Originally, all failed SG_IO requests called scsi_handle_rw_error() to
apply the configured error policy. However, commit f3126d65, which was
supposed to be a mere refactoring for scsi-disk.c, broke this and
accidentally completed the SCSI request without considering the error
policy any more if the error was signalled in the host_status field.
Apart from the commit message not describing the change as intended,
errors indicated in host_status are also obviously backend errors and
not something the guest must deal with independently of the error
policy.
This behaviour means that some recoverable errors (such as a path error
in multipath configurations) were reported to the guest anyway, which
might not expect it and might consider its disk broken.
Make sure that we apply the error policy again for host_status errors,
too. This addresses an existing FIXME comment and allows us to remove
some comments warning that callbacks weren't always called. With this
fix, they are called in all cases again.
The return value passed to the request callback doesn't have more free
values that could be used to indicate host_status errors as well as SAM
status codes and negative errno. Store the value in the host_status
field of the SCSIRequest instead and use -ENODEV as the return value (if
a path hasn't been reachable for a while, blk_aio_ioctl() will return
-ENODEV instead of just setting host_status, so just reuse it here -
it's not necessarily entirely accurate, but it's as good as any errno).
Cc: qemu-stable@nongnu.org
Fixes: f3126d65b3 ('scsi: move host_status handling into SCSI drivers')
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20250407155949.44736-1-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
qcow2_refresh_limits() assumes that s->crypto is non-NULL whenever
bs->encrypted is true. This is actually not the case: qcow2_do_open()
allows to open an image with a missing crypto header for BDRV_O_NO_IO,
and then bs->encrypted is true, but s->crypto is still NULL.
It doesn't make sense to open an invalid image, so remove the exception
for BDRV_O_NO_IO. This catches the problem early and any code that makes
the same assumption is safe now.
At the same time, in the name of defensive programming, we shouldn't
make the assumption in the first place. Let qcow2_refresh_limits() check
s->crypto rather than bs->encrypted. If s->crypto is NULL, it also can't
make any requirement on request alignment.
Finally, start a qcow2-encryption test case that only serves as a
regression test for this crash for now.
Reported-by: Leonid Reviakin <L.reviakin@fobos-nt.ru>
Reported-by: Denis Rastyogin <gerben@altlinux.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20250318201143.70657-1-kwolf@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This error was discovered by fuzzing qemu-img.
This commit fixes a division by zero error in the bench_cb() function
that occurs when using the bench command with a zero-sized image.
The issue arises because b->image_size can be zero, leading to a
division by zero in the modulo operation (b->offset %= b->image_size).
This patch adds a check for b->image_size == 0 and resets b->offset
to 0 in such cases, preventing the error.
Signed-off-by: Denis Rastyogin <gerben@altlinux.org>
Message-ID: <20250318101933.255617-1-gerben@altlinux.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Commit 5b4beba124 ("RISC-V Spike Machines") added the Spike machine and
made it default for qemu-system-riscv32/64. It was the first RISC-V
machine added in QEMU so setting it as default was sensible.
Today we have 7 riscv64 and 6 riscv32 machines and having 'spike' as
default machine is not intuitive. For example, [1] is a bug that was
opened with the 'virt' board in mind, but given that the user didn't
pass a '-machine' option, the user was using 'spike' without knowing.
Being explicit in the command line is desirable when we have a handful
of boards available, so deprecate the default machine setting from
RISC-V.
[1] https://gitlab.com/qemu-project/qemu/-/issues/2467
Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20250404122858.241598-1-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
First, the VMapple machine only works with the ARM 'host' CPU
type, which isn't accepted for QTest:
$ qemu-system-aarch64 -M vmapple -accel qtest
qemu-system-aarch64: The 'host' CPU type can only be used with KVM or HVF
Second, the QTest framework expects machines to be createable
without specifying optional arguments, however the VMapple
machine requires few of them:
$ qemu-system-aarch64 -M vmapple -accel qtest
qemu-system-aarch64: No firmware specified
$ qemu-system-aarch64 -M vmapple -accel qtest -bios /dev/null
qemu-system-aarch64: No AUX device. Please specify one as pflash drive.
Restrict this machine with QTest so we can at least run check-qtest,
otherwise we get:
$ make check-qtest-aarch64
qemu-system-aarch64: The 'host' CPU type can only be used with KVM or HVF
Broken pipe
../tests/qtest/libqtest.c:199: kill_qemu() tried to terminate QEMU process but encountered exit status 1 (expected 0)
...
7/26 qemu:qtest+qtest-aarch64 / qtest-aarch64/test-hmp ERROR 24.71s killed by signal 6 SIGABRT
2/26 qemu:qtest+qtest-aarch64 / qtest-aarch64/qom-test ERROR 71.23s killed by signal 6 SIGABRT
Suggested-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20250403203241.46692-5-philmd@linaro.org>
This reverts commit e2668ba1ed.
This commit made test 162 fail occasionally with:
162 fail [13:06:40] [13:06:40] 0.2s (last: 0.2s) output mismatch
--- tests/qemu-iotests/162.out
+++ tests/qemu-iotests/scratch/qcow2-file-162/162.out.bad
@@ -3,6 +3,7 @@
=== NBD ===
qemu-img: Could not open 'json:{"driver": "nbd", "host": -1}': address
resolution failed for -1:10809: Name or service not known
image: nbd://localhost:PORT
+./common.rc: line 371: kill: (891116) - No such process
image: nbd+unix://?socket=42
The nbd server should normally terminate automatically, so trying to
kill it here now seems to cause a race that will cause a test failure
when the server terminated before the kill command has been executed.
The "Stop NBD server" patch has originally been written to solve another
problem with a hanging nbd server, but since that problem has been properly
solved by commit 3e16834856, we now don't need the "_stop_nbd_server" here
anymore.
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-ID: <20250326143533.932899-1-thuth@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Currently, the QEMU-emulated AMD IOMMU device use PCI vendor id 0x1022
(AMD) with device id zero (undefined). Eventhough this does not cause any
functional issue for AMD IOMMU driver since it normally uses information
in the ACPI IVRS table to probe and initialize the device per
recommendation in the AMD IOMMU specification, the device id zero causes
the Windows Device Manager utility to show the device as an unknown device.
Since Windows only recognizes AMD IOMMU device with device id 0x1419 as
listed in the machine.inf file, modify the QEMU AMD IOMMU model to use
the id 0x1419 to avoid the issue. This advertise the IOMMU as the AMD
IOMMU device for Family 15h (Models 10h-1fh).
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Message-Id: <20250325021140.5676-1-suravee.suthikulpanit@amd.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Yan Vugenfirer <yvugenfi@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When update_pagemask was split from helper_mtc0_pagemask,
we failed to actually write to the new parameter but continue
to write to env->CP0_PageMask. Thus the use within
page_table_walk_refill modifies cpu state and not the local
variable as expected.
Simplify by renaming to compute_pagemask and returning the
value directly. No need for either env or pointer return.
Fixes: 074cfcb4da ("target/mips: Implement hardware page table walker for MIPS32")
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20250328175526.368121-4-richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Cc: qemu-stable@nongnu.org
