Messages in this thread | | | From | "Krcka, Tomas" <> | Subject | Re: [PATCH] iommu/arm-smmu-v3: Fix event queue overflow acknowledgment | Date | Wed, 29 Mar 2023 12:36:26 +0000 |
| |
> On 28. Mar 2023, at 13:56, Will Deacon <will@kernel.org> wrote: > > > On Tue, Mar 28, 2023 at 07:13:52AM +0000, Krcka, Tomas wrote: >> >>> On 27. Mar 2023, at 14:12, Will Deacon <will@kernel.org> wrote: >>> >>> >>> On Wed, Mar 08, 2023 at 09:20:47AM +0000, Tomas Krcka wrote: >>>> When an overflow occurs in the event queue, the SMMU toggles overflow >>>> flag OVFLG in the PROD register. >>>> The evtq thread is supposed to acknowledge the overflow flag by toggling >>>> flag OVACKFLG in the CONS register, otherwise the overflow condition is >>>> still active (OVFLG != OVACKFLG). >>>> >>>> Currently the acknowledge register is toggled after clearing the event >>>> queue but is never propagated to the hardware. It would be done next >>>> time when executing evtq thread. >>>> >>>> The SMMU still adds elements to the queue when the overflow condition is >>>> active but any subsequent overflow information after clearing the event >>>> queue will be lost. >>>> >>>> This change keeps the SMMU in sync as it's expected by design. >>>> >>>> Signed-off-by: Tomas Krcka <krckatom@amazon.de> >>>> Suggested-by: KarimAllah Ahmed <karahmed@amazon.de> >>>> --- >>>> drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 1 + >>>> 1 file changed, 1 insertion(+) >>>> >>>> diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c >>>> index f2425b0f0cd6..acc1ff5ff69b 100644 >>>> --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c >>>> +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c >>>> @@ -1579,6 +1579,7 @@ static irqreturn_t arm_smmu_evtq_thread(int irq, void *dev) >>>> /* Sync our overflow flag, as we believe we're up to speed */ >>>> llq->cons = Q_OVF(llq->prod) | Q_WRP(llq, llq->cons) | >>>> Q_IDX(llq, llq->cons); >>>> + queue_sync_cons_out(q); >>>> return IRQ_HANDLED; >>>> } >>> >>> I think I probably did mean to have something like this, but can we >>> only do the actual h/w update if overflow has occurred? Otherwise I think >>> we're pointlessly writing back the same value most of the time. >>> >>> Will >> >> Yes, we can, but then same applies for the priq as well, there we also write back >> every time. > > Sure, feel free to update both. >
OK, I sent it as new patch: https://lore.kernel.org/lkml/20230329123420.34641-1-tomas.krcka@gmail.com
Amazon Development Center Germany GmbH Krausenstr. 38 10117 Berlin Geschaeftsfuehrung: Christian Schlaeger, Jonathan Weiss Eingetragen am Amtsgericht Charlottenburg unter HRB 149173 B Sitz: Berlin Ust-ID: DE 289 237 879
| |