| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Thu, 13 Jun 2024 00:01:44 +0200
From: Mehdi Djait <mehdi.djait.k@...il.com>
To: Val Packett <val@...kett.cool>
Cc: mchehab@...nel.org, heiko@...ech.de, hverkuil-cisco@...all.nl,
krzysztof.kozlowski+dt@...aro.org, robh+dt@...nel.org,
conor+dt@...nel.org, linux-media@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
thomas.petazzoni@...tlin.com, alexandre.belloni@...tlin.com,
maxime.chevallier@...tlin.com, paul.kocialkowski@...tlin.com,
michael.riesch@...fvision.net, laurent.pinchart@...asonboard.com,
Mehdi Djait <mehdi.djait@...tlin.com>
Subject: Re: [RESEND Patch v13 2/3] media: rockchip: Add a driver for
Rockchip's camera interface
Hi Val :)
And sorry for the late response!
On Wed, Jun 12, 2024 at 02:23:13AM -0300, Val Packett wrote:
> Hi,
>
> a couple more comments on this driver:
>
> On Sun, Feb 11 2024 at 20:03:31 +01:00:00, Mehdi Djait
> <mehdi.djait.k@...il.com> wrote:
> > +static int cif_stream_start(struct cif_stream *stream)
> > +{
> > + u32 val, fmt_type, xfer_mode = 0;
> > + struct cif_device *cif_dev = stream->cifdev;
> > + struct cif_remote *remote_info = &cif_dev->remote;
> > + int ret;
> > + u32 input_mode;
> > +
> > + stream->frame_idx = 0;
> > + stream->frame_phase = 0;
> > +
> > + fmt_type = stream->cif_fmt_in->fmt_type;
> > + input_mode = (remote_info->std == V4L2_STD_NTSC) ?
> > + CIF_FORMAT_INPUT_MODE_NTSC :
> > + CIF_FORMAT_INPUT_MODE_PAL;
>
> Mode logic needs to be expanded for cameras; I'm trying to get it working
> correctly,
> so far managed to get some cursed selfies with the wrong pixel format :) but
> either
> way I could send a patch when I have it working well.
>
Do you mind sharing which sensor is connected on one end to the camera
and the other end to cif ?
I developed this driver using the tw9900 NTSC/PAL/SECAM Video Decoder
and I know that more patches are needed to make this driver work
with other sensors. Look at the series from Michael Riesch where he
adds support for other sensors on top of this series:
https://lore.kernel.org/linux-media/20240220-v6-8-topic-rk3568-vicap-v1-0-2680a1fa640b@wolfvision.net/
PLEASE NOTE: For the moment being I am not working on this driver as I
don't have access to the hardware.
I need a rockchip board that can boot upstream kernel AND that has
DVP (Digital Video Port) camera interface to continue working on CIF.
I have some candidates but I am still not entirely sure how to proceed
further.
ALSO NOTE: This driver still need some more work as you can see in the
last round of reviews: it should be converted to a media device centric
driver that correctly uses the Media Controller.
Look at the discussion here: https://lore.kernel.org/linux-media/ZeWqEkcoYN6gXWS9@valkosipuli.retiisi.eu/
> > +static int subdev_notifier_complete(struct v4l2_async_notifier
> > *notifier)
> > +{
> > + struct cif_device *cif_dev;
> > + struct v4l2_subdev *sd;
> > + int ret;
> > +
> > + cif_dev = container_of(notifier, struct cif_device, notifier);
> > + sd = cif_dev->remote.sd;
> > +
> > + mutex_lock(&cif_dev->media_dev.graph_mutex);
> > +
>
> Potential deadlock with this lock:
I will look more into this when/if I contine working on this driver
Still I am happy that other people are taking a look at this series and
doing something with it!
--
Kind Regards
Mehdi Djait
>
> [ 3.438667] ======================================================
> [ 3.438672] WARNING: possible circular locking dependency detected
> [ 3.438677] 6.10.0-rc1-next-20240531 #151 Not tainted
> [ 3.438684] ------------------------------------------------------
> [ 3.438687] kworker/u8:1/25 is trying to acquire lock:
> [ 3.438695] c152d41c (videodev_lock){+.+.}-{4:4}, at:
> __video_register_device+0xf4/0x15b0
> [ 3.438737]
> [ 3.438737] but task is already holding lock:
> [ 3.438740] c31981fc (&mdev->graph_mutex){+.+.}-{4:4}, at:
> subdev_notifier_complete+0x20/0x80
> [ 3.438765]
> [ 3.438765] which lock already depends on the new lock.
> [ 3.438765]
> [ 3.438769]
> [ 3.438769] the existing dependency chain (in reverse order) is:
> [ 3.438772]
> [ 3.438772] -> #1 (&mdev->graph_mutex){+.+.}-{4:4}:
> [ 3.438786] lock_acquire+0x110/0x374
> [ 3.438809] __mutex_lock+0xac/0xf2c
> [ 3.438828] mutex_lock_nested+0x1c/0x24
> [ 3.438843] media_device_register_entity+0x80/0x1e8
> [ 3.438857] __video_register_device+0xab0/0x15b0
> [ 3.438869] cif_register_stream_vdev+0x158/0x18c
> [ 3.438880] cif_plat_probe+0x20c/0x424
> [ 3.438888] platform_probe+0x5c/0xb0
> [ 3.438905] really_probe+0xc8/0x2cc
> [ 3.438916] __driver_probe_device+0x88/0x1a0
> [ 3.438926] driver_probe_device+0x30/0x108
> [ 3.438936] __driver_attach+0x94/0x184
> [ 3.438946] bus_for_each_dev+0x7c/0xcc
> [ 3.438955] bus_add_driver+0xcc/0x1ec
> [ 3.438964] driver_register+0x7c/0x114
> [ 3.438975] do_one_initcall+0x7c/0x2f8
> [ 3.438989] kernel_init_freeable+0x1b0/0x20c
> [ 3.439010] kernel_init+0x14/0x12c
> [ 3.439024] ret_from_fork+0x14/0x28
> [ 3.439033]
> [ 3.439033] -> #0 (videodev_lock){+.+.}-{4:4}:
> [ 3.439048] check_prev_add+0x134/0x17d8
> [ 3.439065] __lock_acquire+0x17e0/0x21fc
> [ 3.439080] lock_acquire+0x110/0x374
> [ 3.439095] __mutex_lock+0xac/0xf2c
> [ 3.439109] mutex_lock_nested+0x1c/0x24
> [ 3.439123] __video_register_device+0xf4/0x15b0
> [ 3.439135] __v4l2_device_register_subdev_nodes+0xd8/0x1a0
> [ 3.439152] subdev_notifier_complete+0x2c/0x80
> [ 3.439160] __v4l2_async_register_subdev+0xa8/0x1a0
> [ 3.439176] gc0308_probe+0x654/0x6f4
> [ 3.439187] i2c_device_probe+0x168/0x268
> [ 3.439201] really_probe+0xc8/0x2cc
> [ 3.439211] __driver_probe_device+0x88/0x1a0
> [ 3.439221] driver_probe_device+0x30/0x108
> [ 3.439231] __device_attach_driver+0x94/0x10c
> [ 3.439241] bus_for_each_drv+0x90/0xe4
> [ 3.439250] __device_attach+0xac/0x1b0
> [ 3.439260] bus_probe_device+0x8c/0x90
> [ 3.439269] deferred_probe_work_func+0x7c/0xac
> [ 3.439279] process_one_work+0x23c/0x6bc
> [ 3.439295] worker_thread+0x190/0x3d8
> [ 3.439308] kthread+0xf8/0x114
> [ 3.439321] ret_from_fork+0x14/0x28
> [ 3.439330]
> [ 3.439330] other info that might help us debug this:
> [ 3.439330]
> [ 3.439333] Possible unsafe locking scenario:
> [ 3.439333]
> [ 3.439336] CPU0 CPU1
> [ 3.439339] ---- ----
> [ 3.439341] lock(&mdev->graph_mutex);
> [ 3.439349] lock(videodev_lock);
> [ 3.439356] lock(&mdev->graph_mutex);
> [ 3.439363] lock(videodev_lock);
> [ 3.439370]
> [ 3.439370] *** DEADLOCK ***
> [ 3.439370]
> [ 3.439373] 5 locks held by kworker/u8:1/25:
> [ 3.439379] #0: c28106b4 ((wq_completion)events_unbound){+.+.}-{0:0},
> at: process_one_work+0x1ac/0x6bc
> [ 3.439408] #1: f0889f20 (deferred_probe_work){+.+.}-{0:0}, at:
> process_one_work+0x1d4/0x6bc
> [ 3.439436] #2: c303ec9c (&dev->mutex){....}-{4:4}, at:
> __device_attach+0x30/0x1b0
> [ 3.439461] #3: c152d3d0 (list_lock){+.+.}-{4:4}, at:
> __v4l2_async_register_subdev+0x50/0x1a0
> [ 3.439491] #4: c31981fc (&mdev->graph_mutex){+.+.}-{4:4}, at:
> subdev_notifier_complete+0x20/0x80
>
> >
>
>
Powered by blists - more mailing lists