kernel/drivers/vfio/platform, branch linux-6.9.y

vfio/platform: Create persistent IRQ handlers

2024-03-11T19:08:52Z

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

vfio/platform: Disable virqfds on cleanup

2024-03-11T19:08:52Z

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

vfio/platform: Convert to platform remove callback returning void

2024-03-11T18:09:22Z

The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by: Uwe Kleine-König Reviewed-by: Jason Gunthorpe Reviewed-by: Eric Auger Link: https://lore.kernel.org/r/79d3df42fe5b359a05b8061631e72e5ed249b234.1709886922.git.u.kleine-koenig@pengutronix.de Signed-off-by: Alex Williamson

vfio: amba: Rename pl330_ids[] to vfio_amba_ids[]

2024-03-04T22:10:49Z

Obviously drivers/vfio/platform/vfio_amba.c started its life as a simplified copy of drivers/dma/pl330.c, but not all variable names were updated. Signed-off-by: Geert Uytterhoeven Reviewed-by: Jason Gunthorpe Link: https://lore.kernel.org/r/1d1b873b59b208547439225aee1f24d6f2512a1f.1708945194.git.geert+renesas@glider.be Signed-off-by: Alex Williamson

eventfd: simplify eventfd_signal()

2023-11-28T13:08:38Z

Ever since the eventfd type was introduced back in 2007 in commit e1ad7468c77d ("signal/timer/event: eventfd core") the eventfd_signal() function only ever passed 1 as a value for @n. There's no point in keeping that additional argument. Link: https://lore.kernel.org/r/20231122-vfs-eventfd-signal-v2-2-bd549b14ce0c@kernel.org Acked-by: Xu Yilun Acked-by: Andrew Donnellan # ocxl Acked-by: Eric Farman # s390 Reviewed-by: Jan Kara Reviewed-by: Jens Axboe Signed-off-by: Christian Brauner

vfio-iommufd: Add detach_ioas support for physical VFIO devices

2023-07-25T16:19:12Z

This prepares for adding DETACH ioctl for physical VFIO devices. Reviewed-by: Kevin Tian Reviewed-by: Jason Gunthorpe Tested-by: Terrence Xu Tested-by: Nicolin Chen Tested-by: Matthew Rosato Tested-by: Yanting Jiang Tested-by: Shameer Kolothum Tested-by: Zhenzhong Duan Signed-off-by: Yi Liu Link: https://lore.kernel.org/r/20230718135551.6592-14-yi.l.liu@intel.com Signed-off-by: Alex Williamson

vfio/platform: Cleanup Kconfig

2023-06-16T18:23:50Z

Like vfio-pci, there's also a base module here where vfio-amba depends on vfio-platform, when really it only needs vfio-platform-base. Create a sub-menu for platform drivers and a nested menu for reset drivers. Cleanup Makefile to make use of new CONFIG_VFIO_PLATFORM_BASE for building the shared modules and traversing reset modules. Reviewed-by: Cédric Le Goater Reviewed-by: Eric Auger Link: https://lore.kernel.org/r/20230614193948.477036-3-alex.williamson@redhat.com Signed-off-by: Alex Williamson

vfio: platform: ignore missing reset if disabled at module init

2023-02-01T19:25:41Z

If reset requirement was relaxed via module parameter, errors caused by missing reset should not be propagated down to the vfio core. Otherwise initialization will fail. Signed-off-by: Tomasz Duszynski Fixes: 5f6c7e0831a1 ("vfio/platform: Use the new device life cycle helpers") Reviewed-by: Kevin Tian Reviewed-by: Eric Auger Link: https://lore.kernel.org/r/20230131083349.2027189-1-tduszynski@marvell.com Signed-off-by: Alex Williamson

vfio/platform: Use GFP_KERNEL_ACCOUNT for userspace persistent allocations

2023-01-23T18:26:30Z

Use GFP_KERNEL_ACCOUNT for userspace persistent allocations. The GFP_KERNEL_ACCOUNT option lets the memory allocator know that this is untrusted allocation triggered from userspace and should be a subject of kmem accounting, and as such it is controlled by the cgroup mechanism. Signed-off-by: Yishai Hadas Reviewed-by: Jason Gunthorpe Link: https://lore.kernel.org/r/20230108154427.32609-7-yishaih@nvidia.com Signed-off-by: Alex Williamson

vfio: platform: No need to check res again

2023-01-23T18:26:28Z

In function vfio_platform_regions_init(),we did check res implied by using while loop, so no need to check whether res be null or not again. No functional change intended. Signed-off-by: Angus Chen Reviewed-by: Eric Auger Link: https://lore.kernel.org/r/20230107034721.2127-1-angus.chen@jaguarmicro.com Signed-off-by: Alex Williamson