| Message ID | 1534328897-14957-1-git-send-email-jacopo+renesas@jmondi.org (mailing list archive) |
|---|---|
| Headers | show |
| Series | media: i2c: ov5640: Re-work MIPI startup sequence | expand |
On 15 August 2018 at 12:28, Jacopo Mondi <jacopo+renesas@jmondi.org> wrote: > Hello ov5640 people, > this driver has received a lot of attention recently, and this series aims > to fix the CSI-2 interface startup on i.Mx6Q platforms. > > Please refer to the v2 cover letters for more background informations: > https://www.mail-archive.com/linux-media@vger.kernel.org/msg133420.html > > This two patches alone allows the MIPI interface to startup properly, but in > order to capture good images (good as in 'not completely black') exposure and > gain handling should be fixed too. > Hugues Fruchet has a series in review that fixes that issues: > [PATCH v2 0/5] Fix OV5640 exposure & gain > > And I have re-based it on top of this two fixes here: > git://jmondi.org/linux ov5640/timings_exposure > > Steve Longerbeam tested that branch on his I.MX6q SabreSD board and confirms he > can now capture frames (I added his Tested-by tag to this patches). I have > verified the same on Engicam iCore I.MX6q and an Intel Atom based board. > > Ideally I would like to have these two fixes merged, and Hugues' ones then > applied on top. Of course, more testing on other platforms using CSI-2 is very > welcome. > > Thanks > j > > v2 -> v3: > - patch [2/2] was originally sent in a different series, compared to v2 it > removes entries from the blob array instead of adding more. > > Jacopo Mondi (2): > media: ov5640: Re-work MIPI startup sequence > media: ov5640: Fix timings setup code > > drivers/media/i2c/ov5640.c | 141 +++++++++++++++++++++++++++++---------------- > 1 file changed, 92 insertions(+), 49 deletions(-) > > -- > 2.7.4 > Thanks for this work. I've just tested this with a dragonboard-410c (MICPI/CSI) + OV5640 sensor. It works on my side for 1280*720, 1920*1080 and 2592*1944 formats. Tested-by: Loic Poulain <loic.poulain@linaro.org>
Hello Loic, On Tue, Aug 28, 2018 at 05:08:07PM +0200, Loic Poulain wrote: > On 15 August 2018 at 12:28, Jacopo Mondi <jacopo+renesas@jmondi.org> wrote: > > Hello ov5640 people, > > this driver has received a lot of attention recently, and this series aims > > to fix the CSI-2 interface startup on i.Mx6Q platforms. > > > > Please refer to the v2 cover letters for more background informations: > > https://www.mail-archive.com/linux-media@vger.kernel.org/msg133420.html > > > > This two patches alone allows the MIPI interface to startup properly, but in > > order to capture good images (good as in 'not completely black') exposure and > > gain handling should be fixed too. > > Hugues Fruchet has a series in review that fixes that issues: > > [PATCH v2 0/5] Fix OV5640 exposure & gain > > > > And I have re-based it on top of this two fixes here: > > git://jmondi.org/linux ov5640/timings_exposure > > > > Steve Longerbeam tested that branch on his I.MX6q SabreSD board and confirms he > > can now capture frames (I added his Tested-by tag to this patches). I have > > verified the same on Engicam iCore I.MX6q and an Intel Atom based board. > > > > Ideally I would like to have these two fixes merged, and Hugues' ones then > > applied on top. Of course, more testing on other platforms using CSI-2 is very > > welcome. > > > > Thanks > > j > > > > v2 -> v3: > > - patch [2/2] was originally sent in a different series, compared to v2 it > > removes entries from the blob array instead of adding more. > > > > Jacopo Mondi (2): > > media: ov5640: Re-work MIPI startup sequence > > media: ov5640: Fix timings setup code > > > > drivers/media/i2c/ov5640.c | 141 +++++++++++++++++++++++++++++---------------- > > 1 file changed, 92 insertions(+), 49 deletions(-) > > > > -- > > 2.7.4 > > > > Thanks for this work. > I've just tested this with a dragonboard-410c (MICPI/CSI) + OV5640 sensor. > It works on my side for 1280*720, 1920*1080 and 2592*1944 formats. > > Tested-by: Loic Poulain <loic.poulain@linaro.org> Thanks for testing! Just out of curiosity, did the dragonboard-410c CSI-2 interface had issues poperly starting up before this series like the i.MX6q did? Thanks j