On Fri, Sep 25, 2020 at 11:45:05AM +0000, Suravee Suthikulpanit wrote:
> When using 128-bit interrupt-remapping table entry (IRTE) (a.k.a GA mode),
> current driver disables interrupt remapping when it updates the IRTE
> so that the upper and lower 64-bit values can be updated safely.
>
> However, this creates a small window, where the interrupt could
> arrive and result in IO_PAGE_FAULT (for interrupt) as shown below.
>
> IOMMU Driver Device IRQ
> ============ ===========
> irte.RemapEn=0
> ...
> change IRTE IRQ from device ==> IO_PAGE_FAULT !!
> ...
> irte.RemapEn=1
>
> This scenario has been observed when changing irq affinity on a system
> running I/O-intensive workload, in which the destination APIC ID
> in the IRTE is updated.
>
> Instead, use cmpxchg_double() to update the 128-bit IRTE at once without
> disabling the interrupt remapping. However, this means several features,
> which require GA (128-bit IRTE) support will also be affected if cmpxchg16b
> is not supported (which is unprecedented for AMD processors w/ IOMMU).
>
> Cc: stable@xxxxxxxxxxxxxxx
> Fixes: 880ac60e2538 ("iommu/amd: Introduce interrupt remapping ops structure")
> Reported-by: Sean Osborne <sean.m.osborne@xxxxxxxxxx>
> Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@xxxxxxx>
> Tested-by: Erik Rockstrom <erik.rockstrom@xxxxxxxxxx>
> Reviewed-by: Joao Martins <joao.m.martins@xxxxxxxxxx>
> Link: https://lore.kernel.org/r/20200903093822.52012-3-suravee.suthikulpanit@xxxxxxx
> Signed-off-by: Joerg Roedel <jroedel@xxxxxxx>
> ---
> Note: This patch is the back-port on top of the stable branch linux-5.4.y
> for the upstream commit e52d58d54a32 ("iommu/amd: Use cmpxchg_double() when
> updating 128-bit IRTE") since the original patch does not apply cleanly.
Now queued up, thanks.
greg k-h
