Wow, a patchset from 2020.
On 25/05/2022 12:03, Vladimir Zapolskiy wrote:
On 10/22/20 20:47, kholk11@xxxxxxxxx wrote:
From: AngeloGioacchino Del Regno <kholk11@xxxxxxxxx>
Resetting the ISPIF VFE0 context is wrong if we are using the VFE1
for dual-camera or simply because a secondary camera is connected
to it: in this case the reset will always happen on the VFE0 ctx
of the ISPIF, which is .. useless.
Fix this usecase by adding the ISPIF_RST_CMD_1 address and choose
where to do the (or what to) reset based on the VFE line id.
Signed-off-by: AngeloGioacchino Del Regno <kholk11@xxxxxxxxx>
Reviewed-by: Robert Foss <robert.foss@xxxxxxxxxx>
---
.../media/platform/qcom/camss/camss-ispif.c | 85 ++++++++++++-------
.../media/platform/qcom/camss/camss-ispif.h | 2 +-
2 files changed, 56 insertions(+), 31 deletions(-)
[skipped]
@@ -355,7 +379,7 @@ static int ispif_set_power(struct v4l2_subdev *sd,
int on)
goto exit;
}
- ret = ispif_reset(ispif);
+ ret = ispif_reset(ispif, line->vfe_id);
But in fact here is an error.
line->vfe_id is never set.
I'm unable to test any fix, since I don't have a correspondent hardware,
but I can write a fix for someone's testing.
I have a sda660 device, but I don't have cameras attached. So one will
have to use TPG.
--
With best wishes
Dmitry
