Re: [PATCH] soc: ti: k3-ringacc: Fix access mode for k3_ringacc_ring_pop_tail_io()

[Siddharth Vadapalli]

On 21/04/26 21:20, Hari Prasath G E wrote:
> Hello Siddharth,
>
> Thanks for the patch.
>
> On 4/13/2026 12:21 PM, Siddharth Vadapalli wrote:
> > k3_ringacc_ring_pop_tail_io() invokes k3_ringacc_ring_access_io() with the
> > access mode incorrectly set to K3_RINGACC_ACCESS_MODE_POP_HEAD instead of
> > K3_RINGACC_ACCESS_MODE_POP_TAIL. Fix this.
> >
> > Fixes: 3277e8aa2504 ("soc: ti: k3: add navss ringacc driver")
> > Cc: <stable@xxxxxxxxxxxxxxx>
> > Signed-off-by: Siddharth Vadapalli <s-vadapalli@xxxxxx>
> > ---
> >
> > Hello,
> >
> > This patch is based on commit
> > 028ef9c96e96 Linux 7.0
> > of Mainline Linux.
> >
> > I noticed (visually) the incorrect access mode while working on:
> > https://lore.kernel.org/r/20260325123850.638748-1-s-vadapalli@xxxxxx/
> >
> > Regards,
> > Siddharth.
> >
> >   drivers/soc/ti/k3-ringacc.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/soc/ti/k3-ringacc.c b/drivers/soc/ti/k3-ringacc.c
> > index 7602b8a909b0..24f658e8c1dc 100644
> > --- a/drivers/soc/ti/k3-ringacc.c
> > +++ b/drivers/soc/ti/k3-ringacc.c
> > @@ -1083,7 +1083,7 @@ static int k3_ringacc_ring_pop_io(struct k3_ring 
> > *ring, void *elem)
> >   static int k3_ringacc_ring_pop_tail_io(struct k3_ring *ring, void *elem)
> >   {
> >       return k3_ringacc_ring_access_io(ring, elem,
> > -                     K3_RINGACC_ACCESS_MODE_POP_HEAD);
> > +                     K3_RINGACC_ACCESS_MODE_POP_TAIL);
>
> I see that you have noticed this visually and fixed this,was there any 
> impact you faced without this change like data corruption or something ?

I haven't tested it with / without the fix above. But if I have to guess, 
the impact will be out-of-order completions for transmit descriptors and 
out-of-order of packets on receive. Although I have described it in the 
context of Networking, the out-of-order (reversed order to be precise) 
issue will be faced by any user.

> It would be better to mention the impact this change brings-in by doing 
> some analysis.
>
> There is a similar function k3_ringacc_ring_pop_tail_proxy() few lines 
> above where the same change might be required.

Thank you for pointing it out. I will fix that function as well in the v2 
patch.

Regards,
Siddharth.