00:22
grming has quit [Quit: Konversation terminated!]
00:23
ftg has quit [Read error: Connection reset by peer]
00:38
dikiy has joined #linux-sunxi
01:08
apritzel has quit [Ping timeout: 480 seconds]
02:09
cnxsoft has joined #linux-sunxi
02:24
Daanct12 has joined #linux-sunxi
02:49
dikiy has quit [Ping timeout: 480 seconds]
02:54
Daanct12 has quit [Ping timeout: 480 seconds]
03:01
vagrantc has quit [Quit: leaving]
04:35
jernej_ has joined #linux-sunxi
04:35
jernej has quit [Read error: Connection reset by peer]
05:30
JohnDoe_71Rus has joined #linux-sunxi
06:02
jernej_ is now known as jernej
06:14
apritzel has joined #linux-sunxi
06:58
apritzel has quit [Ping timeout: 480 seconds]
07:22
<
smaeul >
apritzel: You could send the vendor/board bindings and DTs without USB nodes, and add those in the USB series, so none of the other series are holding up the new boards
07:24
Daanct12 has joined #linux-sunxi
07:30
<
wens >
anyone know what the "VE_R" port in the H6 IOMMU is for? I seem to have stepped on an IOMMU error on port 1
07:30
<
wens >
also, the Hantro block is missing an IOMMU reference. I will send a patch for that.
07:44
<
smaeul >
is VE_R not the hantro block?
07:51
Daanct12 has quit [Quit: Quitting]
08:25
<
wens >
no, hantro is the "VP9" block, id = 5
08:25
<
wens >
there's VE: port = 3, and VE_R: port = 1
09:24
apritzel has joined #linux-sunxi
09:36
<
wens >
sun50i-iommu 30f0000.iommu: Page fault for 0x00000000fcbd34c0 (master 1, dir rd)
09:36
<
wens >
cedrus 1c0e000.video-codec: frame processing timed out!
09:36
<
wens >
this is what I'm getting
09:38
<
wens >
difference being that the Chrome VDA test is using MEMORY_TYPE_DMABUF on the capture side
09:43
apritzel has quit [Ping timeout: 480 seconds]
09:45
<
wens >
or maybe the Chrome stack is doing something else funny :/
10:15
apritzel has joined #linux-sunxi
10:35
montjoie_ has joined #linux-sunxi
10:37
montjoie has quit [Ping timeout: 480 seconds]
10:40
bauen1 has quit [Ping timeout: 480 seconds]
11:48
bauen1 has joined #linux-sunxi
11:58
evgeny_boger has quit [Ping timeout: 480 seconds]
12:16
grming has joined #linux-sunxi
13:13
DasChaos has joined #linux-sunxi
13:13
DasChaos has left #linux-sunxi [#linux-sunxi]
13:17
bauen1_ has joined #linux-sunxi
13:19
bauen1 has quit [Ping timeout: 480 seconds]
13:21
evgeny_boger has joined #linux-sunxi
13:23
cnxsoft1 has joined #linux-sunxi
13:24
JohnDoe_71Rus has quit []
13:25
cnxsoft has quit [Read error: Connection reset by peer]
13:38
montjoie_ has quit [Ping timeout: 480 seconds]
13:41
montjoie has joined #linux-sunxi
13:49
indy has quit [Ping timeout: 480 seconds]
13:54
indy has joined #linux-sunxi
14:08
evgeny_boger has quit [Ping timeout: 480 seconds]
14:10
gnarface has quit [Read error: Connection reset by peer]
14:11
cnxsoft1 has quit []
14:11
gnarface has joined #linux-sunxi
14:45
JohnDoe_71Rus has joined #linux-sunxi
15:11
dikiy has joined #linux-sunxi
15:22
dikiy has quit [Ping timeout: 480 seconds]
15:26
bauen1 has joined #linux-sunxi
15:28
bauen1_ has quit [Ping timeout: 480 seconds]
15:38
grming has quit [Quit: Konversation terminated!]
15:45
grming has joined #linux-sunxi
16:19
bauen1 has quit [Ping timeout: 480 seconds]
16:22
<
jernej >
wens, smaeul: Cedrus on H6 actually needs both iommu channels
16:23
<
jernej >
that become obvious during HEVC and IOMMU development
16:23
<
jernej >
I got page faults from both channels
16:23
<
jernej >
it seems that MPEG2 and VP8 use only one channel and H264 and HEVC both channels
16:24
<
jernej >
wens: iommu was useless for Cedrus until I fixed all issues with it
16:24
<
jernej >
so you should pick all patches from linux-next, if you don't have it
16:27
<
jernej >
most of them will be merged for 6.2
16:27
pabs has quit [Ping timeout: 480 seconds]
16:30
pabs has joined #linux-sunxi
16:42
bauen1 has joined #linux-sunxi
16:54
bauen1 has quit [Ping timeout: 480 seconds]
17:07
bauen1 has joined #linux-sunxi
17:21
<
wens >
I've picked all the iommu and cedrus patches
17:22
<
wens >
so I guess I need the dts patch to add the second channel
17:26
dikiy has joined #linux-sunxi
17:35
<
jernej >
wens: not really, since it currently doesn't matter
17:36
<
jernej >
if iommu driver is loaded, all iommu channels are enabled
17:37
<
jernej >
callback to enable iommu channel based on DT values is not implemented (yet)
17:37
<
jernej >
wens: but DT fixes, like in hantro case, are appreciated :)
18:06
apritzel has quit [Ping timeout: 480 seconds]
18:08
dikiy has quit [Ping timeout: 480 seconds]
18:12
w0ll3q1uszxabiwo7vet7op4fhjk84 has joined #linux-sunxi
19:45
apritzel has joined #linux-sunxi
19:48
JohnDoe_71Rus has quit []
20:11
bauen1 has quit [Ping timeout: 480 seconds]
20:14
bauen1 has joined #linux-sunxi
21:28
dikiy has joined #linux-sunxi
21:43
vagrantc has joined #linux-sunxi
23:00
grming has quit [Quit: Konversation terminated!]
23:15
<
smaeul >
it will matter as soon as I send v2 of the D1 IOMMU series, since D1 enables bypass by default, and I was told to disable it per channel
23:16
<
smaeul >
so presumably we would enable bypass by default