nobe has quit [Read error: Connection reset by peer]
lurchi_ is now known as lurchi__
lurchi__ is now known as lurchi_
lurchi_ has quit [Read error: Connection reset by peer]
lurchi__ has joined #linux-rockchip
tllim has joined #linux-rockchip
lurchi__ has quit [Quit: Konversation terminated!]
lurchi__ has joined #linux-rockchip
afaerber has quit [Ping timeout: 240 seconds]
afaerber has joined #linux-rockchip
gnufan has quit [Ping timeout: 248 seconds]
gnufan has joined #linux-rockchip
nots has quit [Quit: Page closed]
lurchi_ has joined #linux-rockchip
lurchi__ has quit [Ping timeout: 248 seconds]
vstehle has joined #linux-rockchip
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft has joined #linux-rockchip
tl_lim has joined #linux-rockchip
wadim_ has joined #linux-rockchip
tllim has quit [Ping timeout: 260 seconds]
xerpi has joined #linux-rockchip
tl_lim has quit [Ping timeout: 276 seconds]
tl_lim has joined #linux-rockchip
tl_lim has quit [Read error: Connection reset by peer]
matthias_bgg has joined #linux-rockchip
wouterstreamit has joined #linux-rockchip
<wouterstreamit>
I understand there is a V4L2 driver for the VPU, how does this differ from the MPP? Can I use the V4L2 driver to HW accelerate video decoding with Chromium?
BenG83 has quit [Ping timeout: 260 seconds]
BenG83 has joined #linux-rockchip
poulecaca has joined #linux-rockchip
nighty- has quit [Quit: Disappears in a puff of smoke]
vstehle has quit [Ping timeout: 248 seconds]
xerpi has quit [Quit: Leaving]
afaerber has quit [Quit: Leaving]
nighty- has joined #linux-rockchip
matthias_bgg has quit [Ping timeout: 248 seconds]
afaerber has joined #linux-rockchip
<phh>
wouterstreamit: if i'm not mistaken, the v4l2 driver is no longer supported (since August ~). chromium OS still uses it I think, so I guess it is available to chromium as well? though it can only handle h264 iirc
<wouterstreamit>
phh: Thanks, why was the decision made to develop a custom decoding API if a generic one supported by a good number of programs is available?
<phh>
wouterstreamit: because there is no generic API for this. the API has been submitted to v4l2 team, and it has been refused, because it scales really poorly to multiple codecs
<wouterstreamit>
I have been trying to get the ffmpeg rkmppdec to work with chromium (by adding a video decode accelerator that interacts with ffmpeg's MPP binding, based on the code in Chromium's ffmpeg and DRM bindings) but I'm struggling
cnxsoft has quit [Read error: Connection reset by peer]
<phh>
ayaka: ^^^ I guess you already had customers asking for some mpp/webkit integration?
cnxsoft has joined #linux-rockchip
<phh>
wouterstreamit: ah, I got a reminder from a colleague. I think chromium is using libva, and there is a rockchip libva/v4l2 binding
<wouterstreamit>
Could you point me to some more details of this libva binding? I haven't been able to find anything
<wouterstreamit>
Or should everything work out of the box with the rockchip-linux kernel and standard libva?
matthias_bgg has joined #linux-rockchip
<phh>
uh, looks like they removed it?
<phh>
but no, it's not enough to just have libva, you need rockchip's libva/v4l2 connector
<Ke>
yay, I got picture on screen on Kevin with 4.16-rc1 =o)
<Ke>
it even apparently works with the initrd I got from my rk3288 C201
<maz>
Esmil: I have that in my tree, and it doesn't help.
<Esmil>
:*
<Esmil>
:(
<maz>
Esmil: it does work with 4.15 though. so there is a 4.16 regression.
<Esmil>
I see
<mmind00>
which is interesting as 4.15-4.16 was especially tiny changeset-wise
<mmind00>
related to kevin
BenG83 has quit [Quit: Leaving]
<Esmil>
maz: Did you play with arm's mali driver?
<maz>
Esmil: never. I refuse to get anywhere near this thing.
<Esmil>
maz: I made the r22p0 version compile and boot yesterday, but I still need to install the closed source userspace part to check that it actually works
<Esmil>
Yeah, it's super annoying but it would be nice for browsing a bit more smoothly I guess
<maz>
mmind00: it may not be related to kevin itself, but to the interrupt nightmare in the VOP driver.
<maz>
mmind00: I guess something may have change in the DRM layer that triggers this.
<maz>
Esmil: I don't care much for browsing, and I certainly do not want that ugly piece of close source on my system.
<maz>
mmind00: do we have any public information on the VOP, and specially its interrupts as well as its relationship with the IOMMU interrupt?
<Esmil>
maz: :)
<maz>
mmind00: the bits of TRM I've found are really unhelpful...
droman0 has quit []
<mmind00>
maz: in such detail the TRMs are unhelpful most of the time
<mmind00>
sadly
droman0 has joined #linux-rockchip
<mmind00>
maz: i.e. if you found a TRM containing the VOP description, that is the only level of documentation you will get most of the time without pestering Rockchip developers
<maz>
mmind00: great. I guess I'll need to reverse engineer some of it by looking at the code...
<mmind00>
maz: but I guess talking to for example Sean Paul (on dri-devel for example) might be helpful, as it seems he is working on drm both upstream and in chromeos (including the rk3399)
<mmind00>
maz: rockchip drm driver has diverged quite a bit again between chromeos and mainline, so possibly they have seen your issue already
lurchi_ is now known as lurchi__
<maz>
mmind00: I need to spend some quality time with this interrupt problem, because I see really bizarre things. Once I've investigated it, I'll write it up and post that.
vstehle has joined #linux-rockchip
tllim has joined #linux-rockchip
<Ke>
Esmil: your tree is based on 4.10?
<Esmil>
Ke: You need the kevin branch
<Ke>
ah true
<Ke>
looking at the log there is loads of stuff for kevin
<Ke>
so it could be not just my config
<Esmil>
yeah, but I don't there is anything relevant for the sd-card
<Ke>
well ok only 5000ish lines
<Ke>
a2c010927bac phy: rockchip-emmc: use regmap_read_poll_timeout to poll dllrdy
<Esmil>
yeah, that's the 43 long patch series Thierry sent
<Esmil>
I thought those two emmc patches were for the on-board mmc. kevin confusingly has two different mmc controllers. One is connected to the on-board mmc and one to the sd-card
<Esmil>
different as in needs different drivers. not just 2 instances of the same controller
<Ke>
yes
<Ke>
though the driver is probably not my problem, but some auxiliary thing
<Ke>
whatever voltage or frequency controller or iommu or something
<Esmil>
yeah which is why my first guess would be that you just forgot something in your .config
<Ke>
but I am actually checking all the compatibles from the device tree
<Ke>
be it only with a hacky script
<Ke>
some things obviously are not in device tree though
BenG83 has joined #linux-rockchip
Easyfab has joined #linux-rockchip
lkcl has joined #linux-rockchip
lkcl_ has joined #linux-rockchip
lkcl has quit [Ping timeout: 256 seconds]
lkcl_ has quit [Ping timeout: 256 seconds]
wouterstreamit has quit [Ping timeout: 260 seconds]