linux-stable.git/drivers/vfio/platform, branch linux-5.10.y

vfio/platform: check the bounds of read/write syscalls

2025-03-13T11:47:12+00:00

commit ce9ff21ea89d191e477a02ad7eabf4f996b80a69 upstream.

count and offset are passed from user space and not checked, only
offset is capped to 40 bits, which can be used to read/write out of
bounds of the device.

Fixes: 6e3f26456009 (“vfio/platform: read and write support for the device fd”)
Cc: stable@vger.kernel.org
Reported-by: Mostafa Saleh 
Reviewed-by: Eric Auger 
Reviewed-by: Mostafa Saleh 
Tested-by: Mostafa Saleh 
Signed-off-by: Alex Williamson 
Signed-off-by: Greg Kroah-Hartman

vfio/platform: check the bounds of read/write syscalls

2025-02-01T17:22:33+00:00

commit ce9ff21ea89d191e477a02ad7eabf4f996b80a69 upstream.

count and offset are passed from user space and not checked, only
offset is capped to 40 bits, which can be used to read/write out of
bounds of the device.

Fixes: 6e3f26456009 (“vfio/platform: read and write support for the device fd”)
Cc: stable@vger.kernel.org
Reported-by: Mostafa Saleh 
Reviewed-by: Eric Auger 
Reviewed-by: Mostafa Saleh 
Tested-by: Mostafa Saleh 
Signed-off-by: Alex Williamson 
Signed-off-by: Greg Kroah-Hartman

vfio/platform: Create persistent IRQ handlers

2024-04-13T10:59:21+00:00

[ Upstream commit 675daf435e9f8e5a5eab140a9864dfad6668b375 ]

The vfio-platform SET_IRQS ioctl currently allows loopback triggering of
an interrupt before a signaling eventfd has been configured by the user,
which thereby allows a NULL pointer dereference.

Rather than register the IRQ relative to a valid trigger, register all
IRQs in a disabled state in the device open path.  This allows mask
operations on the IRQ to nest within the overall enable state governed
by a valid eventfd signal.  This decouples @masked, protected by the
@locked spinlock from @trigger, protected via the @igate mutex.

In doing so, it's guaranteed that changes to @trigger cannot race the
IRQ handlers because the IRQ handler is synchronously disabled before
modifying the trigger, and loopback triggering of the IRQ via ioctl is
safe due to serialization with trigger changes via igate.

For compatibility, request_irq() failures are maintained to be local to
the SET_IRQS ioctl rather than a fatal error in the open device path.
This allows, for example, a userspace driver with polling mode support
to continue to work regardless of moving the request_irq() call site.
This necessarily blocks all SET_IRQS access to the failed index.

Cc: Eric Auger 
Cc:  
Fixes: 57f972e2b341 ("vfio/platform: trigger an interrupt via eventfd")
Reviewed-by: Kevin Tian 
Reviewed-by: Eric Auger 
Link: https://lore.kernel.org/r/20240308230557.805580-7-alex.williamson@redhat.com
Signed-off-by: Alex Williamson 
Signed-off-by: Greg Kroah-Hartman

vfio/platform: Disable virqfds on cleanup

2024-04-13T10:58:32+00:00

[ Upstream commit fcdc0d3d40bc26c105acf8467f7d9018970944ae ]

irqfds for mask and unmask that are not specifically disabled by the
user are leaked.  Remove any irqfds during cleanup

Cc: Eric Auger 
Cc:  
Fixes: a7fa7c77cf15 ("vfio/platform: implement IRQ masking/unmasking via an eventfd")
Reviewed-by: Kevin Tian 
Reviewed-by: Eric Auger 
Link: https://lore.kernel.org/r/20240308230557.805580-6-alex.williamson@redhat.com
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

vfio: platform: Do not pass return buffer to ACPI _RST method

2023-01-14T09:15:54+00:00

[ Upstream commit e67e070632a665c932d534b8b800477bb3111449 ]

The ACPI _RST method has no return value, there's no need to pass a return
buffer to acpi_evaluate_object().

Fixes: d30daa33ec1d ("vfio: platform: call _RST method when using ACPI")
Signed-off-by: Rafael Mendonca 
Reviewed-by: Eric Auger 
Link: https://lore.kernel.org/r/20221018152825.891032-1-rafaelmendsr@gmail.com
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

amba: Make the remove callback return void

2022-04-08T12:40:02+00:00

[ Upstream commit 3fd269e74f2feec973f45ee11d822faeda4fe284 ]

All amba drivers return 0 in their remove callback. Together with the
driver core ignoring the return value anyhow, it doesn't make sense to
return a value here.

Change the remove prototype to return void, which makes it explicit that
returning an error value doesn't work as expected. This simplifies changing
the core remove callback to return void, too.

Reviewed-by: Ulf Hansson 
Reviewed-by: Arnd Bergmann 
Acked-by: Alexandre Belloni 
Acked-by: Dmitry Torokhov 
Acked-by: Krzysztof Kozlowski  # for drivers/memory
Acked-by: Mark Brown 
Acked-by: Linus Walleij 
Acked-by: Suzuki K Poulose  # for hwtracing/coresight
Acked-By: Vinod Koul  # for dmaengine
Acked-by: Guenter Roeck  # for watchdog
Acked-by: Wolfram Sang  # for I2C
Acked-by: Takashi Iwai  # for sound
Acked-by: Vladimir Zapolskiy  # for memory/pl172
Acked-by: Greg Kroah-Hartman 
Link: https://lore.kernel.org/r/20210126165835.687514-5-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König 
Signed-off-by: Sasha Levin

vfio: platform: simplify device removal

2022-04-08T12:40:02+00:00

[ Upstream commit 5b495ac8fe03b9e0d2e775f9064c3e2a340ff440 ]

vfio_platform_remove_common() cannot return non-NULL in
vfio_amba_remove() as the latter is only called if vfio_amba_probe()
returned success.

Diagnosed-by: Arnd Bergmann 
Acked-by: Eric Auger 
Link: https://lore.kernel.org/r/20210126165835.687514-4-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König 
Signed-off-by: Sasha Levin

vfio/platform: fix module_put call in error flow

2021-06-10T11:39:15+00:00

[ Upstream commit dc51ff91cf2d1e9a2d941da483602f71d4a51472 ]

The ->parent_module is the one that use in try_module_get. It should
also be the one the we use in module_put during vfio_platform_open().

Fixes: 32a2d71c4e80 ("vfio: platform: introduce vfio-platform-base module")
Signed-off-by: Max Gurtovoy 
Message-Id: <20210518192133.59195-1-mgurtovoy@nvidia.com>
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

vfio: platform: fix reference leak in vfio_platform_open

2020-11-03T18:07:40+00:00

pm_runtime_get_sync() will increment pm usage counter even it
failed. Forgetting to call pm_runtime_put will result in
reference leak in vfio_platform_open, so we should fix it.

Signed-off-by: Zhang Qilong 
Acked-by: Eric Auger 
Signed-off-by: Alex Williamson

vfio: platform: Switch to platform_get_irq_optional()

2020-03-24T15:26:30+00:00

Since commit 7723f4c5ecdb ("driver core: platform: Add an error
message to platform_get_irq*()"), platform_get_irq() calls dev_err()
on an error. As we enumerate all interrupts until platform_get_irq()
fails, we now systematically get a message such as:
"vfio-platform fff51000.ethernet: IRQ index 3 not found" which is
a false positive.

Let's use platform_get_irq_optional() instead.

Signed-off-by: Eric Auger 
Cc: stable@vger.kernel.org # v5.3+
Reviewed-by: Andre Przywara 
Tested-by: Andre Przywara 
Signed-off-by: Alex Williamson