target/arm: add descriptions of CLIDR_EL1, CCSIDR_EL1, CTR_EL0 to cpu.h

Signed-off-by: Leif Lindholm <leif@nuviainc.com>
Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Message-id: 20210108185154.8108-5-leif@nuviainc.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: make ARMCPU.ctr 64-bit

When FEAT_MTE is implemented, the AArch64 view of CTR_EL0 adds the
TminLine field in bits [37:32].
Extend the ctr field to be able to hold this context.

Signed-off-by: Leif Lindholm <leif@nuviainc.com>
Reviewed-by: Hao Wu <wuhaotsh@google.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Message-id: 20210108185154.8108-4-leif@nuviainc.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: make ARMCPU.clidr 64-bit

The AArch64 view of CLIDR_EL1 extends the ICB field to include also bit
32, as well as adding a Ttype<n> field when FEAT_MTE is implemented.
Extend the clidr field to be able to hold this context.

Signed-off-by: Leif Lindholm <leif@nuviainc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Message-id: 20210108185154.8108-3-leif@nuviainc.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: fix typo in cpu.h ID_AA64PFR1 field name

SBSS -> SSBS

Signed-off-by: Leif Lindholm <leif@nuviainc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Message-id: 20210108185154.8108-2-leif@nuviainc.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: enable Small Translation tables in max CPU

Signed-off-by: Rémi Denis-Courmont <remi.denis.courmont@huawei.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: ARMv8.4-TTST extension

This adds for the Small Translation tables extension in AArch64 state.

Signed-off-by: Rémi Denis-Courmont <remi.denis.courmont@huawei.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/huth-gitlab/tags/pull-request-2021-01-11v2' into staging

* Fuzzer improvements
* Add OpenSUSE leap to the gitlab-CI
* Some fixes to get our CI "green" again
* Some initial patches to update bsd-user

# gpg: Signature made Mon 11 Jan 2021 14:00:07 GMT
# 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

* remotes/huth-gitlab/tags/pull-request-2021-01-11v2:
  fuzz: map all BARs and enable PCI devices
  tests/acceptance: Fix race conditions in s390x tests & skip fedora on gitlab-CI
  bsd-user: Update strace.list for FreeBSD's latest syscalls
  bsd-user: move strace OS/arch dependent code to host/arch dirs
  bsd-user: regenerate FreeBSD's system call numbers
  fuzz: heuristic split write based on past IOs
  fuzz: add minimization options
  fuzz: set bits in operand of write/out to zero
  fuzz: remove IO commands iteratively
  fuzz: split write operand using binary approach
  fuzz: double the IOs to remove for every loop
  fuzz: accelerate non-crash detection
  util/oslib-win32: Fix _aligned_malloc() arguments order
  qtest/libqtest: fix heap-buffer-overflow in qtest_cb_for_every_machine()
  gitlab-ci.yml: Add openSUSE Leap 15.2 for gitlab CI/CD

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

fuzz: map all BARs and enable PCI devices

Prior to this patch, the fuzzer found inputs to map PCI device BARs and
enable the device. While it is nice that the fuzzer can do this, it
added significant overhead, since the fuzzer needs to map all the
BARs (regenerating the memory topology), at the start of each input.
With this patch, we do this once, before fuzzing, mitigating some of
this overhead.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201221181203.1853-1-alxndr@bu.edu>
Signed-off-by: Thomas Huth <thuth@redhat.com>

tests/acceptance: Fix race conditions in s390x tests & skip fedora on gitlab-CI

There was a race condition in the first test where there was already the
"crw" output in the dmesg, but the "0.0.4711" entry has not been created
in the /sys fs yet. Fix it by waiting until it is there.

The second test has even more problems on gitlab-CI. Even after adding some
more synchronization points (that wait for some messages in the "dmesg"
output to make sure that the modules got loaded correctly), there are still
occasionally some hangs in this test when it is running in the gitlab-CI.
So far I was unable to reproduce these hangs locally on my computer, so
this issue might take a while to debug. Thus disable the 2nd test in the
gitlab-CI until the problems are better understood and fixed.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Tested-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Message-Id: <20210108185645.86351-1-thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

bsd-user: Update strace.list for FreeBSD's latest syscalls

Update strace.list to include all of FreeBSD's syscalls up through svn
r331280.

Signed-off-by: Stacey Son <sson@FreeBSD.org>
Signed-off-by: Sean Bruno <sburno@FreeBSD.org>
Signed-off-by: Alexander Kabaev <kan@FreeBSD.org>
Signed-off-by: Jung-uk Kim <jkim@FreeBSD.org>
Author: Michal Meloun <mmel@FreeBSD.org>
Signed-off-by: Mikaël Urankar <mikael.urankar@gmail.com>
[imp moved this change to early in the sequence]
Signed-off-by: Warner Losh <imp@FreeBSD.org>
Message-Id: <20201218205451.10559-5-imp@freebsd.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>

bsd-user: move strace OS/arch dependent code to host/arch dirs

This change moves host OS and arch dependent code for the sysarch
system call related to the -strace functionality into the
appropriate host OS and target arch directories.

Signed-off-by: Stacey Son <sson@FreeBSD.org>
Signed-off-by: Sean Bruno <sburno@FreeBSD.org>
[ imp integrated minor build fixes from sbruno ]
Signed-off-by: Warner Losh <imp@FreeBSD.org>
Message-Id: <20201218205451.10559-4-imp@freebsd.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>

bsd-user: regenerate FreeBSD's system call numbers

Recreate the FreeBSD system call numbers from current sys/syscall.h. Since this
isn't guaranteed to be on all systems, continue the practice of generating it
with some variation on:
sed -e s/SYS_/TARGET_NR_/ < $FREEBSD_SRC/sys/syscall.h > syscall_nr.h
until a more comprehensive system can be put in place.

Signed-off-by: Warner Losh <imp@FreeBSD.org>
Message-Id: <20201218205451.10559-2-imp@freebsd.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>

fuzz: heuristic split write based on past IOs

If previous write commands write the same length of data with the same step,
we view it as a hint.

Signed-off-by: Qiuhao Li <Qiuhao.Li@outlook.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <SYCPR01MB3502480AD07811A6A49B8FEAFCAB0@SYCPR01MB3502.ausprd01.prod.outlook.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

fuzz: add minimization options

-M1: remove IO commands iteratively
-M2: try setting bits in operand of write/out to zero

Signed-off-by: Qiuhao Li <Qiuhao.Li@outlook.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <SYCPR01MB350204C52E7A39E6B0EEC870FCAB0@SYCPR01MB3502.ausprd01.prod.outlook.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

fuzz: set bits in operand of write/out to zero

Simplifying the crash cases by opportunistically setting bits in operands of
out/write to zero may help to debug, since usually bit one means turn on or
trigger a function while zero is the default turn-off setting.

Tested bug https://bugs.launchpad.net/qemu/+bug/1908062

Signed-off-by: Qiuhao Li <Qiuhao.Li@outlook.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <SYCPR01MB3502C84B6346A3E3DE708C7BFCAB0@SYCPR01MB3502.ausprd01.prod.outlook.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

fuzz: remove IO commands iteratively

Now we use a one-time scan and remove strategy in the minimizer,
which is not suitable for timing dependent instructions.

For example, instruction A will indicate an address where the config
chunk locates, and instruction B will make the configuration active.
If we have the following instruction sequence:

...
A1
B1
A2
B2
...

A2 and B2 are the actual instructions that trigger the bug.

If we scan from top to bottom, after we remove A1, the behavior of B1
might be unknowable, including not to crash the program. But we will
successfully remove B1 later cause A2 and B2 will crash the process
anyway:

...
A1
A2
B2
...

Now one more trimming will remove A1.

In the perfect case, we would need to be able to remove A and B (or C!) at
the same time. But for now, let's just add a loop around the minimizer.

Since we only remove instructions, this iterative algorithm is converging.

Tested with Bug 1908062.

Signed-off-by: Qiuhao Li <Qiuhao.Li@outlook.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <SYCPR01MB350263004448040ACCB9A9F1FCAB0@SYCPR01MB3502.ausprd01.prod.outlook.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

fuzz: split write operand using binary approach

Currently, we split the write commands' data from the middle. If it does not
work, try to move the pivot left by one byte and retry until there is no
space.

But, this method has two flaws:

1. It may fail to trim all unnecessary bytes on the right side.

For example, there is an IO write command:

  write addr uuxxxxuu

u is the unnecessary byte for the crash. Unlike ram write commands, in most
case, a split IO write won't trigger the same crash, So if we split from the
middle, we will get:

  write addr uu (will be removed in next round)
  write addr xxxxuu

For xxxxuu, since split it from the middle and retry to the leftmost byte
won't get the same crash, we will be stopped from removing the last two
bytes.

2. The algorithm complexity is O(n) since we move the pivot byte by byte.

To solve the first issue, we can try a symmetrical position on the right if
we fail on the left. As for the second issue, instead moving by one byte, we
can approach the boundary exponentially, achieving O(log(n)).

Give an example:

                   xxxxuu len=6
                        +
                        |
                        +
                 xxx,xuu 6/2=3 fail
                        +
         +--------------+-------------+
         |                            |
         +                            +
  xx,xxuu 6/2^2=1 fail         xxxxu,u 6-1=5 success
                                 +   +
         +------------------+----+   |
         |                  |        +-------------+ u removed
         +                  +
   xx,xxu 5/2=2 fail  xxxx,u 6-2=4 success
                           +
                           |
                           +-----------+ u removed

In some rare cases, this algorithm will fail to trim all unnecessary bytes:

  xxxxxxxxxuxxxxxx
  xxxxxxxx-xuxxxxxx Fail
  xxxx-xxxxxuxxxxxx Fail
  xxxxxxxxxuxx-xxxx Fail
  ...

I think the trade-off is worth it.

Signed-off-by: Qiuhao Li <Qiuhao.Li@outlook.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <SYCPR01MB3502D26F1BEB680CBBC169E5FCAB0@SYCPR01MB3502.ausprd01.prod.outlook.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

fuzz: double the IOs to remove for every loop

Instead of removing IO instructions one by one, we can try deleting multiple
instructions at once. According to the locality of reference, we double the
number of instructions to remove for the next round and recover it to one
once we fail.

This patch is usually significant for large input.

Test with quadrupled trace input at:
  https://bugs.launchpad.net/qemu/+bug/1890333/comments/1

Patched 1/6 version:
  real  0m45.904s
  user  0m16.874s
  sys   0m10.042s

Refined version:
  real  0m11.412s
  user  0m6.888s
  sys   0m3.325s

Signed-off-by: Qiuhao Li <Qiuhao.Li@outlook.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <SYCPR01MB350280A67BB55C3FADF173E3FCAB0@SYCPR01MB3502.ausprd01.prod.outlook.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

fuzz: accelerate non-crash detection

We spend much time waiting for the timeout program during the minimization
process until it passes a time limit. This patch hacks the CLOSED (indicates
the redirection file closed) notification in QTest's output if it doesn't
crash.

Test with quadrupled trace input at:
  https://bugs.launchpad.net/qemu/+bug/1890333/comments/1

Original version:
  real 1m37.246s
  user 0m13.069s
  sys 0m8.399s

Refined version:
  real 0m45.904s
  user 0m16.874s
  sys 0m10.042s

Note:

Sometimes the mutated or the same trace may trigger a different crash
summary (second-to-last line) but indicates the same bug. For example, Bug
1910826 [1], which will trigger a stack overflow, may output summaries
like:

SUMMARY: AddressSanitizer: stack-overflow
/home/qiuhao/hack/qemu/build/../softmmu/physmem.c:488 in
flatview_do_translate

or

SUMMARY: AddressSanitizer: stack-overflow
(/home/qiuhao/hack/qemu/build/qemu-system-i386+0x27ca049) in __asan_memcpy

Etc.

If we use the whole summary line as the token, we may be prevented from
further minimization. So in this patch, we only use the first three words
which indicate the type of crash:

SUMMARY: AddressSanitizer: stack-overflow

[1] https://bugs.launchpad.net/qemu/+bug/1910826

Signed-off-by: Qiuhao Li <Qiuhao.Li@outlook.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <SYCPR01MB350251DC04003450348FAF68FCAB0@SYCPR01MB3502.ausprd01.prod.outlook.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

util/oslib-win32: Fix _aligned_malloc() arguments order

Commit dfbd0b873a8 inadvertently swapped the arguments
of _aligned_malloc(), correct it to fix [*]:

  G_TEST_SRCDIR=C:/Users/ContainerAdministrator/AppData/Local/Temp/cirrus-ci-build/tests
  G_TEST_BUILDDIR=C:/Users/ContainerAdministrator/AppData/Local/Temp/cirrus-ci-build/build/tests
  tests/test-qht.exe --tap -k
  ERROR test-qht - too few tests run (expected 2, got 0)
  make: *** [Makefile.mtest:256: run-test-30] Error 1

[*] https://cirrus-ci.com/task/6055645751279616?command=test#L593

Fixes: dfbd0b873a8 ("util/oslib-win32: Use _aligned_malloc for qemu_try_memalign")
Reported-by: Yonggang Luo <luoyonggang@gmail.com>
Reported-by: Volker Rümelin <vr_qemu@t-online.de>
Suggested-by: Volker Rümelin <vr_qemu@t-online.de>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Message-Id: <20210111001606.1122983-1-f4bug@amsat.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>

qtest/libqtest: fix heap-buffer-overflow in qtest_cb_for_every_machine()

When the length of mname is less than 5, memcpy("xenfv", mname, 5) will cause
heap buffer overflow. Therefore, use strncmp to avoid this problem.

The asan showed stack:

ERROR: AddressSanitizer: heap-buffer-overflow on address 0x60200000f2f4 at
pc 0x7f65d8cc2225 bp 0x7ffe93cc5a60 sp 0x7ffe93cc5208 READ of size 5 at
0x60200000f2f4 thread T0
    #0 0x7f65d8cc2224 in memcmp (/lib64/libasan.so.5+0xdf224)
    #1 0x5632c20be95b in qtest_cb_for_every_machine tests/qtest/libqtest.c:1282
    #2 0x5632c20b7995 in main tests/qtest/test-hmp.c:160
    #3 0x7f65d88fed42 in __libc_start_main (/lib64/libc.so.6+0x26d42)
    #4 0x5632c20b72cd in _start (build/tests/qtest/test-hmp+0x542cd)

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Gan Qixin <ganqixin@huawei.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20210106050625.518041-1-ganqixin@huawei.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

gitlab-ci.yml: Add openSUSE Leap 15.2 for gitlab CI/CD

Add build-system-opensuse jobs and opensuse-leap.docker dockerfile.
Use openSUSE Leap 15.2 container image in the gitlab-CI.

Signed-off-by: Cho, Yu-Chen <acho@suse.com>
Tested-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Message-Id: <20201229085046.8536-1-acho@suse.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20210108' into staging

target-arm queue:
 * intc/arm_gic: Fix gic_irq_signaling_enabled() for vCPUs
 * target/arm: Fix MTE0_ACTIVE
 * target/arm: Implement v8.1M and Cortex-M55 model
 * hw/arm/highbank: Drop dead KVM support code
 * util/qemu-timer: Make timer_free() imply timer_del()
 * various devices: Use ptimer_free() in finalize function
 * docs/system: arm: Add sabrelite board description
 * sabrelite: Minor fixes to allow booting U-Boot

# gpg: Signature made Fri 08 Jan 2021 15:34:25 GMT
# gpg:                using RSA key E1A5C593CD419DE28E8315CF3C2525ED14360CDE
# gpg:                issuer "peter.maydell@linaro.org"
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>" [ultimate]
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>" [ultimate]
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>" [ultimate]
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* remotes/pmaydell/tags/pull-target-arm-20210108: (23 commits)
  docs/system: arm: Add sabrelite board description
  hw/arm: sabrelite: Connect the Ethernet PHY at address 6
  hw/msic: imx6_ccm: Correct register value for silicon type
  hw/misc: imx6_ccm: Update PMU_MISC0 reset value
  exynos4210_mct: Use ptimer_free() in the finalize function to avoid memleaks
  musicpal: Use ptimer_free() in the finalize function to avoid memleaks
  mss-timer: Use ptimer_free() in the finalize function to avoid memleaks
  exynos4210_pwm: Use ptimer_free() in the finalize function to avoid memleaks
  exynos4210_rtc: Use ptimer_free() in the finalize function to avoid memleaks
  allwinner-a10-pit: Use ptimer_free() in the finalize function to avoid memleaks
  digic-timer: Use ptimer_free() in the finalize function to avoid memleaks
  target/arm: Remove timer_del()/timer_deinit() before timer_free()
  Remove superfluous timer_del() calls
  scripts/coccinelle: New script to remove unnecessary timer_del() calls
  util/qemu-timer: Make timer_free() imply timer_del()
  hw/arm/highbank: Drop dead KVM support code
  target/arm: Implement Cortex-M55 model
  target/arm: Implement FPCXT_NS fp system register
  target/arm: Correct store of FPSCR value via FPCXT_S
  hw/intc/armv7m_nvic: Correct handling of CCR.BFHFNMIGN
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

docs/system: arm: Add sabrelite board description

This adds the target guide for SABRE Lite board, and documents how
to boot a Linux kernel and U-Boot bootloader.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20210106063504.10841-5-bmeng.cn@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm: sabrelite: Connect the Ethernet PHY at address 6

At present, when booting U-Boot on QEMU sabrelite, we see:

  Net:   Board Net Initialization Failed
  No ethernet found.

U-Boot scans PHY at address 4/5/6/7 (see board_eth_init() in the
U-Boot source: board/boundary/nitrogen6x/nitrogen6x.c). On the real
board, the Ethernet PHY is at address 6. Adjust this by updating the
"fec-phy-num" property of the fsl_imx6 SoC object.

With this change, U-Boot sees the PHY but complains MAC address:

  Net:   using phy at 6
  FEC [PRIME]
  Error: FEC address not set.

This is due to U-Boot tries to read the MAC address from the fuse,
which QEMU does not have any valid content filled in. However this
does not prevent the Ethernet from working in QEMU. We just need to
set up the MAC address later in the U-Boot command shell, by:

  => setenv ethaddr 00:11:22:33:44:55

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20210106063504.10841-4-bmeng.cn@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/msic: imx6_ccm: Correct register value for silicon type

Currently when U-Boot boots, it prints "??" for i.MX processor:

  CPU:   Freescale i.MX?? rev1.0 at 792 MHz

The register that was used to determine the silicon type is
undocumented in the latest IMX6DQRM (Rev. 6, 05/2020), but we
can refer to get_cpu_rev() in arch/arm/mach-imx/mx6/soc.c in
the U-Boot source codes that USB_ANALOG_DIGPROG is used.

Update its reset value to indicate i.MX6Q.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20210106063504.10841-3-bmeng.cn@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/misc: imx6_ccm: Update PMU_MISC0 reset value

U-Boot expects PMU_MISC0 register bit 7 is set (see init_bandgap()
in arch/arm/mach-imx/mx6/soc.c) during boot. This bit indicates the
bandgap has stabilized.

With this change, the latest upstream U-Boot (v2021.01-rc3) for imx6
sabrelite board (mx6qsabrelite_defconfig), with a slight change made
by switching CONFIG_OF_SEPARATE to CONFIG_OF_EMBED, boots to U-Boot
shell on QEMU with the following command:

$ qemu-system-arm -M sabrelite -smp 4 -m 1G -kernel u-boot \
    -display none -serial null -serial stdio

Boot log below:

  U-Boot 2021.01-rc3 (Dec 12 2020 - 17:40:02 +0800)

  CPU:   Freescale i.MX?? rev1.0 at 792 MHz
  Reset cause: POR
  Model: Freescale i.MX6 Quad SABRE Lite Board
  Board: SABRE Lite
  I2C:   ready
  DRAM:  1 GiB
  force_idle_bus: sda=0 scl=0 sda.gp=0x5c scl.gp=0x55
  force_idle_bus: failed to clear bus, sda=0 scl=0
  force_idle_bus: sda=0 scl=0 sda.gp=0x6d scl.gp=0x6c
  force_idle_bus: failed to clear bus, sda=0 scl=0
  force_idle_bus: sda=0 scl=0 sda.gp=0xcb scl.gp=0x5
  force_idle_bus: failed to clear bus, sda=0 scl=0
  MMC:   FSL_SDHC: 0, FSL_SDHC: 1
  Loading Environment from MMC... *** Warning - No block device, using default environment

  In:    serial
  Out:   serial
  Err:   serial
  Net:   Board Net Initialization Failed
  No ethernet found.
  starting USB...
  Bus usb@2184000: usb dr_mode not found
  USB EHCI 1.00
  Bus usb@2184200: USB EHCI 1.00
  scanning bus usb@2184000 for devices... 1 USB Device(s) found
  scanning bus usb@2184200 for devices... 1 USB Device(s) found
         scanning usb for storage devices... 0 Storage Device(s) found
         scanning usb for ethernet devices... 0 Ethernet Device(s) found
  Hit any key to stop autoboot:  0
  =>

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20210106063504.10841-2-bmeng.cn@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

exynos4210_mct: Use ptimer_free() in the finalize function to avoid memleaks

When running device-introspect-test, a memory leak occurred in the
exynos4210_mct_init function, so use ptimer_free() in the finalize function to
avoid it.

ASAN shows memory leak stack:

Indirect leak of 96 byte(s) in 1 object(s) allocated from:
    #0 0xffffab97e1f0 in __interceptor_calloc (/lib64/libasan.so.5+0xee1f0)
    #1 0xffffab256800 in g_malloc0 (/lib64/libglib-2.0.so.0+0x56800)
    #2 0xaaabf555db78 in ptimer_init /qemu/hw/core/ptimer.c:432
    #3 0xaaabf56b01a0 in exynos4210_mct_init /qemu/hw/timer/exynos4210_mct.c:1505
    #4 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #5 0xaaabf633a1e0 in object_new_with_type /qemu/qom/object.c:729
    #6 0xaaabf6375e40 in qmp_device_list_properties /qemu/qom/qom-qmp-cmds.c:153
    #7 0xaaabf653d8ec in qmp_marshal_device_list_properties /qemu/qapi/qapi-commands-qdev.c:59
    #8 0xaaabf6587d08 in do_qmp_dispatch_bh /qemu/qapi/qmp-dispatch.c:110
    #9 0xaaabf6552708 in aio_bh_call /qemu/util/async.c:136
    #10 0xaaabf6552708 in aio_bh_poll /qemu/util/async.c:164
    #11 0xaaabf655f19c in aio_dispatch /qemu/util/aio-posix.c:381
    #12 0xaaabf65523f4 in aio_ctx_dispatch /qemu/util/async.c:306

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Gan Qixin <ganqixin@huawei.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

musicpal: Use ptimer_free() in the finalize function to avoid memleaks

When running device-introspect-test, a memory leak occurred in the
mv88w8618_pit_init function, so use ptimer_free() in the finalize function to
avoid it.

ASAN shows memory leak stack:

Indirect leak of 192 byte(s) in 4 object(s) allocated from:
    #0 0xffffab97e1f0 in __interceptor_calloc (/lib64/libasan.so.5+0xee1f0)
    #1 0xffffab256800 in g_malloc0 (/lib64/libglib-2.0.so.0+0x56800)
    #2 0xaaabf555db84 in timer_new_full /qemu/include/qemu/timer.h:523
    #3 0xaaabf555db84 in timer_new /qemu/include/qemu/timer.h:544
    #4 0xaaabf555db84 in timer_new_ns /qemu/include/qemu/timer.h:562
    #5 0xaaabf555db84 in ptimer_init /qemu/hw/core/ptimer.c:433
    #6 0xaaabf5bb2290 in mv88w8618_timer_init /qemu/hw/arm/musicpal.c:862
    #7 0xaaabf5bb2290 in mv88w8618_pit_init /qemu/hw/arm/musicpal.c:954
    #8 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #9 0xaaabf633a1e0 in object_new_with_type /qemu/qom/object.c:729
    #10 0xaaabf6375e40 in qmp_device_list_properties /qemu/qom/qom-qmp-cmds.c:153
    #11 0xaaabf5a95540 in qdev_device_help /qemu/softmmu/qdev-monitor.c:283
    #12 0xaaabf5a96940 in qmp_device_add /qemu/softmmu/qdev-monitor.c:801

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Gan Qixin <ganqixin@huawei.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

mss-timer: Use ptimer_free() in the finalize function to avoid memleaks

When running device-introspect-test, a memory leak occurred in the
mss_timer_init function, so use ptimer_free() in the finalize function to avoid
it.

ASAN shows memory leak stack:

Indirect leak of 192 byte(s) in 2 object(s) allocated from:
    #0 0xffffab97e1f0 in __interceptor_calloc (/lib64/libasan.so.5+0xee1f0)
    #1 0xffffab256800 in g_malloc0 (/lib64/libglib-2.0.so.0+0x56800)
    #2 0xaaabf555db78 in ptimer_init /qemu/hw/core/ptimer.c:432
    #3 0xaaabf58a0010 in mss_timer_init /qemu/hw/timer/mss-timer.c:235
    #4 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #5 0xaaabf633ca04 in object_initialize_child_with_propsv /qemu/qom/object.c:564
    #6 0xaaabf633cc08 in object_initialize_child_with_props /qemu/qom/object.c:547
    #7 0xaaabf5b8316c in m2sxxx_soc_initfn /qemu/hw/arm/msf2-soc.c:70
    #8 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #9 0xaaabf633a1e0 in object_new_with_type /qemu/qom/object.c:729
    #10 0xaaabf6375e40 in qmp_device_list_properties /qemu/qom/qom-qmp-cmds.c:153
    #11 0xaaabf653d8ec in qmp_marshal_device_list_properties /qemu/qapi/qapi-commands-qdev.c:59
    #12 0xaaabf6587d08 in do_qmp_dispatch_bh /qemu/qapi/qmp-dispatch.c:110

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Gan Qixin <ganqixin@huawei.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

exynos4210_pwm: Use ptimer_free() in the finalize function to avoid memleaks

When running device-introspect-test, a memory leak occurred in the
exynos4210_pwm_init function, so use ptimer_free() in the finalize function to
avoid it.

ASAN shows memory leak stack:

Indirect leak of 240 byte(s) in 5 object(s) allocated from:
    #0 0xffffab97e1f0 in __interceptor_calloc (/lib64/libasan.so.5+0xee1f0)
    #1 0xffffab256800 in g_malloc0 (/lib64/libglib-2.0.so.0+0x56800)
    #2 0xaaabf555db84 in timer_new_full /qemu/include/qemu/timer.h:523
    #3 0xaaabf555db84 in timer_new /qemu/include/qemu/timer.h:544
    #4 0xaaabf555db84 in timer_new_ns /qemu/include/qemu/timer.h:562
    #5 0xaaabf555db84 in ptimer_init /qemu/hw/core/ptimer.c:433
    #6 0xaaabf56a36cc in exynos4210_pwm_init /qemu/hw/timer/exynos4210_pwm.c:401
    #7 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #8 0xaaabf633a1e0 in object_new_with_type /qemu/qom/object.c:729
    #9 0xaaabf6375e40 in qmp_device_list_properties /qemu/qom/qom-qmp-cmds.c:153
    #10 0xaaabf653d8ec in qmp_marshal_device_list_properties /qemu/qapi/qapi-commands-qdev.c:59
    #11 0xaaabf6587d08 in do_qmp_dispatch_bh /qemu/qapi/qmp-dispatch.c:110
    #12 0xaaabf6552708 in aio_bh_call /qemu/util/async.c:136

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Gan Qixin <ganqixin@huawei.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

exynos4210_rtc: Use ptimer_free() in the finalize function to avoid memleaks

When running device-introspect-test, a memory leak occurred in the
exynos4210_rtc_init function, so use ptimer_free() in the finalize function to
avoid it.

ASAN shows memory leak stack:

Indirect leak of 96 byte(s) in 1 object(s) allocated from:
    #0 0xffffab97e1f0 in __interceptor_calloc (/lib64/libasan.so.5+0xee1f0)
    #1 0xffffab256800 in g_malloc0 (/lib64/libglib-2.0.so.0+0x56800)
    #2 0xaaabf555db78 in ptimer_init /qemu/hw/core/ptimer.c:432
    #3 0xaaabf57b3934 in exynos4210_rtc_init /qemu/hw/rtc/exynos4210_rtc.c:567
    #4 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #5 0xaaabf633a1e0 in object_new_with_type /qemu/qom/object.c:729
    #6 0xaaabf6375e40 in qmp_device_list_properties /qemu/qom/qom-qmp-cmds.c:153
    #7 0xaaabf653d8ec in qmp_marshal_device_list_properties /qemu/qapi/qapi-commands-qdev.c:59
    #8 0xaaabf6587d08 in do_qmp_dispatch_bh /qemu/qapi/qmp-dispatch.c:110
    #9 0xaaabf6552708 in aio_bh_call /qemu/util/async.c:136
    #10 0xaaabf6552708 in aio_bh_poll /qemu/util/async.c:164
    #11 0xaaabf655f19c in aio_dispatch /qemu/util/aio-posix.c:381
    #12 0xaaabf65523f4 in aio_ctx_dispatch /qemu/util/async.c:306

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Gan Qixin <ganqixin@huawei.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

allwinner-a10-pit: Use ptimer_free() in the finalize function to avoid memleaks

When running device-introspect-test, a memory leak occurred in the a10_pit_init
function, so use ptimer_free() in the finalize function to avoid it.

ASAN shows memory leak stack:

Indirect leak of 288 byte(s) in 6 object(s) allocated from:
    #0 0xffffab97e1f0 in __interceptor_calloc (/lib64/libasan.so.5+0xee1f0)
    #1 0xffffab256800 in g_malloc0 (/lib64/libglib-2.0.so.0+0x56800)
    #2 0xaaabf555db84 in timer_new_full /qemu/include/qemu/timer.h:523
    #3 0xaaabf555db84 in timer_new /qemu/include/qemu/timer.h:544
    #4 0xaaabf555db84 in timer_new_ns /qemu/include/qemu/timer.h:562
    #5 0xaaabf555db84 in ptimer_init /qemu/hw/core/ptimer.c:433
    #6 0xaaabf57415e8 in a10_pit_init /qemu/hw/timer/allwinner-a10-pit.c:278
    #7 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #8 0xaaabf633ca04 in object_initialize_child_with_propsv /qemu/qom/object.c:564
    #9 0xaaabf633cc08 in object_initialize_child_with_props /qemu/qom/object.c:547
    #10 0xaaabf5b94680 in aw_a10_init /qemu/hw/arm/allwinner-a10.c:49
    #11 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #12 0xaaabf633a1e0 in object_new_with_type /qemu/qom/object.c:729

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Gan Qixin <ganqixin@huawei.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

digic-timer: Use ptimer_free() in the finalize function to avoid memleaks

When running device-introspect-test, a memory leak occurred in the
digic_timer_init function, so use ptimer_free() in the finalize function to
avoid it.

ASAN shows memory leak stack:

Indirect leak of 288 byte(s) in 3 object(s) allocated from:
    #0 0xffffab97e1f0 in __interceptor_calloc (/lib64/libasan.so.5+0xee1f0)
    #1 0xffffab256800 in g_malloc0 (/lib64/libglib-2.0.so.0+0x56800)
    #2 0xaaabf555db78 in ptimer_init /qemu/hw/core/ptimer.c:432
    #3 0xaaabf5b04084 in digic_timer_init /qemu/hw/timer/digic-timer.c:142
    #4 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #5 0xaaabf633ca04 in object_initialize_child_with_propsv /qemu/qom/object.c:564
    #6 0xaaabf633cc08 in object_initialize_child_with_props /qemu/qom/object.c:547
    #7 0xaaabf5b40e84 in digic_init /qemu/hw/arm/digic.c:46
    #8 0xaaabf6339f6c in object_initialize_with_type /qemu/qom/object.c:515
    #9 0xaaabf633a1e0 in object_new_with_type /qemu/qom/object.c:729
    #10 0xaaabf6375e40 in qmp_device_list_properties /qemu/qom/qom-qmp-cmds.c:153
    #11 0xaaabf653d8ec in qmp_marshal_device_list_properties /qemu/qapi/qapi-commands-qdev.c:59
    #12 0xaaabf6587d08 in do_qmp_dispatch_bh /qemu/qapi/qmp-dispatch.c:110

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Gan Qixin <ganqixin@huawei.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: Remove timer_del()/timer_deinit() before timer_free()

The Arm CPU finalize function uses a sequence of timer_del(), timer_deinit(),
timer_free() to free the timer. The timer_deinit() step in this was always
unnecessary, and now the timer_del() is implied by timer_free(), so we can
collapse this down to simply calling timer_free().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201215154107.3255-5-peter.maydell@linaro.org

Remove superfluous timer_del() calls

This commit is the result of running the timer-del-timer-free.cocci
script on the whole source tree.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Corey Minyard <cminyard@mvista.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201215154107.3255-4-peter.maydell@linaro.org

scripts/coccinelle: New script to remove unnecessary timer_del() calls

Now that timer_free() implicitly calls timer_del(), sequences
  timer_del(mytimer);
  timer_free(mytimer);

can be simplified to just
  timer_free(mytimer);

Add a Coccinelle script to do this transformation.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201215154107.3255-3-peter.maydell@linaro.org

util/qemu-timer: Make timer_free() imply timer_del()

Currently timer_free() is a simple wrapper for g_free().  This means
that the timer being freed must not be currently active, as otherwise
QEMU might crash later when the active list is processed and still
has a pointer to freed memory on it.  As a result almost all calls to
timer_free() are preceded by a timer_del() call, as can be seen in
the output of
  git grep -B1 '\<timer_free\>'

This is unfortunate API design as it makes it easy to accidentally
misuse (by forgetting the timer_del()), and the correct use is
annoyingly verbose.

Make timer_free() imply a timer_del().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201215154107.3255-2-peter.maydell@linaro.org

hw/arm/highbank: Drop dead KVM support code

Support for running KVM on 32-bit Arm hosts was removed in commit
82bf7ae84ce739e.  You can still run a 32-bit guest on a 64-bit Arm
host CPU, but because Arm KVM requires the host and guest CPU types
to match, it is not possible to run a guest that requires a Cortex-A9
or Cortex-A15 CPU there.  That means that the code in the
highbank/midway board models to support KVM is no longer used, and we
can delete it.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20201215144215.28482-1-peter.maydell@linaro.org

target/arm: Implement Cortex-M55 model

Now that we have implemented all the features needed by the v8.1M
architecture, we can add the model of the Cortex-M55.  This is the
configuration without MVE support; we'll add MVE later.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201210201433.26262-5-peter.maydell@linaro.org

target/arm: Implement FPCXT_NS fp system register

Implement the v8.1M FPCXT_NS floating-point system register.  This is
a little more complicated than FPCXT_S, because it has specific
handling for "current FP state is inactive", and it only wants to do
PreserveFPState(), not the full set of actions done by
ExecuteFPCheck() which vfp_access_check() implements.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201210201433.26262-4-peter.maydell@linaro.org

target/arm: Correct store of FPSCR value via FPCXT_S

In commit 64f863baeedc8659 we implemented the v8.1M FPCXT_S register,
but we got the write behaviour wrong. On read, this register reads
bits [27:0] of FPSCR plus the CONTROL.SFPA bit. On write, it doesn't
just write back those bits -- it writes a value to the whole FPSCR,
whose upper 4 bits are zeroes.

We also incorrectly implemented the write-to-FPSCR as a simple store
to vfp.xregs; this skips the "update the softfloat flags" part of
the vfp_set_fpscr helper so the value would read back correctly but
not actually take effect.

Fix both of these things by doing a complete write to the FPSCR
using the helper function.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201210201433.26262-3-peter.maydell@linaro.org

hw/intc/armv7m_nvic: Correct handling of CCR.BFHFNMIGN

The CCR is a register most of whose bits are banked between security
states but where BFHFNMIGN is not, and we keep it in the non-secure
entry of the v7m.ccr[] array.  The logic which tries to handle this
bit fails to implement the "RAZ/WI from Nonsecure if AIRCR.BFHFNMINS
is zero" requirement; correct the omission.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201210201433.26262-2-peter.maydell@linaro.org

target/arm: Fix MTE0_ACTIVE

In 50244cc76abc we updated mte_check_fail to match the ARM
pseudocode, using the correct EL to select the TCF field.
But we failed to update MTE0_ACTIVE the same way, which led
to g_assert_not_reached().

Cc: qemu-stable@nongnu.org
Buglink: https://bugs.launchpad.net/bugs/1907137
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201221204426.88514-1-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/virt: Remove virt machine state 'smp_cpus'

virt machine's 'smp_cpus' and machine->smp.cpus must always have the
same value. And, anywhere we have virt machine state we have machine
state. So let's remove the redundancy. Also, to make it easier to see
that machine->smp is the true source for "smp_cpus" and "max_cpus",
avoid passing them in function parameters, preferring instead to get
them from the state.

No functional change intended.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: David Edmondson <david.edmondson@oracle.com>
Reviewed-by: Ying Fang <fangying1@huawei.com>
Message-id: 20201215174815.51520-1-drjones@redhat.com
[PMM: minor formatting tweak to smp_cpus variable declaration]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

intc/arm_gic: Fix gic_irq_signaling_enabled() for vCPUs

Correct the indexing into s->cpu_ctlr for vCPUs.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Luc Michel <luc.michel@greensocs.com>
Message-id: 20201214222154.3480243-2-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/rth-gitlab/tags/pull-tcg-20210107' into staging

Build fix for ppc64 centos7.
Reduce the use of scratch registers for tcg/i386.
Use _aligned_malloc for Win32.
Enable split w^x code gen buffers.

# gpg: Signature made Thu 07 Jan 2021 20:06:38 GMT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* remotes/rth-gitlab/tags/pull-tcg-20210107: (47 commits)
  tcg: Constify TCGLabelQemuLdst.raddr
  tcg: Constify tcg_code_gen_epilogue
  tcg: Remove TCG_TARGET_SUPPORT_MIRROR
  tcg/arm: Support split-wx code generation
  tcg/mips: Support split-wx code generation
  tcg/mips: Do not assert on relocation overflow
  accel/tcg: Add mips support to alloc_code_gen_buffer_splitwx_memfd
  tcg/riscv: Support split-wx code generation
  tcg/riscv: Remove branch-over-branch fallback
  tcg/riscv: Fix branch range checks
  tcg/s390: Support split-wx code generation
  tcg/s390: Use tcg_tbrel_diff
  tcg/sparc: Support split-wx code generation
  tcg/sparc: Use tcg_tbrel_diff
  tcg/ppc: Support split-wx code generation
  tcg/ppc: Use tcg_out_mem_long to reset TCG_REG_TB
  tcg/ppc: Use tcg_tbrel_diff
  tcg: Introduce tcg_tbrel_diff
  tcg/tci: Push const down through bytecode reading
  disas: Push const down through host disassembly
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

tcg: Constify TCGLabelQemuLdst.raddr

Now that all native tcg hosts support splitwx,
make this pointer const.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Constify tcg_code_gen_epilogue

Now that all native tcg hosts support splitwx,
make this pointer const.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Remove TCG_TARGET_SUPPORT_MIRROR

Now that all native tcg hosts support splitwx, remove the define.
Replace the one use with a test for CONFIG_TCG_INTERPRETER.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/arm: Support split-wx code generation

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/mips: Support split-wx code generation

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/mips: Do not assert on relocation overflow

This target was not updated with 7ecd02a06f8, and so did
not allow re-compilation with relocation overflow.

Remove reloc_26 and reloc_26_val as unused.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Add mips support to alloc_code_gen_buffer_splitwx_memfd

Re-use the 256MiB region handling from alloc_code_gen_buffer_anon,
and replace that with the shared file mapping.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/riscv: Support split-wx code generation

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/riscv: Remove branch-over-branch fallback

Since 7ecd02a06f8, we are prepared to re-start code generation
with a smaller TB if a relocation is out of range.  We no longer
need to leave a nop in the stream Just In Case.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/riscv: Fix branch range checks

The offset even checks were folded into the range check incorrectly.
By offsetting by 1, and not decrementing the width, we silently
allowed out of range branches.

Assert that the offset is always even instead.  Move tcg_out_goto
down into the CONFIG_SOFTMMU block so that it is not unused.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/s390: Support split-wx code generation

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/s390: Use tcg_tbrel_diff

Use tcg_tbrel_diff when we need a displacement to a label,
and with a NULL argument when we need the normalizing addend.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/sparc: Support split-wx code generation

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/sparc: Use tcg_tbrel_diff

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/ppc: Support split-wx code generation

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/ppc: Use tcg_out_mem_long to reset TCG_REG_TB

The maximum TB code gen size is UINT16_MAX, which the current
code does not support.  Use our utility function to optimally
add an arbitrary constant.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/ppc: Use tcg_tbrel_diff

Use tcg_tbrel_diff when we need a displacement to a label,
and with a NULL argument when we need the normalizing addend.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Introduce tcg_tbrel_diff

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/tci: Push const down through bytecode reading

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

disas: Push const down through host disassembly

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/aarch64: Support split-wx code generation

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/aarch64: Use B not BL for tcg_out_goto_long

A typo generated a branch-and-link insn instead of plain branch.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Support split-wx code generation

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Return the TB pointer from the rx region from exit_tb

This produces a small pc-relative displacement within the
generated code to the TB structure that preceeds it.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Support split-wx for darwin/iOS with vm_remap

Cribbed from code posted by Joelle van Dyne <j@getutm.app>,
and rearranged to a cleaner structure.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Support split-wx for linux with memfd

We cannot use a real temp file, because we would need to find
a filesystem that does not have noexec enabled.  However, a
memfd is not associated with any filesystem.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Add --accel tcg,split-wx property

Plumb the value through to alloc_code_gen_buffer.  This is not
supported by any os or tcg backend, so for now enabling it will
result in an error.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Use Error with alloc_code_gen_buffer

Report better error messages than just "could not allocate".
Let alloc_code_gen_buffer set ctx->code_gen_buffer_size
and ctx->code_gen_buffer, and simply return bool.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Make tb arg to synchronize_from_tb const

There is nothing within the translators that ought to be
changing the TranslationBlock data, so make it const.

This does not actually use the read-only copy of the
data structure that exists within the rx region.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Make DisasContextBase.tb const

There is nothing within the translators that ought to be
changing the TranslationBlock data, so make it const.

This does not actually use the read-only copy of the
data structure that exists within the rx region.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Adjust tb_target_set_jmp_target for split-wx

Pass both rx and rw addresses to tb_target_set_jmp_target.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Adjust tcg_register_jit for const

We must change all targets at once, since all must match
the declaration in tcg.c.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Adjust tcg_out_label for const

Simplify the arguments to always use s->code_ptr instead of
take it as an argument.  That makes it easy to ensure that
the value_ptr is always the rx version.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Adjust tcg_out_call for const

We must change all targets at once, since all must match
the declaration in tcg.c.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Adjust TCGLabel for const

Change TCGLabel.u.value_ptr to const, and initialize it with
tcg_splitwx_to_rx.  Propagate const through tcg/host/ only
as far as needed to avoid errors from the value_ptr change.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Introduce tcg_splitwx_to_{rx,rw}

Add two helper functions, using a global variable to hold
the displacement.  The displacement is currently always 0,
so no change in behaviour.

Begin using the functions in tcg common code only.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Add in_code_gen_buffer

Create a function to determine if a pointer is within the buffer.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Move tcg epilogue pointer out of TCGContext

This value is constant across all thread-local copies of TCGContext,
so we might as well move it out of thread-local storage.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Move tcg prologue pointer out of TCGContext

This value is constant across all thread-local copies of TCGContext,
so we might as well move it out of thread-local storage.

Use the correct function pointer type, and name the variable
tcg_qemu_tb_exec, which means that we are able to remove the
macro that does the casting.

Replace HAVE_TCG_QEMU_TB_EXEC with CONFIG_TCG_INTERPRETER,
as this is somewhat clearer in intent.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

util: Specialize flush_idcache_range for aarch64

For darwin, the CTR_EL0 register is not accessible, but there
are system routines that we can use.

For other hosts, copy the single pointer implementation from
libgcc and modify it to support the double pointer interface
we require.  This halves the number of cache operations required
when split-rwx is enabled.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

util: Enhance flush_icache_range with separate data pointer

We are shortly going to have a split rw/rx jit buffer.  Depending
on the host, we need to flush the dcache at the rw data pointer and
flush the icache at the rx code pointer.

For now, the two passed pointers are identical, so there is no
effective change in behaviour.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Do not flush icache for interpreter

This is currently a no-op within tci/tcg-target.h, but
is about to be moved to a more generic location.

Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

util/oslib: Assert qemu_try_memalign() alignment is a power of 2

qemu_try_memalign() expects a power of 2 alignment:

- posix_memalign(3):

  The address of the allocated memory will be a multiple of alignment,
  which must be a power of two and a multiple of sizeof(void *).

- _aligned_malloc()

  The alignment value, which must be an integer power of 2.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20201021173803.2619054-3-philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

util/oslib-win32: Use _aligned_malloc for qemu_try_memalign

We do not need or want to be allocating page sized quanta.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Message-Id: <20201018164836.1149452-1-richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Introduce INDEX_op_qemu_st8_i32

Enable this on i386 to restrict the set of input registers
for an 8-bit store, as required by the architecture.  This
removes the last use of scratch registers for user-only mode.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Adjust TCG_TARGET_HAS_MEMORY_BSWAP

Always true when movbe is available, otherwise leave
this to generic code.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Conditionalize TUNSETVNETLE

This fixes the build for older ppc64 kernel headers.

Fixes: 6addf06a3c4
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge remote-tracking branch 'remotes/stsquad/tags/pull-testing-060121-4' into staging

Testing updates (back to green)

  - include ccache in Debian 10 docker image
  - iotests: drop 312 from auto group
  - bound reading of s390x framebuffer file
  - cirrus: drop non-x86 tests so we complete

# gpg: Signature made Wed 06 Jan 2021 17:31:14 GMT
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [full]
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* remotes/stsquad/tags/pull-testing-060121-4:
  cirrus: don't run full qtest on macOS
  tests/acceptance: bound the size of readline in s390_ccw_virtio
  tests/iotests: drop test 312 from auto group
  tests/docker: Include 'ccache' in Debian base image

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

cirrus: don't run full qtest on macOS

The Cirrus CI macOS build hosts have exhibited a serious performance
degradation in recent months. For example the "qom-test" qtest takes
over an hour for only the qemu-system-aarch64 binary. This is as much
20-40 times slower than other environments. The other qtests all show
similar performance degradation, as do many of the unit tests.

This does not appear related to QEMU code changes, since older git
commits which were known to fully complete in less than 1 hour on
Cirrus CI now also show similar bad performance. Either Cirrus CI
performance has degraded, or an change in its environment has exposed
a latent bug widely affecting all of QEMU. Debugging the qom-test
showed no easily identified large bottleneck - every step of the
test execution was simply slower.

macOS builds/tests run outside Cirrus CI show normal performance.

With an inability to identify any obvious problem, the only viable
way to get a reliably completing Cirrus CI macOS job is to cut out
almost all of the qtests. We choose to run the x86_64 target only,
since that has very few machine types and thus is least badly
impacted in the qom-test execution.

With this change, the macOS jobs complete in approx 35 minutes.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Message-Id: <20210106114159.981538-1-berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>

tests/acceptance: bound the size of readline in s390_ccw_virtio

The read binary data as text via a PPM export of the frame buffer
seems a bit sketchy and it did blow up in the real world when the
assertion failed:

  https://gitlab.com/qemu-project/qemu/-/jobs/943183183

However short of cleaning up the test to be more binary focused at
least limit the attempt to dump the whole file as hexified zeros in
the logs.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Acked-by: Halil Pasic <pasic@linux.ibm.com>
Acked-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20210105124405.15424-1-alex.bennee@linaro.org>

tests/iotests: drop test 312 from auto group

The "auto" documentation states:

  That means they should run with every QEMU binary (also non-x86)

which is not the case as the check-system-fedora build which only
includes a rag tag group of rare and deprecated targets doesn't
support the virtio device required.

Fixes: ef9bba1484b ("quorum: Implement bdrv_co_block_status()")
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210105100402.12350-1-alex.bennee@linaro.org>

tests/docker: Include 'ccache' in Debian base image

Include the 'ccache' package to speed up compilation.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20201213211601.253530-1-f4bug@amsat.org>
Fixes: d6db2a1cdf ("docker: add debian-buster-arm64-cross")
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>

Merge remote-tracking branch 'remotes/bonzini-gitlab/tags/for-upstream' into staging

From Alex's pull request:
* improve cross-build KVM coverage
* new --without-default-features configure flag
* add __repr__ for ConsoleSocket for debugging
* build tcg tests with -Werror
* test 32 bit builds with fedora
* remove last traces of debian9
* hotfix for centos8 powertools repo

* Move lots of feature detection code to meson (Alex, myself)
* CFI and LTO support (Daniele)
* test-char dangling pointer (Eduardo)
* Build system and win32 fixes (Marc-André)
* Initialization fixes (myself)
* TCG include cleanup (Richard, myself)
* x86 'int N' fix (Peter)

# gpg: Signature made Wed 06 Jan 2021 09:21:25 GMT
# 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

* remotes/bonzini-gitlab/tags/for-upstream: (52 commits)
  win32: drop fd registration to the main-loop on setting non-block
  configure: move tests/qemu-iotests/common.env generation to meson
  meson.build: convert --with-default-devices to meson
  libattr: convert to meson
  cap_ng: convert to meson
  virtfs: convert to meson
  seccomp: convert to meson
  zstd: convert to meson
  lzfse: convert to meson
  snappy: convert to meson
  lzo: convert to meson
  rbd: convert to meson
  libnfs: convert to meson
  libiscsi: convert to meson
  bzip2: convert to meson
  glusterfs: convert to meson
  curl: convert to meson
  curl: remove compatibility code, require 7.29.0
  brlapi: convert to meson
  configure: remove CONFIG_FILEVERSION and CONFIG_PRODUCTVERSION
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
# Conflicts:
# trace/meson.build