[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CADL8D3ZcvynQCGLCcbK=U9-2WB758abLcKaNkTtXN8Y7s=dyqQ@mail.gmail.com>
Date: Mon, 21 Oct 2024 11:21:32 -0400
From: Jon Cormier <jcormier@...ticallink.com>
To: Tomi Valkeinen <tomi.valkeinen@...asonboard.com>
Cc: Devarsh Thakkar <devarsht@...com>, Jyri Sarha <jyri.sarha@....fi>,
Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>, Maxime Ripard <mripard@...nel.org>,
Thomas Zimmermann <tzimmermann@...e.de>, David Airlie <airlied@...il.com>, Daniel Vetter <daniel@...ll.ch>,
dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
Bin Liu <b-liu@...com>, stable@...r.kernel.org
Subject: Re: [PATCH 1/7] drm/tidss: Fix issue in irq handling causing
irq-flood issue
On Mon, Oct 21, 2024 at 10:08 AM Tomi Valkeinen
<tomi.valkeinen@...asonboard.com> wrote:
>
> It has been observed that sometimes DSS will trigger an interrupt and
> the top level interrupt (DISPC_IRQSTATUS) is not zero, but the VP and
> VID level interrupt-statuses are zero.
>
> As the top level irqstatus is supposed to tell whether we have VP/VID
> interrupts, the thinking of the driver authors was that this particular
> case could never happen. Thus the driver only clears the DISPC_IRQSTATUS
> bits which has corresponding interrupts in VP/VID status. So when this
> issue happens, the driver will not clear DISPC_IRQSTATUS, and we get an
> interrupt flood.
>
> It is unclear why the issue happens. It could be a race issue in the
> driver, but no such race has been found. It could also be an issue with
> the HW. However a similar case can be easily triggered by manually
> writing to DISPC_IRQSTATUS_RAW. This will forcibly set a bit in the
> DISPC_IRQSTATUS and trigger an interrupt, and as the driver never clears
> the bit, we get an interrupt flood.
>
> To fix the issue, always clear DISPC_IRQSTATUS. The concern with this
> solution is that if the top level irqstatus is the one that triggers the
> interrupt, always clearing DISPC_IRQSTATUS might leave some interrupts
> unhandled if VP/VID interrupt statuses have bits set. However, testing
> shows that if any of the irqstatuses is set (i.e. even if
> DISPC_IRQSTATUS == 0, but a VID irqstatus has a bit set), we will get an
> interrupt.
>
> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@...asonboard.com>
> Co-developed-by: Bin Liu <b-liu@...com>
> Signed-off-by: Bin Liu <b-liu@...com>
> Co-developed-by: Devarsh Thakkar <devarsht@...com>
> Signed-off-by: Devarsh Thakkar <devarsht@...com>
> Co-developed-by: Jonathan Cormier <jcormier@...ticallink.com>
> Signed-off-by: Jonathan Cormier <jcormier@...ticallink.com>
> Fixes: 32a1795f57ee ("drm/tidss: New driver for TI Keystone platform Display SubSystem")
> Cc: stable@...r.kernel.org
> ---
> drivers/gpu/drm/tidss/tidss_dispc.c | 12 ++++--------
> 1 file changed, 4 insertions(+), 8 deletions(-)
I assume a reviewed by doesn't make sense since I co-developed this
patch but adding my tested by, hopefully, that makes sense.
Tested an equivalent patch for several weeks.
Tested-by: Jonathan Cormier <jcormier@...ticallink.com>
>
> diff --git a/drivers/gpu/drm/tidss/tidss_dispc.c b/drivers/gpu/drm/tidss/tidss_dispc.c
> index 1ad711f8d2a8..f81111067578 100644
> --- a/drivers/gpu/drm/tidss/tidss_dispc.c
> +++ b/drivers/gpu/drm/tidss/tidss_dispc.c
> @@ -780,24 +780,20 @@ static
> void dispc_k3_clear_irqstatus(struct dispc_device *dispc, dispc_irq_t clearmask)
> {
> unsigned int i;
> - u32 top_clear = 0;
>
> for (i = 0; i < dispc->feat->num_vps; ++i) {
> - if (clearmask & DSS_IRQ_VP_MASK(i)) {
> + if (clearmask & DSS_IRQ_VP_MASK(i))
> dispc_k3_vp_write_irqstatus(dispc, i, clearmask);
> - top_clear |= BIT(i);
> - }
> }
> for (i = 0; i < dispc->feat->num_planes; ++i) {
> - if (clearmask & DSS_IRQ_PLANE_MASK(i)) {
> + if (clearmask & DSS_IRQ_PLANE_MASK(i))
> dispc_k3_vid_write_irqstatus(dispc, i, clearmask);
> - top_clear |= BIT(4 + i);
> - }
> }
> if (dispc->feat->subrev == DISPC_K2G)
> return;
>
> - dispc_write(dispc, DISPC_IRQSTATUS, top_clear);
> + /* always clear the top level irqstatus */
> + dispc_write(dispc, DISPC_IRQSTATUS, dispc_read(dispc, DISPC_IRQSTATUS));
>
> /* Flush posted writes */
> dispc_read(dispc, DISPC_IRQSTATUS);
>
> --
> 2.43.0
>
Powered by blists - more mailing lists