| + S0rin (~S0rin@82-64-116-66.subs.proxad.net) | 00:08 | |
| - S0rin (QUIT: Ping timeout: 264 seconds) (~S0rin@82-64-116-66.subs.proxad.net) | 00:14 | |
| + S0rin (~S0rin@82-64-116-66.subs.proxad.net) | 00:16 | |
| - freakazoid333 (QUIT: Read error: Connection reset by peer) (~freakazoi@2001:5b0:2a65:fa88:42f:154a:842c:dca2) | 00:30 | |
| + freakazoid333 (~freakazoi@2001:5b0:2a65:fa88:42f:154a:842c:dca2) | 00:30 | |
| - adjtm_ (QUIT: Ping timeout: 272 seconds) (~adjtm@188.26.215.143) | 02:29 | |
| + adjtm (~adjtm@2a0c:5a80:3d0c:7700:49ef:d0f0:57d6:3082) | 02:41 | |
| - Neelfyn (QUIT: Quit: Connection closed for inactivity) (uid180106@id-180106.stonehaven.irccloud.com) | 05:03 | |
| + specing_ (~specing@APN-123-240-240-gprs.simobil.net) | 06:28 | |
| - specing (QUIT: Ping timeout: 245 seconds) (~specing@user/specing) | 06:29 | |
| - adjtm (QUIT: Ping timeout: 244 seconds) (~adjtm@2a0c:5a80:3d0c:7700:49ef:d0f0:57d6:3082) | 09:20 | |
| + mjw (~mjw_@2001:1c06:2488:200:9e5c:8eff:fe8f:a440) | 10:07 | |
| + patrick (~patrick@2001:8b0:ca70:32b2:428d:5cff:fea4:5806) | 11:04 | |
| + rasmus-mk (~rasmus@c83-253-223-217.bredband.tele2.se) | 11:06 | |
| - patrick (QUIT: Quit: WeeChat 3.0.1) (~patrick@2001:8b0:ca70:32b2:428d:5cff:fea4:5806) | 11:09 | |
| - jryans (QUIT: Ping timeout: 244 seconds) (~jryansmat@2001:470:69fc:105::1d) | 11:20 | |
| + ephase (~ephase@185-82-190-109.dsl.ovh.fr) | 11:46 | |
| ephase | hello | 11:46 |
|---|---|---|
| + Neelfyn (uid180106@id-180106.stonehaven.irccloud.com) | 11:53 | |
| + jomo7 (~jomo@user/jomo) | 11:55 | |
| - FUZxxl (QUIT: *.net *.split) (~fuz@fuz.su) | 12:01 | |
| - marlun (QUIT: *.net *.split) (~marlun@46.101.246.245) | 12:01 | |
| - jomo (QUIT: *.net *.split) (~jomo@user/jomo) | 12:01 | |
| - marex (QUIT: *.net *.split) (~marex@195.140.253.37) | 12:01 | |
| - jfred (QUIT: *.net *.split) (quassel@libera/sponsor/jfred) | 12:01 | |
| - mntmn (QUIT: *.net *.split) (~mntirc@softboy.mntmn.com) | 12:01 | |
| - _Bnu (QUIT: *.net *.split) (~beeanyew@89-160-120-72.cust.bredband2.com) | 12:01 | |
| * jomo7 -> jomo | 12:01 | |
| + FUZxxl (~fuz@fuz.su) | 12:11 | |
| + marex (~marex@195.140.253.37) | 12:11 | |
| + marlun (~marlun@46.101.246.245) | 12:11 | |
| + jfred (quassel@libera/sponsor/jfred) | 12:11 | |
| + mntmn (~mntirc@softboy.mntmn.com) | 12:11 | |
| + _Bnu (~beeanyew@89-160-120-72.cust.bredband2.com) | 12:11 | |
| - lexik (QUIT: Quit: No Ping reply in 180 seconds.) (~lexik@171.25.222.217) | 13:55 | |
| + lexik (~lexik@171.25.222.217) | 13:58 | |
| - ephase (QUIT: Quit: WeeChat 3.1) (~ephase@185-82-190-109.dsl.ovh.fr) | 14:10 | |
| - lexik (QUIT: Quit: No Ping reply in 180 seconds.) (~lexik@171.25.222.217) | 14:18 | |
| - rasmus-mk (QUIT: Quit: WeeChat 3.0.1) (~rasmus@c83-253-223-217.bredband.tele2.se) | 14:21 | |
| + lexik (~lexik@171.25.222.217) | 14:22 | |
| + jryans (~jryansmat@2001:470:69fc:105::1d) | 14:34 | |
| - cryptix (QUIT: Quit: node-irc says goodbye) (~cryptxxma@2001:470:69fc:105::94a) | 15:24 | |
| - jryans (QUIT: Quit: node-irc says goodbye) (~jryansmat@2001:470:69fc:105::1d) | 15:24 | |
| mntmn | hmm there might be a trick at least to test if reform can revive deep discharged batteries... by disabling the LPC firmware temporarily | 15:25 |
| + jryans (~jryansmat@2001:470:69fc:105::1d) | 15:25 | |
| mntmn | not sure yet if it is safe, but i shall try | 15:25 |
| + cryptix (~cryptxxma@2001:470:69fc:105::94a) | 15:28 | |
| + rasmus-mk (~rasmus@c83-253-223-217.bredband.tele2.se) | 15:31 | |
| specing_ | that should be accompanied with cell disassembly and checking for dendride growth | 15:32 |
| mntmn | optimally yes | 15:37 |
| ezequielg | marex: +1 | 15:41 |
| ezequielg | any luck so far? | 15:41 |
| - rasmus-mk (QUIT: Quit: WeeChat 3.0.1) (~rasmus@c83-253-223-217.bredband.tele2.se) | 15:46 | |
| + rasmus-mk (~rasmus@c83-253-223-217.bredband.tele2.se) | 15:47 | |
| - rasmus-mk (QUIT: Client Quit) (~rasmus@c83-253-223-217.bredband.tele2.se) | 15:47 | |
| + rasmus-mk (~rasmus@c83-253-223-217.bredband.tele2.se) | 15:47 | |
| - rasmus-mk (QUIT: Client Quit) (~rasmus@c83-253-223-217.bredband.tele2.se) | 15:49 | |
| + rasmus-mk (~rasmus@c83-253-223-217.bredband.tele2.se) | 15:49 | |
| - rasmus-mk (QUIT: Quit: WeeChat 3.0.1) (~rasmus@c83-253-223-217.bredband.tele2.se) | 16:07 | |
| - natalie (QUIT: Ping timeout: 244 seconds) (~natalie@user/natalie) | 16:07 | |
| marex | ezequielg: nope | 16:15 |
| marex | ezequielg: I'm hitting G1 watchdog | 16:15 |
| ezequielg | hm. | 16:15 |
| ezequielg | not good. | 16:15 |
| marex | ezequielg: sounds familiar ? | 16:15 |
| ezequielg | well, means no interrupt. | 16:15 |
| marex | ezequielg: well, I do get 1 interrupt before the WDT | 16:16 |
| marex | ezequielg: at least per /proc/interrupts | 16:16 |
| ezequielg | ah. | 16:16 |
| ezequielg | which status? | 16:16 |
| marex | ezequielg: where do I get that information ? patch teh driver and printk it ? | 16:16 |
| ezequielg | yeah, sorry. | 16:17 |
| marex | ezequielg: well no problem, I think I've tried compiling the kernel before ;-) | 16:17 |
| ezequielg | once or twice. | 16:18 |
| marex | ezequielg: yeah, something like that | 16:18 |
| marex | ezequielg: well ... maybe I need to undo some of my power domain digging | 16:23 |
| marex | [ 57.959021] hantro_g1_irq[21] status=80110100 | 16:28 |
| marex | here it is | 16:28 |
| marex | lets see | 16:28 |
| ezequielg | you are doing h264 right? | 16:34 |
| marex | yep | 16:34 |
| ezequielg | SW_DEC_ERROR_INT | 16:35 |
| ezequielg | hm | 16:35 |
| marex | ffmpeg -i big_buck_bunny_1080p_h264.mov -s:v 800x480 -r 25 -c:v libx264 -threads 64 bbb.h264 ... should spit out valid h264 cbp, no ? | 16:35 |
| ezequielg | gstreamer can understand the mov. | 16:35 |
| ezequielg | why you'd use ffmpeg? | 16:36 |
| marex | ezequielg: because I wasn't sure which exact elements to use in gstreamer to feed the slh264dec | 16:36 |
| ezequielg | ah | 16:36 |
| marex | ezequielg: and I transcoded it into smaller resolution too | 16:36 |
| ezequielg | -v filesrc location=$1 ! parsebin ! decodebin3 ! fakevideosink -> one option | 16:37 |
| marex | gst-launch-1.0 filesrc location=big_buck_bunny_1080p_h264.mov ! parsebin ! h264parse ! v4l2slh264dec ! fakevideosink | 16:37 |
| ezequielg | filesrc ! parsebin ! h264parse ! v4l2slh264dec -> another option | 16:37 |
| marex | same thing btw | 16:37 |
| ezequielg | right. | 16:37 |
| ezequielg | what's the SoC again? | 16:38 |
| marex | 8mm | 16:38 |
| marex | I'm about to set up 8mq | 16:38 |
| marex | I mean, there it should work, so I should be able to validate userspace there | 16:38 |
| ezequielg | hm hm, my SDK doesn't have 8mm :P | 16:38 |
| ezequielg | I suppose you have some NXP SDK checkout? | 16:39 |
| marex | ezequielg: what's that ? | 16:39 |
| ezequielg | e.g. https://git.buildroot.net/buildroot/tree/package/freescale-imx/imx-vpu-hantro | 16:40 |
| marex | ezequielg: oh, broken downstream stuff you mean ... no :) | 16:40 |
| marex | ezequielg: next + patches picked from ML | 16:40 |
| ezequielg | that's my go-to-reference to see how things are "supposed to be". | 16:42 |
| marex | ezequielg: ... that's not how things are supposed to be :( | 16:42 |
| ezequielg | haha. in any case, confirming 8mq works would be great. | 16:42 |
| marex | ezequielg: those things are supposed to be upstream, not in some BSP with 1M+ LoC of changes in kernel itself and custom stuff all over the place :-( | 16:43 |
| marex | ezequielg: yeah | 16:43 |
| ezequielg | marex: fwiw, Pull Request for Hantro G2 was _just_ sent. | 16:43 |
| ezequielg | so, it's slowly going upstream. | 16:43 |
| marex | ezequielg: I saw some of that | 16:43 |
| marex | ezequielg: I mean, it should've been done by NXP long ago | 16:44 |
| marex | ezequielg: hm ok, it seems the latest gpcv2 patches broke the VPU power domain on mx8mq , so I need to debug that first | 16:52 |
| ezequielg | are those mainline already? | 16:53 |
| marex | ezequielg: some of it is in next | 16:53 |
| + GBenji (~benjamin@88.123.46.161) | 16:57 | |
| marex | ezequielg: so yeah, on the 8mq I'm getting | 17:06 |
| marex | [ 35.106977] hantro_g1_irq[21] status=1001100 | 17:06 |
| marex | [ 35.146673] hantro_g1_irq[21] status=1100 | 17:06 |
| marex | which is better | 17:06 |
| ezequielg | clearly something is different with 8mm g1. | 17:06 |
| marex | ezequielg: so it seems | 17:08 |
| marex | ezequielg: or not ... I got it working on 8mm now too | 17:12 |
| marex | ezequielg: just a missing control register write as a result of too much digging | 17:12 |
| ezequielg | ah! that's the expected outcome. these hantro cores are expected to be integrated pretty much the same way. | 17:13 |
| ezequielg | this means no regressions in linux-next? if you report any, could you Cc me? | 17:13 |
| marex | ezequielg: once the gpcv2 mess settles, I will start sending out whatever is possibly left for 8mm/mn, so yeah | 17:14 |
| + natalie (~natalie@user/natalie) | 17:15 | |
| marex | ezequielg: so the only difference in the hantro stuff on 8mm is apparently even more shenanigans with gpcv2 ... | 17:19 |
| + rasmus-mk (~rasmus@c83-253-223-217.bredband.tele2.se) | 18:10 | |
| - Neelfyn (QUIT: Quit: Connection closed for inactivity) (uid180106@id-180106.stonehaven.irccloud.com) | 18:25 | |
| + specing (~specing@user/specing) | 18:27 | |
| - specing_ (QUIT: Ping timeout: 272 seconds) (~specing@APN-123-240-240-gprs.simobil.net) | 18:29 | |
| mntmn | marex: did you have to use a custom build of ffmpeg? | 18:32 |
| mntmn | or can it work in ffmpeg out of the box now? | 18:33 |
| mntmn | cc ezequielg | 18:33 |
| ezequielg | mntmn: only gstreamer for now. | 18:33 |
| ezequielg | ffmpeg hasn't merged any of the patches, afaik. | 18:33 |
| ezequielg | libreleec people hang out on a slack channel, if you manage to find your way there, you can ask them about the status. | 18:34 |
| ezequielg | they're all pretty friendly. | 18:34 |
| mntmn | i see, thanks | 18:34 |
| mntmn | btw do i remember correctly that WPE uses gstreamer or something? | 18:35 |
| - GBenji (PART: !!unknown attribute: msg!!) (~benjamin@88.123.46.161) | 18:44 | |
| - mjw (QUIT: Quit: Leaving) (~mjw_@2001:1c06:2488:200:9e5c:8eff:fe8f:a440) | 19:07 | |
| - sknebel (QUIT: Quit: sknebel) (~quassel@v22016013254630973.happysrv.de) | 19:15 | |
| + sknebel (~quassel@v22016013254630973.happysrv.de) | 19:15 | |
| marex | mntmn: I used x86 computer for the conversion | 19:19 |
| mntmn | marex: sorry, misunderstood | 19:19 |
| marex | mntmn: no worries | 19:19 |
| marex | mntmn: also, you'd need the h1 for encoding | 19:21 |
| mntmn | ok | 19:21 |
| marex | mntmn: which is not supported on 8mm and not present on 8mq ;-) | 19:22 |
| mntmn | marex: coincidentally my favorite bar is also called 8mm (i don't drink anymore though) | 19:22 |
| marex | mntmn: 8mn might convince you otherwise, and if not 8mn, then 8mp | 19:23 |
| mntmn | haha | 19:23 |
| mntmn | i'm ignoring those until imx9 | 19:24 |
| marex | mntmn: oh yes, the imx9 ... | 19:24 |
| mntmn | next week is NXP tech week or something... maybe they will tell us about the cores? | 19:24 |
| marex | mntmn: ooooooh ... I wouldn't hold my breath | 19:25 |
| mntmn | i don't really | 19:25 |
| marex | mntmn: cortex A55 finally ? | 19:27 |
| mntmn | i guess so | 19:27 |
| marex | https://www.renesas.com/us/en/products/microcontrollers-microprocessors/rz-arm-based-high-end-32-64-bit-mpus/rzg2l-general-purpose-microprocessors-dual-core-arm-cortex-a55-12-ghz-cpus-3d-graphics-and-video-codec like this ? | 19:27 |
| mntmn | what's the biglittle partner of a55 again? a73? | 19:27 |
| mntmn | naja i ~am~ looking forward to 12nm process | 19:28 |
| mntmn | also they want to do some 5nm thing | 19:28 |
| mntmn | lets hope it's not just a mcu | 19:28 |
| marex | mntmn: the mx9 ? no ... it's likely gonna be similar to the g2l though | 19:28 |
| marex | (except for the software, more BSP-only downstream goo) | 19:29 |
| mntmn | hmm i hadn't really considered renesas for a longer while... how are their docs? | 19:30 |
| marex | mntmn: the g2l seems to be some new development with non-powervr GPU, that one looks nice | 19:31 |
| mntmn | yeah, mali | 19:31 |
| mntmn | cool indeed | 19:32 |
| marex | mntmn: the rzg docs were public I think, and well, upstream support is nice | 19:32 |
| mntmn | oh ok | 19:32 |
| marex | mntmn: the g2l is new , the datasheet isnt public yet I think | 19:32 |
| mntmn | no pcie though | 19:34 |
| marex | mntmn: some g2h would be nice, since 2xa55 isnt much | 19:34 |
| mntmn | yeah | 19:34 |
| mntmn | all kind of reforms in the wild https://twitter.com/SawaizSyed/status/1400512502589542409 | 20:02 |
| swivel | mntmn: cool! you must be proud to have produced something worthy of such exhibition by a third party, I would be | 20:13 |
| mntmn | genode sculpt OS on reform, incl. image! https://genodians.org/skalk/2021-06-03-mnt-reform | 21:05 |
| mntmn | cc bluerise, this could be interesting for you as it enables the internal display | 21:35 |
| + Neelfyn (uid180106@id-180106.stonehaven.irccloud.com) | 21:35 | |
| - freakazoid333 (QUIT: Read error: Connection reset by peer) (~freakazoi@2001:5b0:2a65:fa88:42f:154a:842c:dca2) | 21:54 | |
| + freakazoid333 (~freakazoi@2001:5b0:2a65:fa88:98db:2c2d:b456:5c52) | 21:57 | |
| + ephase (~ephase@2a01:e0a:168:1211:f3e4:d6ab:6cbf:f594) | 22:01 | |
| bluerise | hmhm? | 22:02 |
| ephase | hellp | 22:02 |
| bluerise | mntmn: the genode stuff? | 22:02 |
| mntmn | bluerise: yep. i just checked the code and some of it is wrapping linux drivers though i think. | 22:04 |
| mntmn | ephase: what's up? | 22:04 |
| ephase | hey mntmn allright and you? | 22:04 |
| mntmn | good good... working on the FPGA module for reform at the moment | 22:05 |
| + mtm (~mtm@c-174-58-99-93.hsd1.fl.comcast.net) | 22:06 | |
| + mjw (~mark@herd.wildebeest.org) | 22:15 | |
| - S0rin (QUIT: Ping timeout: 264 seconds) (~S0rin@82-64-116-66.subs.proxad.net) | 22:21 | |
| + S0rin (~S0rin@82-64-116-66.subs.proxad.net) | 22:33 | |
| specing | cool | 22:34 |
| - rasmus-mk (QUIT: Quit: WeeChat 3.0.1) (~rasmus@c83-253-223-217.bredband.tele2.se) | 22:38 | |
| + odnes (~odnes@109-178-243-26.pat.ren.cosmote.net) | 22:39 | |
| ephase | mntmn: sounds greats! I receive back my motherboard last week and all work great thank you. Need to continue working on my image creator | 22:41 |
| mntmn | ephase: ahh, good to hear | 22:42 |
| mjw | I often don't have a sound device, is that a software or hardware issue? | 22:43 |
| mntmn | mjw: could be either or. i _think_ it appeared at a certain kernel rev. maybe just some i2c timing or reset issue with the wm8960 | 22:49 |
| ephase | mntmn: have you tried to compile kernel 5.12 for the reform? | 22:49 |
| mntmn | ephase: maybe, i can't remember atm | 22:49 |
| ephase | mntmn: here is my mkuserland script and dockerfile : https://git.epha.se/ephase/reform3_mkuserland, all based on mmdebstrap and its hook system | 23:01 |
| mjw | so is there a known "good" kernel? | 23:02 |
| mntmn | mjw: the one in the reform-system-image is the reference | 23:02 |
| mjw | I appear to be running 5.11.0-rc7+ | 23:03 |
| mntmn | mjw: https://source.mnt.re/reform/reform-system-image/-/blob/main/reform2-imx8mq/mkkernel.sh#L18 | 23:03 |
| mntmn | my guess is 5.12 will work fine, but you have to include the same patches | 23:03 |
| mntmn | ephase: cool, interesting | 23:05 |
| mjw | ACTION will work in silence for a bit | 23:06 |
| ephase | thanks mjw | 23:08 |
| ephase | thanks mntmn | 23:08 |
| ephase | mjw:I mean mntmn | 23:08 |
| mjw | :) /me happily gets thanks by osmosis | 23:09 |
| mntmn | ;3 | 23:18 |
| - odnes (QUIT: Ping timeout: 272 seconds) (~odnes@109-178-243-26.pat.ren.cosmote.net) | 23:25 | |
| + erle (~erle@dynamic-046-114-034-096.46.114.pool.telefonica.de) | 23:42 | |
| - erlehmann (QUIT: Killed (NickServ (GHOST command used by erle!~erle@dynamic-046-114-034-096.46.114.pool.telefonica.de))) (~erle@dynamic-046-114-032-022.46.114.pool.telefonica.de) | 23:42 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!