On Tue, 2 Dec 2025 15:09:42 +0000
Akash Goel <[email protected]> wrote:
> Hi Boris,
>
>
> On 12/2/25 13:35, Boris Brezillon wrote:
> > An AS can be disabled in the middle of a VM operation (VM being
> > evicted from an AS slot, for instance). In that case, we need the
> > locked section to be unlocked before releasing the slot.
> >
> > Fixes: 6e2d3b3e8589 ("drm/panthor: Add support for atomic page table
> > updates")
> > Signed-off-by: Boris Brezillon <[email protected]>
> > ---
> > drivers/gpu/drm/panthor/panthor_mmu.c | 11 +++++++++++
> > 1 file changed, 11 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/panthor/panthor_mmu.c
> > b/drivers/gpu/drm/panthor/panthor_mmu.c
> > index 3644af1a8e56..4ed28bd1b8ec 100644
> > --- a/drivers/gpu/drm/panthor/panthor_mmu.c
> > +++ b/drivers/gpu/drm/panthor/panthor_mmu.c
> > @@ -588,6 +588,7 @@ static int panthor_mmu_as_enable(struct panthor_device
> > *ptdev, u32 as_nr,
> > static int panthor_mmu_as_disable(struct panthor_device *ptdev, u32 as_nr,
> > bool recycle_slot)
> > {
> > + struct panthor_vm *vm = ptdev->mmu->as.slots[as_nr].vm;
> > int ret;
> >
>
> Thanks for the fix.
>
> Please can you consider adding a lockdep_assert_held for
> mmu->as.slots_lock. That would in a way make it explicit that it is safe
> to de-reference vm pointer here.
Sure, I can do that.
>
> Best regards
> Akash
>
>
> > /* Flush+invalidate RW caches, invalidate RO ones. */
> > @@ -596,6 +597,16 @@ static int panthor_mmu_as_disable(struct
> > panthor_device *ptdev, u32 as_nr,
> > if (ret)
> > return ret;
> >
> > + if (vm && vm->locked_region.size) {
> > + /* Unlock the region if there a lock pending. */
> > + ret = as_send_cmd_and_wait(ptdev, vm->as.id, AS_COMMAND_UNLOCK);
> > + if (ret)
> > + return ret;
> > +
> > + vm->locked_region.start = 0;
> > + vm->locked_region.size = 0;
> > + }
> > +
> > /* If the slot is going to be used immediately, don't bother changing
> > * the config.
> > */
