diff options
author | Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> | 2022-04-26 09:06:18 +0200 |
---|---|---|
committer | Mauro Carvalho Chehab <mchehab@kernel.org> | 2022-08-30 14:27:05 +0200 |
commit | d91105e0567dabf487f81019e8b609706334d5ed (patch) | |
tree | d06893fa44abc03bb213b9d7583751724bebc774 | |
parent | 0872dc04cf65fdfaa1d67453d535f8181982848b (diff) | |
download | linux-d91105e0567dabf487f81019e8b609706334d5ed.tar.bz2 |
media: ti: cal: fix wdma irq for metadata
CAL HW interrupts are inherently racy. If we get both start and end
interrupts, we don't know what has happened: did the DMA for a single
frame start and end, or did one frame end and a new frame start?
Usually for normal pixel frames we get the interrupts separately. If
we do get both, we have to guess. The assumption in the code is that the
active vertical area is larger than the blanking vertical area, and thus
it is more likely that we get the end of the old frame and the start of
a new frame.
However, for embedded data, which is only a few lines high, we always
get both interrupts. Here the assumption is that we get both for the
same frame.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
-rw-r--r-- | drivers/media/platform/ti/cal/cal.c | 31 |
1 files changed, 27 insertions, 4 deletions
diff --git a/drivers/media/platform/ti/cal/cal.c b/drivers/media/platform/ti/cal/cal.c index 910ff179e597..56b61c0583cf 100644 --- a/drivers/media/platform/ti/cal/cal.c +++ b/drivers/media/platform/ti/cal/cal.c @@ -717,11 +717,34 @@ static inline void cal_irq_wdma_end(struct cal_ctx *ctx) static void cal_irq_handle_wdma(struct cal_ctx *ctx, bool start, bool end) { - if (end) - cal_irq_wdma_end(ctx); + /* + * CAL HW interrupts are inherently racy. If we get both start and end + * interrupts, we don't know what has happened: did the DMA for a single + * frame start and end, or did one frame end and a new frame start? + * + * Usually for normal pixel frames we get the interrupts separately. If + * we do get both, we have to guess. The assumption in the code below is + * that the active vertical area is larger than the blanking vertical + * area, and thus it is more likely that we get the end of the old frame + * and the start of a new frame. + * + * However, for embedded data, which is only a few lines high, we always + * get both interrupts. Here the assumption is that we get both for the + * same frame. + */ + if (ctx->v_fmt.fmt.pix.height < 10) { + if (start) + cal_irq_wdma_start(ctx); - if (start) - cal_irq_wdma_start(ctx); + if (end) + cal_irq_wdma_end(ctx); + } else { + if (end) + cal_irq_wdma_end(ctx); + + if (start) + cal_irq_wdma_start(ctx); + } } static irqreturn_t cal_irq(int irq_cal, void *data) |