is not supported anymore by QEMU
since years. Seems like nobody is using (and refreshing) this, and it's
easier to test this via a container anyway, so let's remove this now.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220822175317.190551-1-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220914155950.804707-18-alex.bennee@linaro.org>

22 months agotests/docker: remove amd64 qemu/debian10 dependency
Alex Bennée [Wed, 14 Sep 2022 15:59:36 +0000 (16:59 +0100)]
tests/docker: remove amd64 qemu/debian10 dependency

We missed removing this dependency when we flattened the build.

Fixes 9e19fd7d4a (tests/docker: update debian-amd64 with lcitool)

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-17-alex.bennee@linaro.org>

22 months agotests/docker: remove tricore qemu/debian10 dependency
Alex Bennée [Wed, 14 Sep 2022 15:59:35 +0000 (16:59 +0100)]
tests/docker: remove tricore qemu/debian10 dependency

We missed removing this dependency when we flattened the build.

Fixes: 39ce923732 (gitlab: enable a very minimal build with the tricore container)
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-16-alex.bennee@linaro.org>

22 months agotests/docker: flatten debian-powerpc-test-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:34 +0000 (16:59 +0100)]
tests/docker: flatten debian-powerpc-test-cross

Flatten into a single dockerfile. We really don't need the rest of the
stuff from the QEMU base image just to compile test images.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-15-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-sparc64-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:33 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-sparc64-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile. We really don't need the rest of the stuff from
the QEMU base image just to compile test images.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-14-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-sh4-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:32 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-sh4-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile. We really don't need the rest of the stuff from
the QEMU base image just to compile test images.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-13-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-mips64-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:31 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-mips64-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile. We really don't need the rest of the stuff from
the QEMU base image just to compile test images.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-12-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-m68k-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:30 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-m68k-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile. We really don't need the rest of the stuff from
the QEMU base image just to compile test images.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-11-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-hppa-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:29 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-hppa-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile. We really don't need the rest of the stuff from
the QEMU base image just to compile test images.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-10-alex.bennee@linaro.org>

22 months agotests/docker: update and flatten debian-alpha-cross
Alex Bennée [Wed, 14 Sep 2022 15:59:28 +0000 (16:59 +0100)]
tests/docker: update and flatten debian-alpha-cross

Update to the latest stable Debian. While we are at it flatten into a
single dockerfile. We really don't need the rest of the stuff from
the QEMU base image just to compile test images.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-9-alex.bennee@linaro.org>

22 months agotests/avocado: reduce the default timeout to 120s
Alex Bennée [Wed, 14 Sep 2022 15:59:27 +0000 (16:59 +0100)]
tests/avocado: reduce the default timeout to 120s

We should be aiming to keep our tests under 2 minutes so lets reduce
the default timeout to that. Tests that we know take longer should
explicitly set a longer timeout.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-8-alex.bennee@linaro.org>

22 months agotests/avocado: split the AST2x00Machine classes
Alex Bennée [Wed, 14 Sep 2022 15:59:26 +0000 (16:59 +0100)]
tests/avocado: split the AST2x00Machine classes

The SDK tests take a lot longer to run and hence need a longer
timeout. As they run well over the 60 second maximum for CI lets also
disable them for CI as well.

I suspect they also suffer from the inability to detect the login
prompt due to no newlines being processed.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-7-alex.bennee@linaro.org>

22 months agotests/avocado: add explicit timeout for ppc64le TCG tests
Alex Bennée [Wed, 14 Sep 2022 15:59:25 +0000 (16:59 +0100)]
tests/avocado: add explicit timeout for ppc64le TCG tests

We don't want to rely on the soon to be reduced default time. These
tests are still slow for something we want to run in CI though.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-6-alex.bennee@linaro.org>

22 months agotests/avocado: add explicit timeout for s390 TCG tests
Alex Bennée [Wed, 14 Sep 2022 15:59:24 +0000 (16:59 +0100)]
tests/avocado: add explicit timeout for s390 TCG tests

We don't want to rely on the soon to be reduced default time. These
tests are still slow for something we want to run in CI though.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220914155950.804707-5-alex.bennee@linaro.org>

22 months agotests/avocado: add explicit timeout for Aarch64 TCG tests
Alex Bennée [Wed, 14 Sep 2022 15:59:23 +0000 (16:59 +0100)]
tests/avocado: add explicit timeout for Aarch64 TCG tests

We don't want to rely on the soon to be reduced default time. These
tests are still slow for something we want to run in CI though.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-4-alex.bennee@linaro.org>

22 months agotests/avocado/boot_linux_console: Fix the test_aarch64_xlnx_versal_virt test
Thomas Huth [Wed, 14 Sep 2022 15:59:22 +0000 (16:59 +0100)]
tests/avocado/boot_linux_console: Fix the test_aarch64_xlnx_versal_virt test

The assets that this test tries to download have been removed from the
server. Update to a newer version to get it working again.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20220829080940.110831-1-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220914155950.804707-3-alex.bennee@linaro.org>

22 months agogitlab: reduce targets in cross_user_build_job
Alex Bennée [Wed, 14 Sep 2022 15:59:21 +0000 (16:59 +0100)]
gitlab: reduce targets in cross_user_build_job

We already limit the scope of the cross system build to reduce the
cross build times. With the recent addition of more targets we are
also running into timeout issues for some of the cross user builds.

I've selected a few of those linux-user targets which are less likely
to be in common use as distros don't have pre-built rootfs for them.
I've also added the same CROSS_SKIP_TARGETS variable as is
occasionally used to further limit cross system builds.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220914155950.804707-2-alex.bennee@linaro.org>

22 months agotarget/ppc: Clear fpstatus flags on helpers missing it
Víctor Colombo [Tue, 6 Sep 2022 12:55:23 +0000 (09:55 -0300)]
target/ppc: Clear fpstatus flags on helpers missing it

In ppc emulation, exception flags are not cleared at the end of an
instruction. Instead, the next instruction is responsible to clear
it before its emulation. However, some helpers are not doing it,
causing an issue where the previously set exception flags are being
used and leading to incorrect values being set in FPSCR.
Fix this by clearing fp_status before doing the instruction 'real' work
for the following helpers that were missing this behavior:

- VSX_CVT_INT_TO_FP_VECTOR
- VSX_CVT_FP_TO_FP
- VSX_CVT_FP_TO_INT_VECTOR
- VSX_CVT_FP_TO_INT2
- VSX_CVT_FP_TO_INT
- VSX_CVT_FP_TO_FP_HP
- VSX_CVT_FP_TO_FP_VECTOR
- VSX_CMP
- VSX_ROUND
- xscvqpdp
- xscvdpsp[n]

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220906125523.38765-9-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Zero second doubleword of VSR registers for FPR insns
Víctor Colombo [Tue, 6 Sep 2022 12:55:22 +0000 (09:55 -0300)]
target/ppc: Zero second doubleword of VSR registers for FPR insns

FPR register are mapped to the first doubleword of the VSR registers.
Since PowerISA v3.1, the second doubleword of the target register
must be zeroed for FP instructions.

This patch does it by writting 0 to the second dw everytime the
first dw is being written using set_fpr.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220906125523.38765-8-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Set OV32 when OV is set
Víctor Colombo [Tue, 6 Sep 2022 12:55:21 +0000 (09:55 -0300)]
target/ppc: Set OV32 when OV is set

According to PowerISA: "OV32 is set whenever OV is implicitly set, and
is set to the same value that OV is defined to be set to in 32-bit
mode".

This patch changes helper_update_ov_legacy to set/clear ov32 when
applicable.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220906125523.38765-7-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Zero second doubleword for VSX madd instructions
Víctor Colombo [Tue, 6 Sep 2022 12:55:20 +0000 (09:55 -0300)]
target/ppc: Zero second doubleword for VSX madd instructions

In 205eb5a89e we updated most VSX instructions to zero the
second doubleword, as is requested by PowerISA since v3.1.
However, VSX_MADD helper was left behind unchanged, while it
is also affected and should be fixed as well.

This patch applies the fix for MADD instructions.

Fixes: 205eb5a89e ("target/ppc: Change VSX instructions behavior to fill with zeros")
Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220906125523.38765-6-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Set result to QNaN for DENBCD when VXCVI occurs
Víctor Colombo [Tue, 6 Sep 2022 12:55:19 +0000 (09:55 -0300)]
target/ppc: Set result to QNaN for DENBCD when VXCVI occurs

According to the ISA, for instruction DENBCD:
"If an invalid BCD digit or sign code is detected in the source
operand, an invalid-operation exception (VXCVI) occurs."

In the Invalid Operation Exception section, there is the situation:
"When Invalid Operation Exception is disabled (VE=0) and Invalid
Operation occurs (...) If the operation is an (...) or format the
target FPR is set to a Quiet NaN". This was not being done in
QEMU.

This patch sets the result to QNaN when the instruction DENBCD causes
an Invalid Operation Exception.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220906125523.38765-5-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Zero second doubleword in DFP instructions
Víctor Colombo [Tue, 6 Sep 2022 12:55:18 +0000 (09:55 -0300)]
target/ppc: Zero second doubleword in DFP instructions

Starting at PowerISA v3.1, the second doubleword of the registers
used to store results in DFP instructions are supposed to be zeroed.

From the ISA, chapter 7.2.1.1 Floating-Point Registers:
"""
Chapter 4. Floating-Point Facility provides 32 64-bit
FPRs. Chapter 5. Decimal Floating-Point also employs
FPRs in decimal floating-point (DFP) operations. When
VSX is implemented, the 32 FPRs are mapped to
doubleword 0 of VSRs 0-31. (...)
All instructions that operate on an FPR are redefined
to operate on doubleword element 0 of the
corresponding VSR. (...)
and the contents of doubleword element 1 of the
VSR corresponding to the target FPR or FPR pair for these
instructions are set to 0.
"""

Before, the result stored at doubleword 1 was said to be undefined.

With that, this patch changes the DFP facility to zero doubleword 1
when using set_dfp64 and set_dfp128. This fixes the behavior for ISA
3.1 while keeping the behavior correct for previous ones.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220906125523.38765-4-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Remove unused xer_* macros
Víctor Colombo [Tue, 6 Sep 2022 12:55:17 +0000 (09:55 -0300)]
target/ppc: Remove unused xer_* macros

The macros xer_ov, xer_ca, xer_ov32, and xer_ca32 are both unused and
hiding the usage of env. Remove them.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220906125523.38765-3-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Remove extra space from s128 field in ppc_vsr_t
Víctor Colombo [Tue, 6 Sep 2022 12:55:16 +0000 (09:55 -0300)]
target/ppc: Remove extra space from s128 field in ppc_vsr_t

Very trivial rogue space removal. There are two spaces between Int128
and s128 in ppc_vsr_t struct, where it should be only one.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220906125523.38765-2-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Merge fsqrt and fsqrts helpers
Víctor Colombo [Mon, 5 Sep 2022 12:37:46 +0000 (09:37 -0300)]
target/ppc: Merge fsqrt and fsqrts helpers

These two helpers are almost identical, differing only by the softfloat
operation it calls. Merge them into one using a macro.
Also, take this opportunity to capitalize the helper name as we moved
the instruction to decodetree in a previous patch.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220905123746.54659-4-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Move fsqrts to decodetree
Víctor Colombo [Mon, 5 Sep 2022 12:37:45 +0000 (09:37 -0300)]
target/ppc: Move fsqrts to decodetree

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220905123746.54659-3-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Move fsqrt to decodetree
Víctor Colombo [Mon, 5 Sep 2022 12:37:44 +0000 (09:37 -0300)]
target/ppc: Move fsqrt to decodetree

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220905123746.54659-2-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Implement hashstp and hashchkp
Víctor Colombo [Fri, 15 Jul 2022 20:54:39 +0000 (17:54 -0300)]
target/ppc: Implement hashstp and hashchkp

Implementation for instructions hashstp and hashchkp, the privileged
versions of hashst and hashchk, which were added in Power ISA 3.1B.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Lucas Mateus Castro <lucas.araujo@eldorado.org.br>
Message-Id: <20220715205439.161110-4-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Implement hashst and hashchk
Víctor Colombo [Fri, 15 Jul 2022 20:54:38 +0000 (17:54 -0300)]
target/ppc: Implement hashst and hashchk

Implementation for instructions hashst and hashchk, which were added
in Power ISA 3.1B.

It was decided to implement the hash algorithm from ground up in this
patch exactly as described in Power ISA.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Lucas Mateus Castro <lucas.araujo@eldorado.org.br>
Message-Id: <20220715205439.161110-3-victor.colombo@eldorado.org.br>
[danielhb: fix block comment in excp_helper.c]
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agotarget/ppc: Add HASHKEYR and HASHPKEYR SPRs
Víctor Colombo [Fri, 15 Jul 2022 20:54:37 +0000 (17:54 -0300)]
target/ppc: Add HASHKEYR and HASHPKEYR SPRs

Add the Special Purpose Registers HASHKEYR and HASHPKEYR, which were
introduced by the Power ISA 3.1B. They are used by the new instructions
hashchk(p) and hashst(p).

The ISA states that the Operating System should generate the value for
these registers when creating a process, so it's its responsability to
do so. We initialize it with 0 for qemu-softmmu, and set a random 64
bits value for linux-user.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Lucas Mateus Castro <lucas.araujo@eldorado.org.br>
Message-Id: <20220715205439.161110-2-victor.colombo@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
22 months agoqga: Replace 'blacklist' and 'whitelist' in the guest agent sources
Thomas Huth [Wed, 27 Jul 2022 09:21:34 +0000 (11:21 +0200)]
qga: Replace 'blacklist' and 'whitelist' in the guest agent sources

Let's use better, more inclusive wording here.

Message-Id: <20220727092135.302915-3-thuth@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Konstantin Kostiuk <kkostiuk@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
22 months agoqga: Replace 'blacklist' command line and config file options by 'block-rpcs'
Thomas Huth [Wed, 27 Jul 2022 09:21:33 +0000 (11:21 +0200)]
qga: Replace 'blacklist' command line and config file options by 'block-rpcs'

Let's use a more appropriate wording for this command line and config
file option. The old ones are still accepted for compatibility reasons,
but marked as deprecated now so that it could be removed in a future
version of QEMU.

This change is based on earlier patches from Philippe Mathieu-Daudé,
with the idea for the new option name suggested by BALATON Zoltan.

And while we're at it, replace the "?" in the help text with "help"
since that does not have the problem of conflicting with the wildcard
character of the shells.

Message-Id: <20220727092135.302915-2-thuth@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Konstantin Kostiuk <kkostiuk@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
22 months agogitlab-ci: Update the FreeBSD 13 job from 13.0 to 13.1
Thomas Huth [Tue, 20 Sep 2022 09:27:10 +0000 (11:27 +0200)]
gitlab-ci: Update the FreeBSD 13 job from 13.0 to 13.1

The FreeBSD 13 job in our CI started failing since the python port
stopped working after 13.1 has been released. Thus update our CI
job to FreeBSD 13.1 to get it working again.

Suggested-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220920102041.45067-1-thuth@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
22 months agotests: sb16 has both pc and q35 tests
Juan Quintela [Fri, 2 Sep 2022 17:34:49 +0000 (19:34 +0200)]
tests: sb16 has both pc and q35 tests

Check that the machines are compiled in before calling it

Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20220902173452.1904-6-quintela@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
22 months agotests: Only run intel-hda-tests if machine type is compiled in
Juan Quintela [Fri, 2 Sep 2022 17:34:48 +0000 (19:34 +0200)]
tests: Only run intel-hda-tests if machine type is compiled in

Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20220902173452.1904-5-quintela@redhat.com>
Reviewed-by: Ani Sinha <ani@anisinha.ca>
Signed-off-by: Thomas Huth <thuth@redhat.com>
22 months agobios-tables-test: Only run test for machine types compiled in
Juan Quintela [Fri, 2 Sep 2022 17:34:47 +0000 (19:34 +0200)]
bios-tables-test: Only run test for machine types compiled in

Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20220902173452.1904-4-quintela@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>
22 months agobios-tables-test: Sort all x86_64 tests by machine type
Juan Quintela [Fri, 2 Sep 2022 17:34:46 +0000 (19:34 +0200)]
bios-tables-test: Sort all x86_64 tests by machine type

No code change here, just move test around.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20220902173452.1904-3-quintela@redhat.com>
Reviewed-by: Ani Sinha <ani@anisinha.ca>
Signed-off-by: Thomas Huth <thuth@redhat.com>
22 months agobios-tables-test: Make oem-fields tests be consistent
Juan Quintela [Fri, 2 Sep 2022 17:34:45 +0000 (19:34 +0200)]
bios-tables-test: Make oem-fields tests be consistent

Every other test function is named:

test_acpi_<machine>_<test>()

Just make this test the same.  Once there, rename "acpi/oem-fields" to
"acpi/piix4/oem-fields" so it is consistent with everything else.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20220902173452.1904-2-quintela@redhat.com>
Reviewed-by: Ani Sinha <ani@anisinha.ca>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>
22 months agomeson-build: Enable CONFIG_REPLICATION only when replication is set
Juan Quintela [Fri, 2 Sep 2022 16:51:25 +0000 (18:51 +0200)]
meson-build: Enable CONFIG_REPLICATION only when replication is set

Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20220902165126.1482-8-quintela@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>