OSDN Git Service

qmiga/qemu.git
2 years agoMerge remote-tracking branch 'remotes/kwolf-gitlab/tags/for-upstream' into staging
Peter Maydell [Mon, 14 Feb 2022 19:54:00 +0000 (19:54 +0000)]
Merge remote-tracking branch 'remotes/kwolf-gitlab/tags/for-upstream' into staging

Block layer patches

- Fix crash in blockdev-reopen with iothreads
- fdc-isa: Respect QOM properties when building AML

# gpg: Signature made Fri 11 Feb 2022 17:44:52 GMT
# gpg:                using RSA key DC3DEB159A9AF95D3D7456FE7F09B272C88F2FD6
# gpg:                issuer "kwolf@redhat.com"
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>" [full]
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* remotes/kwolf-gitlab/tags/for-upstream:
  hw/block/fdc-isa: Respect QOM properties when building AML
  iotests: Test blockdev-reopen with iothreads and throttling
  block: Lock AioContext for drain_end in blockdev-reopen

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2 years agoMerge remote-tracking branch 'remotes/rth-gitlab/tags/pull-tcg-20220211' into staging
Peter Maydell [Mon, 14 Feb 2022 15:24:26 +0000 (15:24 +0000)]
Merge remote-tracking branch 'remotes/rth-gitlab/tags/pull-tcg-20220211' into staging

Fix safe_syscall_base for sparc64.
Fix host signal handling for sparc64-linux.
Speedups for jump cache and work list probing.
Fix for exception replays.
Raise guest SIGBUS for user-only misaligned accesses.

# gpg: Signature made Fri 11 Feb 2022 01:27:16 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-20220211: (34 commits)
  tests/tcg/multiarch: Add sigbus.c
  tcg/sparc: Support unaligned access for user-only
  tcg/sparc: Add tcg_out_jmpl_const for better tail calls
  tcg/sparc: Use the constant pool for 64-bit constants
  tcg/sparc: Convert patch_reloc to return bool
  tcg/sparc: Improve code gen for shifted 32-bit constants
  tcg/sparc: Add scratch argument to tcg_out_movi_int
  tcg/sparc: Split out tcg_out_movi_imm32
  tcg/sparc: Use tcg_out_movi_imm13 in tcg_out_addsub2_i64
  tcg/mips: Support unaligned access for softmmu
  tcg/mips: Support unaligned access for user-only
  tcg/arm: Support raising sigbus for user-only
  tcg/arm: Reserve a register for guest_base
  tcg/arm: Support unaligned access for softmmu
  tcg/arm: Check alignment for ldrd and strd
  tcg/arm: Remove use_armv6_instructions
  tcg/arm: Remove use_armv5t_instructions
  tcg/arm: Drop support for armv4 and armv5 hosts
  tcg/loongarch64: Support raising sigbus for user-only
  tcg/tci: Support raising sigbus for user-only
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2 years agoMerge remote-tracking branch 'remotes/legoater/tags/pull-ppc-20220210' into staging
Peter Maydell [Sun, 13 Feb 2022 20:33:28 +0000 (20:33 +0000)]
Merge remote-tracking branch 'remotes/legoater/tags/pull-ppc-20220210' into staging

ppc-7.0 queue

* Exception model rework (Fabiano)
* Unused CPU models removal (Fabiano and Cédric)
* Fix for VOF installation (Alexey)
* Misc fixes

# gpg: Signature made Thu 10 Feb 2022 12:59:07 GMT
# gpg:                using RSA key A0F66548F04895EBFE6B0B6051A343C7CFFBECA1
# gpg: Good signature from "Cédric Le Goater <clg@kaod.org>" [undefined]
# 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: A0F6 6548 F048 95EB FE6B  0B60 51A3 43C7 CFFB ECA1

* remotes/legoater/tags/pull-ppc-20220210: (42 commits)
  spapr/vof: Install rom and nvram binaries
  docs: rstfy confidential guest documentation
  target/ppc: Change VSX instructions behavior to fill with zeros
  target/ppc: books: Remove excp_model argument from ppc_excp_apply_ail
  target/ppc: Assert if MSR bits differ from msr_mask during exceptions
  target/ppc: powerpc_excp: Move common code to the caller function
  target/ppc: Remove powerpc_excp_legacy
  target/ppc: 7xx: Set SRRs directly in exception code
  target/ppc: 7xx: Software TLB cleanup
  target/ppc: 7xx: System Reset cleanup
  target/ppc: 7xx: System Call exception cleanup
  target/ppc: 7xx: Program exception cleanup
  target/ppc: 7xx: External interrupt cleanup
  target/ppc: 7xx: Machine Check exception cleanup
  target/ppc: Simplify powerpc_excp_7xx
  target/ppc: Introduce powerpc_excp_7xx
  target/ppc: Merge 7x5 and 7x0 exception model IDs
  target/ppc: 6xx: Set SRRs directly in exception code
  target/ppc: 6xx: Software TLB exceptions cleanup
  target/ppc: 6xx: System Reset interrupt cleanup
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2 years agoMerge remote-tracking branch 'remotes/vsementsov/tags/pull-nbd-2022-02-09-v2' into...
Peter Maydell [Sat, 12 Feb 2022 22:04:07 +0000 (22:04 +0000)]
Merge remote-tracking branch 'remotes/vsementsov/tags/pull-nbd-2022-02-09-v2' into staging

nbd: handle AioContext change correctly

v2: add my s-o-b marks to each commit

# gpg: Signature made Fri 11 Feb 2022 13:14:55 GMT
# gpg:                using RSA key 8B9C26CDB2FD147C880E86A1561F24C1F19F79FB
# gpg: Good signature from "Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.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: 8B9C 26CD B2FD 147C 880E  86A1 561F 24C1 F19F 79FB

* remotes/vsementsov/tags/pull-nbd-2022-02-09-v2:
  iotests/281: Let NBD connection yield in iothread
  block/nbd: Move s->ioc on AioContext change
  iotests/281: Test lingering timers
  iotests.py: Add QemuStorageDaemon class
  block/nbd: Assert there are no timers when closed
  block/nbd: Delete open timer when done
  block/nbd: Delete reconnect delay timer when done

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2 years agohw/block/fdc-isa: Respect QOM properties when building AML
Bernhard Beschow [Wed, 9 Feb 2022 19:15:58 +0000 (20:15 +0100)]
hw/block/fdc-isa: Respect QOM properties when building AML

Other ISA devices such as serial-isa use the properties in their
build_aml functions. fdc-isa not using them is probably an oversight.

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Message-Id: <20220209191558.30393-1-shentey@gmail.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2 years agoMerge remote-tracking branch 'remotes/stsquad/tags/pull-testing-and-plugins-090222...
Peter Maydell [Fri, 11 Feb 2022 13:11:49 +0000 (13:11 +0000)]
Merge remote-tracking branch 'remotes/stsquad/tags/pull-testing-and-plugins-090222-1' into staging

Testing and plugin updates:

  - include vhost tests in qtest
  - clean-up gcov ephemera in clean/.gitignore
  - lcitool and docker updates
  - mention .editorconfig in devel notes
  - switch Centos8 to Centos Stream 8
  - remove TCG tracing support
  - add coverage plugin using drcov format
  - expand abilities of libinsn.so plugin
  - use correct logging for i386 int cases
  - move reset of plugin data to start of block
  - deprecate ppc6432abi
  - fix TARGET_ABI_FMT_ptr for softmmu builds

# gpg: Signature made Wed 09 Feb 2022 14:13:14 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-and-plugins-090222-1: (28 commits)
  include/exec: fix softmmu version of TARGET_ABI_FMT_lx
  linux-user: Remove the deprecated ppc64abi32 target
  plugins: move reset of plugin data to tb_start
  target/i386: use CPU_LOG_INT for IRQ servicing
  tests/plugins: add instruction matching to libinsn.so
  tests/plugin: allow libinsn.so per-CPU counts
  contrib/plugins: add a drcov plugin
  plugins: add helper functions for coverage plugins
  tracing: excise the tcg related from tracetool
  tracing: remove the trace-tcg includes from the build
  tracing: remove TCG memory access tracing
  docs: remove references to TCG tracing
  tests/tcg/sh4: disable another unreliable test
  tests: Update CentOS 8 container to CentOS Stream 8
  tests/lcitool: Allow lcitool-refresh in out-of-tree builds, too
  gitlab: fall back to commit hash in qemu-setup filename
  docs/devel: mention our .editorconfig
  tests/lcitool: Install libibumad to cover RDMA on Debian based distros
  tests: Manually remove libxml2 on MSYS2 runners
  tests/lcitool: Refresh submodule and remove libxml2
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2 years agoiotests/281: Let NBD connection yield in iothread
Hanna Reitz [Fri, 4 Feb 2022 11:10:12 +0000 (12:10 +0100)]
iotests/281: Let NBD connection yield in iothread

Put an NBD block device into an I/O thread, and then read data from it,
hoping that the NBD connection will yield during that read.  When it
does, the coroutine must be reentered in the block device's I/O thread,
which will only happen if the NBD block driver attaches the connection's
QIOChannel to the new AioContext.  It did not do that after 4ddb5d2fde
("block/nbd: drop connection_co") and prior to "block/nbd: Move s->ioc
on AioContext change", which would cause an assertion failure.

To improve our chances of yielding, the NBD server is throttled to
reading 64 kB/s, and the NBD client reads 128 kB, so it should yield at
some point.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
2 years agoblock/nbd: Move s->ioc on AioContext change
Hanna Reitz [Fri, 4 Feb 2022 11:10:11 +0000 (12:10 +0100)]
block/nbd: Move s->ioc on AioContext change

s->ioc must always be attached to the NBD node's AioContext.  If that
context changes, s->ioc must be attached to the new context.

Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=2033626
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
2 years agoiotests/281: Test lingering timers
Hanna Reitz [Fri, 4 Feb 2022 11:10:10 +0000 (12:10 +0100)]
iotests/281: Test lingering timers

Prior to "block/nbd: Delete reconnect delay timer when done" and
"block/nbd: Delete open timer when done", both of those timers would
remain scheduled even after successfully (re-)connecting to the server,
and they would not even be deleted when the BDS is deleted.

This test constructs exactly this situation:
(1) Configure an @open-timeout, so the open timer is armed, and
(2) Configure a @reconnect-delay and trigger a reconnect situation
    (which succeeds immediately), so the reconnect delay timer is armed.
Then we immediately delete the BDS, and sleep for longer than the
@open-timeout and @reconnect-delay.  Prior to said patches, this caused
one (or both) of the timer CBs to access already-freed data.

Accessing freed data may or may not crash, so this test can produce
false successes, but I do not know how to show the problem in a better
or more reliable way.  If you run this test on "block/nbd: Assert there
are no timers when closed" and without the fix patches mentioned above,
you should reliably see an assertion failure.
(But all other tests that use the reconnect delay timer (264 and 277)
will fail in that configuration, too; as will nbd-reconnect-on-open,
which uses the open timer.)

Remove this test from the quick group because of the two second sleep
this patch introduces.

(I decided to put this test case into 281, because the main bug this
series addresses is in the interaction of the NBD block driver and I/O
threads, which is precisely the scope of 281.  The test case for that
other bug will also be put into the test class added here.

Also, excuse the test class's name, I couldn't come up with anything
better.  The "yield" part will make sense two patches from now.)

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
2 years agoiotests.py: Add QemuStorageDaemon class
Hanna Reitz [Fri, 4 Feb 2022 11:10:09 +0000 (12:10 +0100)]
iotests.py: Add QemuStorageDaemon class

This is a rather simple class that allows creating a QSD instance
running in the background and stopping it when no longer needed.

The __del__ handler is a safety net for when something goes so wrong in
a test that e.g. the tearDown() method is not called (e.g. setUp()
launches the QSD, but then launching a VM fails).  We do not want the
QSD to continue running after the test has failed, so __del__() will
take care to kill it.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
2 years agoblock/nbd: Assert there are no timers when closed
Hanna Reitz [Fri, 4 Feb 2022 11:10:08 +0000 (12:10 +0100)]
block/nbd: Assert there are no timers when closed

Our two timers must not remain armed beyond nbd_clear_bdrvstate(), or
they will access freed data when they fire.

This patch is separate from the patches that actually fix the issue
(HEAD^^ and HEAD^) so that you can run the associated regression iotest
(281) on a configuration that reproducibly exposes the bug.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
2 years agoblock/nbd: Delete open timer when done
Hanna Reitz [Fri, 4 Feb 2022 11:10:07 +0000 (12:10 +0100)]
block/nbd: Delete open timer when done

We start the open timer to cancel the connection attempt after a while.
Once nbd_do_establish_connection() has returned, the attempt is over,
and we no longer need the timer.

Delete it before returning from nbd_open(), so that it does not persist
for longer.  It has no use after nbd_open(), and just like the reconnect
delay timer, it might well be dangerous if it were to fire afterwards.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
2 years agoblock/nbd: Delete reconnect delay timer when done
Hanna Reitz [Fri, 4 Feb 2022 11:10:06 +0000 (12:10 +0100)]
block/nbd: Delete reconnect delay timer when done

We start the reconnect delay timer to cancel the reconnection attempt
after a while.  Once nbd_co_do_establish_connection() has returned, this
attempt is over, and we no longer need the timer.

Delete it before returning from nbd_reconnect_attempt(), so that it does
not persist beyond the I/O request that was paused for reconnecting; we
do not want it to fire in a drained section, because all sort of things
can happen in such a section (e.g. the AioContext might be changed, and
we do not want the timer to fire in the wrong context; or the BDS might
even be deleted, and so the timer CB would access already-freed data).

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
2 years agoiotests: Test blockdev-reopen with iothreads and throttling
Kevin Wolf [Thu, 3 Feb 2022 14:05:34 +0000 (15:05 +0100)]
iotests: Test blockdev-reopen with iothreads and throttling

The 'throttle' block driver implements .bdrv_co_drain_end, so
blockdev-reopen will have to wait for it to complete in the polling
loop at the end of qmp_blockdev_reopen(). This makes AIO_WAIT_WHILE()
release the AioContext lock, which causes a crash if the lock hasn't
correctly been taken.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20220203140534.36522-3-kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2 years agoblock: Lock AioContext for drain_end in blockdev-reopen
Kevin Wolf [Thu, 3 Feb 2022 14:05:33 +0000 (15:05 +0100)]
block: Lock AioContext for drain_end in blockdev-reopen

bdrv_subtree_drained_end() requires the caller to hold the AioContext
lock for the drained node. Not doing this for nodes outside of the main
AioContext leads to crashes when AIO_WAIT_WHILE() needs to wait and
tries to temporarily release the lock.

Fixes: 3908b7a8994fa5ef7a89aa58cd5a02fc58141592
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2046659
Reported-by: Qing Wang <qinwang@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20220203140534.36522-2-kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2 years agoinclude/exec: fix softmmu version of TARGET_ABI_FMT_lx
Alex Bennée [Tue, 8 Feb 2022 12:43:33 +0000 (12:43 +0000)]
include/exec: fix softmmu version of TARGET_ABI_FMT_lx

TARGET_ABI_FMT_lx isn't available for softmmu which causes confusion
when trying to print. As abi_ptr == target_ulong use its format string
instead.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
2 years agolinux-user: Remove the deprecated ppc64abi32 target
Thomas Huth [Wed, 12 Jan 2022 11:27:22 +0000 (11:27 +0000)]
linux-user: Remove the deprecated ppc64abi32 target

It's likely broken, and nobody cared for picking it up again
during the deprecation phase, so let's remove this now.

Since this is the last entry in deprecated_targets_list, remove
the related code in the configure script, too.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Cédric Le Goater <clg@kaod.org>
Acked-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20211215084958.185214-1-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220112112722.3641051-32-alex.bennee@linaro.org>

2 years agoplugins: move reset of plugin data to tb_start
Alex Bennée [Fri, 4 Feb 2022 20:43:35 +0000 (20:43 +0000)]
plugins: move reset of plugin data to tb_start

We can't always guarantee we get to the end of a translator loop.
Although this can happen for a variety of reasons it does happen more
often on x86 system emulation when an instruction spans across to an
un-faulted page. This caused confusion of the instruction tracking
data resulting in apparent reverse execution (at least from the
plugins point of view).

Fix this by moving the reset code to plugin_gen_tb_start so we always
start with a clean slate.

We unconditionally reset tcg_ctx->plugin_insn as the
plugin_insn_append code uses this as a proxy for knowing if plugins
are enabled for the current instruction. Otherwise we can hit a race
where a previously instrumented thread leaves a stale value after the
main thread exits and disables instrumentation.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/824
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220204204335.1689602-27-alex.bennee@linaro.org>

2 years agotarget/i386: use CPU_LOG_INT for IRQ servicing
Alex Bennée [Fri, 4 Feb 2022 20:43:34 +0000 (20:43 +0000)]
target/i386: use CPU_LOG_INT for IRQ servicing

I think these have been wrong since f193c7979c (do not depend on
thunk.h - more log items). Fix them so as not to confuse other
debugging.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220204204335.1689602-26-alex.bennee@linaro.org>

2 years agotests/plugins: add instruction matching to libinsn.so
Alex Bennée [Fri, 4 Feb 2022 20:43:33 +0000 (20:43 +0000)]
tests/plugins: add instruction matching to libinsn.so

This adds simple instruction matching to the libinsn.so plugin which
is useful for examining the execution distance between instructions.
For example to track how often we flush in ARM due to TLB updates:

  -plugin ./tests/plugin/libinsn.so,match=tlbi

which leads to output like this:

  0xffffffc01019a918, 'tlbi vale1is, x1', 5702 hits, 31825 match hits, Δ+8112 since last match, 68859 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5703 hits, 56593 match hits, Δ+17712125 since last match, 33455 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5704 hits, 56594 match hits, Δ+12689 since last match, 33454 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5705 hits, 56595 match hits, Δ+12585 since last match, 33454 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5706 hits, 56596 match hits, Δ+10491 since last match, 33454 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5707 hits, 56597 match hits, Δ+4721 since last match, 33453 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5708 hits, 56598 match hits, Δ+10733 since last match, 33453 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5709 hits, 56599 match hits, Δ+61959 since last match, 33453 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5710 hits, 56600 match hits, Δ+55235 since last match, 33454 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5711 hits, 56601 match hits, Δ+54373 since last match, 33454 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5712 hits, 56602 match hits, Δ+2705 since last match, 33453 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5713 hits, 56603 match hits, Δ+17262 since last match, 33453 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5714 hits, 56604 match hits, Δ+17206 since last match, 33453 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5715 hits, 56605 match hits, Δ+28940 since last match, 33453 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5716 hits, 56606 match hits, Δ+7370 since last match, 33452 avg insns/match
  0xffffffc01019a918, 'tlbi vale1is, x1', 5717 hits, 56607 match hits, Δ+7066 since last match, 33452 avg insns/match

showing we do some sort of TLBI invalidation every 33 thousand
instructions.

Cc: Vasilev Oleg <vasilev.oleg@huawei.com>
Cc: Richard Henderson <richard.henderson@linaro.org>
Cc: Emilio Cota <cota@braap.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220204204335.1689602-25-alex.bennee@linaro.org>

2 years agotests/plugin: allow libinsn.so per-CPU counts
Alex Bennée [Fri, 4 Feb 2022 20:43:32 +0000 (20:43 +0000)]
tests/plugin: allow libinsn.so per-CPU counts

We won't go fully flexible but for most system emulation 8 vCPUs
resolution should be enough for anybody ;-)

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220204204335.1689602-24-alex.bennee@linaro.org>

2 years agocontrib/plugins: add a drcov plugin
Ivanov Arkady [Fri, 4 Feb 2022 20:43:31 +0000 (20:43 +0000)]
contrib/plugins: add a drcov plugin

This patch adds the ability to generate files in drcov format. Primary
goal this script is to have coverage logfiles thatwork in Lighthouse.

Signed-off-by: Ivanov Arkady <arkadiy.ivanov@ispras.ru>
Message-Id: <163491884553.304355.13246023070235438959.stgit@pc-System-Product-Name>
[AJB: use g_ptr_array instead of slist]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220204204335.1689602-23-alex.bennee@linaro.org>

2 years agoplugins: add helper functions for coverage plugins
Ivanov Arkady [Fri, 4 Feb 2022 20:43:30 +0000 (20:43 +0000)]
plugins: add helper functions for coverage plugins

Which provide information about:
- start_code.
- end_code.
- entry.
- path to the executable binary.

Signed-off-by: Ivanov Arkady <arkadiy.ivanov@ispras.ru>
Message-Id: <163491883461.304355.8210754161847179432.stgit@pc-System-Product-Name>
[AJB: reword title, better descriptions, defaults, rm export, fix include]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220204204335.1689602-22-alex.bennee@linaro.org>

2 years agotracing: excise the tcg related from tracetool
Alex Bennée [Fri, 4 Feb 2022 20:43:29 +0000 (20:43 +0000)]
tracing: excise the tcg related from tracetool

Now we have no TCG trace events and no longer handle them in the code
we can remove the handling from the tracetool to generate them. vcpu
tracing is still available although the existing syscall event is an
exercise in redundancy (plugins and -strace can also get the
information).

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Cc: Luis Vilanova <vilanova@imperial.ac.uk>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20220204204335.1689602-21-alex.bennee@linaro.org>

2 years agotracing: remove the trace-tcg includes from the build
Alex Bennée [Fri, 4 Feb 2022 20:43:28 +0000 (20:43 +0000)]
tracing: remove the trace-tcg includes from the build

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Cc: Luis Vilanova <vilanova@imperial.ac.uk>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220204204335.1689602-20-alex.bennee@linaro.org>

2 years agotracing: remove TCG memory access tracing
Alex Bennée [Fri, 4 Feb 2022 20:43:27 +0000 (20:43 +0000)]
tracing: remove TCG memory access tracing

If you really want to trace all memory operations TCG plugins gives
you a more flexible interface for doing so.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Cc: Luis Vilanova <vilanova@imperial.ac.uk>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20220204204335.1689602-19-alex.bennee@linaro.org>

2 years agodocs: remove references to TCG tracing
Alex Bennée [Fri, 4 Feb 2022 20:43:26 +0000 (20:43 +0000)]
docs: remove references to TCG tracing

Users wanting this sort of functionality should turn to TCG plugins
instead.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Cc: Luis Vilanova <vilanova@imperial.ac.uk>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20220204204335.1689602-18-alex.bennee@linaro.org>

2 years agotests/tcg/sh4: disable another unreliable test
Alex Bennée [Fri, 4 Feb 2022 20:43:25 +0000 (20:43 +0000)]
tests/tcg/sh4: disable another unreliable test

Given the other failures it looks like general thread handling on sh4
is sketchy. It fails more often on CI than on my developer machine
though. See https://gitlab.com/qemu-project/qemu/-/issues/856 for more
details.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Laurent Vivier <laurent@vivier.eu>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220204204335.1689602-17-alex.bennee@linaro.org>

2 years agotests: Update CentOS 8 container to CentOS Stream 8
Thomas Huth [Fri, 4 Feb 2022 20:43:24 +0000 (20:43 +0000)]
tests: Update CentOS 8 container to CentOS Stream 8

Support for CentOS 8 has stopped at the end of 2021, so let's
switch to the Stream variant instead.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220201101911.97900-1-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220204204335.1689602-16-alex.bennee@linaro.org>

2 years agotests/lcitool: Allow lcitool-refresh in out-of-tree builds, too
Thomas Huth [Fri, 4 Feb 2022 20:43:23 +0000 (20:43 +0000)]
tests/lcitool: Allow lcitool-refresh in out-of-tree builds, too

When running "make lcitool-refresh" in an out-of-tree build, it
currently fails with an error message from git like this:

 fatal: not a git repository (or any parent up to mount point /)
 Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).

Fix it by changing to the source directory first before updating
the submodule.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220201085554.85733-1-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220204204335.1689602-15-alex.bennee@linaro.org>

2 years agogitlab: fall back to commit hash in qemu-setup filename
Stefan Hajnoczi [Fri, 4 Feb 2022 20:43:22 +0000 (20:43 +0000)]
gitlab: fall back to commit hash in qemu-setup filename

Personal repos may not have release tags (v6.0.0, v6.1.0, etc) and this
causes cross_system_build_job to fail when pretty-printing a unique
qemu-setup-*.exe name:

  version="$(git describe --match v[0-9]*)";
             ^^^^^^^^^^ fails ^^^^^^^^^^^
  mv -v qemu-setup*.exe qemu-setup-${version}.exe;

Fall back to the short commit hash if necessary. This fixes CI failures
that Greg Kurz and I experienced in our personal repos.

Cc: Greg Kurz <groug@kaod.org>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220125173454.10381-1-stefanha@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220204204335.1689602-14-alex.bennee@linaro.org>

2 years agodocs/devel: mention our .editorconfig
Alex Bennée [Fri, 4 Feb 2022 20:43:21 +0000 (20:43 +0000)]
docs/devel: mention our .editorconfig

Ideally we should keep all our automatic formatting gubins in here.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220204204335.1689602-13-alex.bennee@linaro.org>

2 years agotests/lcitool: Install libibumad to cover RDMA on Debian based distros
Philippe Mathieu-Daudé [Fri, 4 Feb 2022 20:43:20 +0000 (20:43 +0000)]
tests/lcitool: Install libibumad to cover RDMA on Debian based distros

On Debian we also need libibumad to enable RDMA:

  $ ../configure --enable-rdma

  ERROR:  OpenFabrics librdmacm/libibverbs/libibumad not present.
          Your options:
           (1) Fast: Install infiniband packages (devel) from your distro.
           (2) Cleanest: Install libraries from www.openfabrics.org
           (3) Also: Install softiwarp if you don't have RDMA hardware

Add the dependency to lcitool's qemu.yml (where librdmacm and
libibverbs are already listed) and refresh the generated files
by running:

      $ make lcitool-refresh

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220121154134.315047-8-f4bug@amsat.org>
Message-Id: <20220204204335.1689602-12-alex.bennee@linaro.org>

2 years agotests: Manually remove libxml2 on MSYS2 runners
Philippe Mathieu-Daudé [Fri, 4 Feb 2022 20:43:19 +0000 (20:43 +0000)]
tests: Manually remove libxml2 on MSYS2 runners

lcitool doesn't support MSYS2 targets, so manually remove
this now unnecessary library.

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220121154134.315047-7-f4bug@amsat.org>
Message-Id: <20220204204335.1689602-11-alex.bennee@linaro.org>

2 years agotests/lcitool: Refresh submodule and remove libxml2
Philippe Mathieu-Daudé [Fri, 4 Feb 2022 20:43:18 +0000 (20:43 +0000)]
tests/lcitool: Refresh submodule and remove libxml2

The previous commit removed all uses of libxml2.

Refresh lcitool submodule, update qemu.yml and refresh the generated
files by running:

  $ make lcitool-refresh

Note: This refreshment also removes libudev dependency on Fedora
and CentOS due to libvirt-ci commit 18bfaee ("mappings: Improve
mapping for libudev"), since "The udev project has been absorbed
by the systemd project", and lttng-ust on FreeBSD runners due to
libvirt-ci commit 6dd9b6f ("guests: drop lttng-ust from FreeBSD
platform").

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220121154134.315047-6-f4bug@amsat.org>
Message-Id: <20220204204335.1689602-10-alex.bennee@linaro.org>

2 years agodrop libxml2 checks since libxml is not actually used (for parallels)
Michael Tokarev [Fri, 4 Feb 2022 20:43:17 +0000 (20:43 +0000)]
drop libxml2 checks since libxml is not actually used (for parallels)

For a long time, we assumed that libxml2 is necessary for parallels
block format support (block/parallels*). However, this format actually
does not use libxml [*]. Since this is the only user of libxml2 in
whole QEMU tree, we can drop all libxml2 checks and dependencies too.

It is even more: --enable-parallels configure option was the only
option which was silently ignored when it's (fake) dependency
(libxml2) isn't installed.

Drop all mentions of libxml2.

[*] Actually the basis for libxml use were introduced in commit
    ed279a06c53 ("configure: add dependency") but the implementation
    was never merged:
    https://lore.kernel.org/qemu-devel/70227bbd-a517-70e9-714f-e6e0ec431be9@openvz.org/

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20220119090423.149315-1-mjt@msgid.tls.msk.ru>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
[PMD: Updated description and adapted to use lcitool]
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220121154134.315047-5-f4bug@amsat.org>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20220204204335.1689602-9-alex.bennee@linaro.org>

2 years agotests/lcitool: Include local qemu.yml when refreshing cirrus-ci files
Philippe Mathieu-Daudé [Fri, 4 Feb 2022 20:43:16 +0000 (20:43 +0000)]
tests/lcitool: Include local qemu.yml when refreshing cirrus-ci files

The script only include the local qemu.yml for Dockerfiles.
Since we want to keep the Cirrus-CI generated files in sync,
also use the --data-dir option in generate_cirrus().

Fixes: c45a540f4bd (".gitlab-ci.d/cirrus: auto-generate variables with lcitool")
Reported-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220121154134.315047-4-f4bug@amsat.org>
Message-Id: <20220204204335.1689602-8-alex.bennee@linaro.org>

2 years agogitmodules: Correct libvirt-ci submodule URL
Philippe Mathieu-Daudé [Fri, 4 Feb 2022 20:43:15 +0000 (20:43 +0000)]
gitmodules: Correct libvirt-ci submodule URL

Correct the libvirt-ci repository URL to avoid this warning when
cloning / refreshing the submodule:

  warning: redirecting to https://gitlab.com/libvirt/libvirt-ci.git/

Fixes: 4ebb040f1fd ("tests: integrate lcitool for generating build env manifests")
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220121154134.315047-3-f4bug@amsat.org>
Message-Id: <20220204204335.1689602-7-alex.bennee@linaro.org>

2 years agoMAINTAINERS: Cover lcitool submodule with build test / automation
Philippe Mathieu-Daudé [Fri, 4 Feb 2022 20:43:14 +0000 (20:43 +0000)]
MAINTAINERS: Cover lcitool submodule with build test / automation

lcitool is used by build test / automation, we want maintainers
to get notified if the submodule is updated.

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220121154134.315047-2-f4bug@amsat.org>
Message-Id: <20220204204335.1689602-6-alex.bennee@linaro.org>

2 years ago.gitignore: add .gcov pattern
Alex Bennée [Fri, 4 Feb 2022 20:43:13 +0000 (20:43 +0000)]
.gitignore: add .gcov pattern

The gcovr tool is very messy and can leave a lot of crap in the source
tree even when using build directories.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220204204335.1689602-5-alex.bennee@linaro.org>

2 years agoMakefile: also remove .gcno files when cleaning
Alex Bennée [Fri, 4 Feb 2022 20:43:12 +0000 (20:43 +0000)]
Makefile: also remove .gcno files when cleaning

Left over .gcno files from old builds can really confuse gcov and the
user expects a clean slate after "make clean". Make clean mean clean.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220204204335.1689602-4-alex.bennee@linaro.org>

2 years agotests/qtest: enable more vhost-user tests by default
Alex Bennée [Fri, 4 Feb 2022 20:43:11 +0000 (20:43 +0000)]
tests/qtest: enable more vhost-user tests by default

If this starts causing failures again we should probably fix that.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220204204335.1689602-3-alex.bennee@linaro.org>

2 years agotests/Makefile.include: clean-up old code
Alex Bennée [Fri, 4 Feb 2022 20:43:10 +0000 (20:43 +0000)]
tests/Makefile.include: clean-up old code

This is no longer needed since a2ce7dbd91 ("meson: convert tests/qtest
to meson", 2020-08-21)

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220204204335.1689602-2-alex.bennee@linaro.org>

2 years agospapr/vof: Install rom and nvram binaries
Alexey Kardashevskiy [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
spapr/vof: Install rom and nvram binaries

This installs VOF-related binaries (the firmware and the preformatted
NVRAM) as those were left out when the VOF was submitted initially.

Fixes: fc8c745d5015 ("spapr: Implement Open Firmware client interface")
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Message-Id: <20220208103751.1587902-1-aik@ozlabs.ru>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agodocs: rstfy confidential guest documentation
Cornelia Huck [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
docs: rstfy confidential guest documentation

Also rstfy the documentation for AMD SEV, and link it.

The documentation for PEF had been merged into the pseries doc,
fix the reference.

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220204161251.241877-1-cohuck@redhat.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Change VSX instructions behavior to fill with zeros
Víctor Colombo [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: Change VSX instructions behavior to fill with zeros

ISA v3.1 changed some VSX instructions behavior by changing what the
other words/doubleword in the result should contain when the result is
only one word/doubleword. e.g. xsmaxdp operates on doubleword 0 and
saves the result also in doubleword 0.
Before, the second doubleword result was undefined according to the
ISA, but now it's stated that it should be zeroed.

Even tough the result was undefined before, hardware implementing these
instructions already filled these fields with 0s. Changing every ISA
version in QEMU to this behavior makes the results match what happens
in hardware.

Signed-off-by: Víctor Colombo <victor.colombo@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220204181944.65063-1-victor.colombo@eldorado.org.br>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: books: Remove excp_model argument from ppc_excp_apply_ail
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: books: Remove excp_model argument from ppc_excp_apply_ail

We don't really need to check for exception model while applying
AIL. We can check the lpcr_mask for the presence of
LPCR_AIL/LPCR_HAIL.

This removes one more instance of passing the exception model ID
around.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220207183036.1507882-5-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Assert if MSR bits differ from msr_mask during exceptions
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: Assert if MSR bits differ from msr_mask during exceptions

We currently abort QEMU during the dispatch of an interrupt if we try
to set MSR_HV without having MSR_HVB in the msr_mask. I think we
should verify this for all MSR bits. There is no reason to ever have a
MSR bit set if the corresponding bit is not set in that CPU's
msr_mask.

Note that this is not about the emulated code setting reserved
bits. We clear the new_msr when starting to dispatch an exception, so
if we end up with bits not present in the msr_mask that is a QEMU
programming error.

I kept the HSRR verification for BookS because it is the only CPU
family that has HSRRs.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220207183036.1507882-4-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: powerpc_excp: Move common code to the caller function
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: powerpc_excp: Move common code to the caller function

Make the cpu-specific powerpc_excp_* functions a bit simpler by moving
the bounds check and logging to powerpc_excp.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220207183036.1507882-3-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Remove powerpc_excp_legacy
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: Remove powerpc_excp_legacy

Now that all CPU families have their own separate exception
dispatching code we can remove powerpc_excp_legacy.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220207183036.1507882-2-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 7xx: Set SRRs directly in exception code
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: 7xx: Set SRRs directly in exception code

The 7xx CPUs don't have alternate/hypervisor Save and Restore
Registers, so we can set SRR0 and SRR1 directly.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-11-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 7xx: Software TLB cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: 7xx: Software TLB cleanup

This code applies only to the 7xx CPUs, so we can remove the switch
statement.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-10-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 7xx: System Reset cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: 7xx: System Reset cleanup

Thre is no HV support in the 7xx.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-9-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 7xx: System Call exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: 7xx: System Call exception cleanup

Remove the BookE code and add a comment explaining why we need to keep
hypercall support even though this CPU does not have a hypervisor
mode.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-8-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 7xx: Program exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: 7xx: Program exception cleanup

There's no ESR in the 7xx.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-7-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 7xx: External interrupt cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: 7xx: External interrupt cleanup

There is no MSR_HV in the 7xx so remove the LPES0 handling.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-6-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 7xx: Machine Check exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: 7xx: Machine Check exception cleanup

There's no MSR_HV in the 7xx.

Also remove 40x and BookE code.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-5-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Simplify powerpc_excp_7xx
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: Simplify powerpc_excp_7xx

Differences from the generic powerpc_excp code:

- Not BookE, so some MSR bits are cleared at interrupt dispatch;
- No MSR_HV;
- No power saving states;
- No Hypervisor Emulation Assistance;
- Not 64 bits;
- No System call vectored;
- No Alternate Interrupt Location.

Exceptions used:

POWERPC_EXCP_ALIGN
POWERPC_EXCP_DECR
POWERPC_EXCP_DLTLB
POWERPC_EXCP_DSI
POWERPC_EXCP_DSTLB
POWERPC_EXCP_EXTERNAL
POWERPC_EXCP_FPU
POWERPC_EXCP_IABR
POWERPC_EXCP_IFTLB
POWERPC_EXCP_ISI
POWERPC_EXCP_MCHECK
POWERPC_EXCP_PERFM
POWERPC_EXCP_PROGRAM
POWERPC_EXCP_RESET
POWERPC_EXCP_SMI
POWERPC_EXCP_SYSCALL
POWERPC_EXCP_THERM
POWERPC_EXCP_TRACE

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-4-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Introduce powerpc_excp_7xx
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: Introduce powerpc_excp_7xx

Introduce a new powerpc_excp function specific for PowerPC 7xx CPUs
(740, 745, 750, 750cl, 750cx, 750fx, 750gx, 755). This commit copies
powerpc_excp_legacy verbatim so the next one has a clean diff.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-3-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Merge 7x5 and 7x0 exception model IDs
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: Merge 7x5 and 7x0 exception model IDs

Since we've split the exception code by exception model, the exception
model IDs are becoming less useful. These two can be merged.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220204173430.1457358-2-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 6xx: Set SRRs directly in exception code
Fabiano Rosas [Wed, 9 Feb 2022 08:08:56 +0000 (09:08 +0100)]
target/ppc: 6xx: Set SRRs directly in exception code

The 6xx CPUs don't have alternate/hypervisor Save and Restore
Registers, so we can set SRR0 and SRR1 directly.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-12-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 6xx: Software TLB exceptions cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: 6xx: Software TLB exceptions cleanup

This code applies only to the 6xx CPUs, so we can remove the switch
statement.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-11-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 6xx: System Reset interrupt cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: 6xx: System Reset interrupt cleanup

There is no HV support in the 6xx.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-10-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 6xx: System Call exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: 6xx: System Call exception cleanup

There is no Hypervisor mode in the 6xx CPUs.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-9-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 6xx: Program exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: 6xx: Program exception cleanup

There's no ESR in the 6xx CPUs.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-8-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 6xx: External interrupt cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: 6xx: External interrupt cleanup

There's no Hypervisor mode in the 6xx, so remove all LPES0 logic.

Also remove BookE IRQ code.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-7-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 6xx: Machine Check exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: 6xx: Machine Check exception cleanup

There's no MSR_HV in the 6xx CPUs.

Also remove the 40x and BookE code.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-6-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: 6xx: Critical exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: 6xx: Critical exception cleanup

This only applies to the G2s, the other 6xx CPUs will not have this
vector registered.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-5-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Simplify powerpc_excp_6xx
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: Simplify powerpc_excp_6xx

Differences from the generic powerpc_excp code:

- Not BookE, so some MSR bits are cleared at interrupt dispatch;
- No MSR_HV;
- No power saving states;
- No Hypervisor Emulation Assistance;
- Not 64 bits;
- No System call vectored;
- No Alternate Interrupt Location.

Exceptions used:

POWERPC_EXCP_ALIGN
POWERPC_EXCP_CRITICAL
POWERPC_EXCP_DABR
POWERPC_EXCP_DECR
POWERPC_EXCP_DLTLB
POWERPC_EXCP_DSI
POWERPC_EXCP_DSTLB
POWERPC_EXCP_DTLB
POWERPC_EXCP_EXTERNAL
POWERPC_EXCP_FPA
POWERPC_EXCP_FPU
POWERPC_EXCP_IABR
POWERPC_EXCP_IFTLB
POWERPC_EXCP_ISI
POWERPC_EXCP_ITLB
POWERPC_EXCP_MCHECK
POWERPC_EXCP_MEXTBR
POWERPC_EXCP_NMEXTBR
POWERPC_EXCP_PROGRAM
POWERPC_EXCP_RESET
POWERPC_EXCP_SMI
POWERPC_EXCP_SYSCALL
POWERPC_EXCP_TRACE

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-4-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Introduce powerpc_excp_6xx
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: Introduce powerpc_excp_6xx

Introduce a new powerpc_excp function specific for PowerPC 6xx CPUs
(603, 604, G2, MPC5xx, MCP8xx). This commit copies powerpc_excp_legacy
verbatim so the next one has a clean diff.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-3-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Merge exception model IDs for 6xx CPUs
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: Merge exception model IDs for 6xx CPUs

We don't need three separate exception model IDs for the 603, 604 and
G2.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203200957.1434641-2-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Remove PowerPC 601 CPUs
Cédric Le Goater [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: Remove PowerPC 601 CPUs

The PowerPC 601 processor is the first generation of processors to
implement the PowerPC architecture. It was designed as a bridge
processor and also could execute most of the instructions of the
previous POWER architecture. It was found on the first Macs and IBM
RS/6000 workstations.

There is not much interest in keeping the CPU model of this
POWER-PowerPC bridge processor. We have the 603 and 604 CPU models of
the 60x family which implement the complete PowerPC instruction set.

Cc: "Hervé Poussineau" <hpoussin@reactos.org>
Cc: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220203142756.1302515-1-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Fix radix logging
Cédric Le Goater [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: Fix radix logging

ppc_radix64_partition_scoped_xlate() logs the host page protection
bits variable but it is uninitialized. The value is set later on in
ppc_radix64_check_prot(). Remove the output.

Fixes: Coverity CID 1468942
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Message-Id: <20220203142145.1301749-1-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: System Reset exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: System Reset exception cleanup

There is no MSR_HV in BookE, so remove all of the HV logic.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-12-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: Watchdog Timer interrupt
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: Watchdog Timer interrupt

Remove the switch as this function applies to BookE only.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-11-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: System Call exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: System Call exception cleanup

QEMU does not support BookE as a hypervisor.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-10-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: Alignment interrupt cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: Alignment interrupt cleanup

BookE has no DSISR or DAR. The proper registers ESR and DEAR were
already set at this point.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-9-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: External interrupt cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: External interrupt cleanup

There is no LPES0 in BookE and no MSR_HV.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-8-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: Instruction storage exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: Instruction storage exception cleanup

The SRR1 should be set to the MSR value. There are no diagnostic bits
in the SRR1 for BookE.

Note that this fixes a bug where MSR_GS would be set and Linux would
go into KVM code when there's no KVM guest.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-7-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: Data Storage exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: Data Storage exception cleanup

There is no DSISR or DAR in BookE. Change to ESR and DEAR.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-6-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: Machine Check cleanups
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: Machine Check cleanups

There's no MSR_HV in BookE.

Also remove 40x code.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-5-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: booke: Critical exception cleanup
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: booke: Critical exception cleanup

Remove 40x and G2 code.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-4-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Simplify powerpc_excp_booke
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: Simplify powerpc_excp_booke

Differences from the generic powerpc_excp code:

- No MSR bits are cleared at interrupt dispatch;
- No MSR_HV;
- No power saving states;
- No Hypervisor Emulation Assistance;
- SPEU needs special handling;
- Big endian only;
- Both 64 and 32 bits;
- No System call vectored;
- No Alternate Interrupt Location.

Exceptions used:

POWERPC_EXCP_ALIGN
POWERPC_EXCP_APU
POWERPC_EXCP_CRITICAL
POWERPC_EXCP_DEBUG
POWERPC_EXCP_DECR
POWERPC_EXCP_DSI
POWERPC_EXCP_DTLB
POWERPC_EXCP_EFPDI
POWERPC_EXCP_EFPRI
POWERPC_EXCP_EXTERNAL
POWERPC_EXCP_FIT
POWERPC_EXCP_FPU
POWERPC_EXCP_ISI
POWERPC_EXCP_ITLB
POWERPC_EXCP_MCHECK
POWERPC_EXCP_PROGRAM
POWERPC_EXCP_RESET
POWERPC_EXCP_SPEU
POWERPC_EXCP_SYSCALL
POWERPC_EXCP_WDT

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-3-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Introduce powerpc_excp_booke
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: Introduce powerpc_excp_booke

Introduce a new powerpc_excp function specific for BookE CPUs. This
commit copies powerpc_excp_legacy verbatim so the next one has a clean
diff.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220128224018.1228062-2-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotarget/ppc: Remove 440x4 CPU
Fabiano Rosas [Wed, 9 Feb 2022 08:08:55 +0000 (09:08 +0100)]
target/ppc: Remove 440x4 CPU

This CPU was partially removed due to lack of support in 2017 by commit
aef7796057 ("ppc: remove non implemented cpu models").

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220128221611.1221715-1-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
2 years agotests/tcg/multiarch: Add sigbus.c
Richard Henderson [Wed, 28 Jul 2021 21:32:24 +0000 (11:32 -1000)]
tests/tcg/multiarch: Add sigbus.c

A mostly generic test for unaligned access raising SIGBUS.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/sparc: Support unaligned access for user-only
Richard Henderson [Thu, 5 Aug 2021 00:54:30 +0000 (03:54 +0300)]
tcg/sparc: Support unaligned access for user-only

This is kinda sorta the opposite of the other tcg hosts, where
we get (normal) alignment checks for free with host SIGBUS and
need to add code to support unaligned accesses.

This inline code expansion is somewhat large, but it takes quite
a few instructions to make a function call to a helper anyway.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/sparc: Add tcg_out_jmpl_const for better tail calls
Richard Henderson [Wed, 4 Aug 2021 23:23:05 +0000 (02:23 +0300)]
tcg/sparc: Add tcg_out_jmpl_const for better tail calls

Due to mapping changes, we now rarely place the code_gen_buffer
near the main executable.  Which means that direct calls will
now rarely be in range.

So, always use indirect calls for tail calls, which allows us to
avoid clobbering %o7, and therefore we need not save and restore it.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/sparc: Use the constant pool for 64-bit constants
Richard Henderson [Thu, 5 Aug 2021 01:34:20 +0000 (04:34 +0300)]
tcg/sparc: Use the constant pool for 64-bit constants

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/sparc: Convert patch_reloc to return bool
Richard Henderson [Sat, 5 Feb 2022 05:12:30 +0000 (08:12 +0300)]
tcg/sparc: Convert patch_reloc to return bool

Since 7ecd02a06f8, if patch_reloc fails we restart translation
with a smaller TB.  SPARC had its function signature changed,
but not the logic.  Replace assert with return false.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/sparc: Improve code gen for shifted 32-bit constants
Richard Henderson [Thu, 5 Aug 2021 01:09:15 +0000 (04:09 +0300)]
tcg/sparc: Improve code gen for shifted 32-bit constants

We had code for checking for 13 and 21-bit shifted constants,
but we can do better and allow 32-bit shifted constants.
This is still 2 insns shorter than the full 64-bit sequence.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/sparc: Add scratch argument to tcg_out_movi_int
Richard Henderson [Wed, 4 Aug 2021 23:16:05 +0000 (02:16 +0300)]
tcg/sparc: Add scratch argument to tcg_out_movi_int

This will allow us to control exactly what scratch register is
used for loading the constant.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/sparc: Split out tcg_out_movi_imm32
Richard Henderson [Mon, 7 Feb 2022 23:46:40 +0000 (10:46 +1100)]
tcg/sparc: Split out tcg_out_movi_imm32

Handle 32-bit constants with a separate function, so that
tcg_out_movi_int does not need to recurse.  This slightly
rearranges the order of tests for small constants, but
produces the same output.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/sparc: Use tcg_out_movi_imm13 in tcg_out_addsub2_i64
Richard Henderson [Mon, 7 Feb 2022 23:15:58 +0000 (10:15 +1100)]
tcg/sparc: Use tcg_out_movi_imm13 in tcg_out_addsub2_i64

When BH is constant, it is constrained to 11 bits for use in MOVCC.
For the cases in which we must load the constant BH into a register,
we do not need the full logic of tcg_out_movi; we can use the simpler
function for emitting a 13 bit constant.

This eliminates the only case in which TCG_REG_T2 was passed to
tcg_out_movi, which will shortly become invalid.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/mips: Support unaligned access for softmmu
Richard Henderson [Fri, 6 Aug 2021 16:55:14 +0000 (06:55 -1000)]
tcg/mips: Support unaligned access for softmmu

We can use the routines just added for user-only to emit
unaligned accesses in softmmu mode too.

Tested-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/mips: Support unaligned access for user-only
Richard Henderson [Fri, 6 Aug 2021 15:49:16 +0000 (05:49 -1000)]
tcg/mips: Support unaligned access for user-only

This is kinda sorta the opposite of the other tcg hosts, where
we get (normal) alignment checks for free with host SIGBUS and
need to add code to support unaligned accesses.

Fortunately, the ISA contains pairs of instructions that are
used to implement unaligned memory accesses.  Use them.

Tested-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/arm: Support raising sigbus for user-only
Richard Henderson [Tue, 10 Aug 2021 05:18:27 +0000 (19:18 -1000)]
tcg/arm: Support raising sigbus for user-only

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/arm: Reserve a register for guest_base
Richard Henderson [Tue, 10 Aug 2021 03:44:50 +0000 (17:44 -1000)]
tcg/arm: Reserve a register for guest_base

Reserve a register for the guest_base using aarch64 for reference.
By doing so, we do not have to recompute it for every memory load.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/arm: Support unaligned access for softmmu
Richard Henderson [Sun, 8 Aug 2021 04:19:14 +0000 (18:19 -1000)]
tcg/arm: Support unaligned access for softmmu

From armv6, the architecture supports unaligned accesses.
All we need to do is perform the correct alignment check
in tcg_out_tlb_read.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/arm: Check alignment for ldrd and strd
Richard Henderson [Mon, 3 Jan 2022 05:26:17 +0000 (21:26 -0800)]
tcg/arm: Check alignment for ldrd and strd

We will shortly allow the use of unaligned memory accesses,
and these require proper alignment.  Use get_alignment_bits
to verify and remove USING_SOFTMMU.

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