|
Description
atirage21
2018-06-24 19:18:15 UTC
To debug this, I'd like you to attach to this bug report: - A better description of what you're trying to do. I suspect you're trying to set a pstate, but please be clear about the steps that you follow to get the result you claim to get. - A copy of your video BIOS (/sys/kernel/debug/dri/<number>/vbios.rom - A peek of the strap register (nvapeek 0x101000) - A full unfiltered dmesg from kernel 4.17 after failure like you did for the 3.19 kernel, preferably booted with the nouveau.debug="PMU=debug" parameter. (In reply to atirage21 from comment #0) > Maybe it is connected on your changes from 2016: > https://cgit.freedesktop.org/nouveau/xf86-video-nouveau/commit/ > ?id=a24ded627e42d453546c553d609edc073f59cd57 This patch... no, this whole repository is unrelated to your problems. Created attachment 140305 [details]
vBIOS.rom (0x0e73b0a2, GK107, NVE7, Family NVE0, version 80.07.40.00.02
can you print the output of "nvapeek 0x20340 0x8" please? (In reply to Karol Herbst from comment #3) > can you print the output of "nvapeek 0x20340 0x8" please? Please send me exatly guide how, because command "nvapeek 0x20340 0x8" not found. Created attachment 140306 [details]
dmesg message from kernel 4.15 (as case study from kernel 4.x) live ubuntu 18.04
It was generated through live ubuntu 18.04 (kernel 4.15)
Created attachment 140307 [details]
dmesg message from kernel 4.4 (on same OS Linux Mint 17.3 like with good kernel 3.19)
(In reply to atirage21 from comment #4) > (In reply to Karol Herbst from comment #3) > > can you print the output of "nvapeek 0x20340 0x8" please? > > Please send me exatly guide how, because command "nvapeek 0x20340 0x8" not > found. Output from mmiotrace: R 4 127.904981 11 0xd8309624 0x203406e 0x0 0 Created attachment 140311 [details]
MMIOTRACEfrom nvidia driver 340.102
(In reply to atirage21 from comment #4) > (In reply to Karol Herbst from comment #3) > > can you print the output of "nvapeek 0x20340 0x8" please? > > Please send me exatly guide how, because command "nvapeek 0x20340 0x8" not > found. you would need to install envytools, but the trace is actually more helpful, because there we could simply lookup how nvidia is doing stuff @atirage21: were you actually able to reclock the GPU before in a stable way? It seems like we have to support a way of setting the voltage we don't implement yet (and never had). It is the first time actually seeing such a GPU, so I am not quite sure how we are able to get that working for now. Of course we could provide patches and hope it kind of works, but this would require testing from your side as well. Created attachment 140318 [details] attachment-26224-0.html I was not successfull in reclock the gpu. OK 2018-06-25 14:03 GMT+02:00 <bugzilla-daemon@freedesktop.org>: > *Comment # 10 <https://bugs.freedesktop.org/show_bug.cgi?id=107016#c10> on > bug 107016 <https://bugs.freedesktop.org/show_bug.cgi?id=107016> from Karol > Herbst <karolherbst@gmail.com> * > > @atirage21: > > were you actually able to reclock the GPU before in a stable way? It seems like > we have to support a way of setting the voltage we don't implement yet (and > never had). > > It is the first time actually seeing such a GPU, so I am not quite sure how we > are able to get that working for now. Of course we could provide patches and > hope it kind of works, but this would require testing from your side as well. > > ------------------------------ > You are receiving this mail because: > > - You reported the bug. > > Created attachment 140326 [details] attachment-20771-0.html But also is question: Why i had only 405 Mhz of core under kernel 4.x, if under kernel 3.19 was more options to see (and same more options is with propretiary driver): [ 4.311074] nouveau [ CLK][0000:01:00.0] 07: core 270-405 MHz memory 810 MHz [ 4.311090] nouveau [ CLK][0000:01:00.0] 0f: core 270-745 MHz memory 1800 MHz [ 4.311233] nouveau [ CLK][0000:01:00.0] --: core 405 MHz memory 648 MHz 2018-06-25 14:03 GMT+02:00 <bugzilla-daemon@freedesktop.org>: > *Comment # 10 <https://bugs.freedesktop.org/show_bug.cgi?id=107016#c10> on > bug 107016 <https://bugs.freedesktop.org/show_bug.cgi?id=107016> from Karol > Herbst <karolherbst@gmail.com> * > > @atirage21: > > were you actually able to reclock the GPU before in a stable way? It seems like > we have to support a way of setting the voltage we don't implement yet (and > never had). > > It is the first time actually seeing such a GPU, so I am not quite sure how we > are able to get that working for now. Of course we could provide patches and > hope it kind of works, but this would require testing from your side as well. > > ------------------------------ > You are receiving this mail because: > > - You reported the bug. > > (In reply to atirage21 from comment #12) > Created attachment 140326 [details] > attachment-20771-0.html > > But also is question: Why i had only 405 Mhz of core under kernel 4.x, > if under kernel 3.19 was more options to see (and same more options is > with propretiary driver): > > [ 4.311074] nouveau [ CLK][0000:01:00.0] 07: core 270-405 MHz > memory 810 MHz > [ 4.311090] nouveau [ CLK][0000:01:00.0] 0f: core 270-745 MHz > memory 1800 MHz > [ 4.311233] nouveau [ CLK][0000:01:00.0] --: core 405 MHz memory 648 > MHz > That's just reduced verbosity of the driver in newer kernels. There's no need to report that during boot, just clutters the log. You'll find all those entries in /sys/kernel/debug/dri/<number>/pstate. Created attachment 140327 [details] attachment-28061-0.html Not truth! I see only 405 MHz of core in newer kernel 4.x in /sys/kernel/debug/dri/<number>/pstate. It was first argument for me, that something is not good compared to kernel 3.19. 2018-06-25 16:25 GMT+02:00 <bugzilla-daemon@freedesktop.org>: > *Comment # 13 <https://bugs.freedesktop.org/show_bug.cgi?id=107016#c13> on > bug 107016 <https://bugs.freedesktop.org/show_bug.cgi?id=107016> from Roy > <nouveau@spliet.org> * > > (In reply to atirage21 from comment #12 <https://bugs.freedesktop.org/show_bug.cgi?id=107016#c12>)> Created attachment 140326 [details] <https://bugs.freedesktop.org/attachment.cgi?id=140326> [details] <https://bugs.freedesktop.org/attachment.cgi?id=140326&action=edit> > > attachment-20771-0.html > > > > But also is question: Why i had only 405 Mhz of core under kernel 4.x, > > if under kernel 3.19 was more options to see (and same more options is > > with propretiary driver): > > > > [ 4.311074] nouveau [ CLK][0000:01:00.0] 07: core 270-405 MHz > > memory 810 MHz > > [ 4.311090] nouveau [ CLK][0000:01:00.0] 0f: core 270-745 MHz > > memory 1800 MHz > > [ 4.311233] nouveau [ CLK][0000:01:00.0] --: core 405 MHz memory 648 > > MHz > > > > That's just reduced verbosity of the driver in newer kernels. There's no need > to report that during boot, just clutters the log. You'll find all those > entries in /sys/kernel/debug/dri/<number>/pstate. > > ------------------------------ > You are receiving this mail because: > > - You reported the bug. > > Created attachment 140332 [details] attachment-14056-0.html In live ubuntu 18.04 with kernel 4.15 Only boot frequency of core sudo cat /sys/kernel/debug/dri/0/pstate 07: core 405 MHz memory 810 MHz 0f: core 405 MHz memory 1800 MHz AC: core 405 MHz memory 648 MHz 2018-06-25 14:51 GMT+00:00 Attila Tóth <tothsoft@gmail.com>: > Not truth! I see only 405 MHz of core in newer kernel 4.x in > > /sys/kernel/debug/dri/<number>/pstate. It was first argument for me, that something is not good compared to kernel 3.19. > > > > 2018-06-25 16:25 GMT+02:00 <bugzilla-daemon@freedesktop.org>: > >> *Comment # 13 <https://bugs.freedesktop.org/show_bug.cgi?id=107016#c13> >> on bug 107016 <https://bugs.freedesktop.org/show_bug.cgi?id=107016> from >> Roy <nouveau@spliet.org> * >> >> (In reply to atirage21 from comment #12 <https://bugs.freedesktop.org/show_bug.cgi?id=107016#c12>)> Created attachment 140326 [details] <https://bugs.freedesktop.org/attachment.cgi?id=140326> [details] <https://bugs.freedesktop.org/attachment.cgi?id=140326&action=edit> >> > attachment-20771-0.html >> > >> > But also is question: Why i had only 405 Mhz of core under kernel 4.x, >> > if under kernel 3.19 was more options to see (and same more options is >> > with propretiary driver): >> > >> > [ 4.311074] nouveau [ CLK][0000:01:00.0] 07: core 270-405 MHz >> > memory 810 MHz >> > [ 4.311090] nouveau [ CLK][0000:01:00.0] 0f: core 270-745 MHz >> > memory 1800 MHz >> > [ 4.311233] nouveau [ CLK][0000:01:00.0] --: core 405 MHz memory 648 >> > MHz >> > >> >> That's just reduced verbosity of the driver in newer kernels. There's no need >> to report that during boot, just clutters the log. You'll find all those >> entries in /sys/kernel/debug/dri/<number>/pstate. >> >> ------------------------------ >> You are receiving this mail because: >> >> - You reported the bug. >> >> > > > -- > S pozdravom / Mit freundlichen Grüßen / Baráti udvozlettel / Best regards > > Ing. Attila Tóth, PhD. > > tel : +421 910 299 189 (4Ka - Slovak Republic) > > Please consider your environmental responsibility before printing this > e-mail > > > > > (In reply to atirage21 from comment #15) > Created attachment 140332 [details] > attachment-14056-0.html > > In live ubuntu 18.04 with kernel 4.15 > > Only boot frequency of core > sudo cat /sys/kernel/debug/dri/0/pstate > > > > 07: core 405 MHz memory 810 MHz > 0f: core 405 MHz memory 1800 MHz > AC: core 405 MHz memory 648 MHz > > > 2018-06-25 14:51 GMT+00:00 Attila Tóth <tothsoft@gmail.com>: > > > Not truth! I see only 405 MHz of core in newer kernel 4.x in > > > > /sys/kernel/debug/dri/<number>/pstate. It was first argument for me, that something is not good compared to kernel 3.19. > > > > > > > > 2018-06-25 16:25 GMT+02:00 <bugzilla-daemon@freedesktop.org>: > > > >> *Comment # 13 <https://bugs.freedesktop.org/show_bug.cgi?id=107016#c13> > >> on bug 107016 <https://bugs.freedesktop.org/show_bug.cgi?id=107016> from > >> Roy <nouveau@spliet.org> * > >> > >> (In reply to atirage21 from comment #12 <https://bugs.freedesktop.org/show_bug.cgi?id=107016#c12>)> Created attachment 140326 [details] <https://bugs.freedesktop.org/attachment.cgi?id=140326> [details] <https://bugs.freedesktop.org/attachment.cgi?id=140326&action=edit> > >> > attachment-20771-0.html > >> > > >> > But also is question: Why i had only 405 Mhz of core under kernel 4.x, > >> > if under kernel 3.19 was more options to see (and same more options is > >> > with propretiary driver): > >> > > >> > [ 4.311074] nouveau [ CLK][0000:01:00.0] 07: core 270-405 MHz > >> > memory 810 MHz > >> > [ 4.311090] nouveau [ CLK][0000:01:00.0] 0f: core 270-745 MHz > >> > memory 1800 MHz > >> > [ 4.311233] nouveau [ CLK][0000:01:00.0] --: core 405 MHz memory 648 > >> > MHz > >> > > >> > >> That's just reduced verbosity of the driver in newer kernels. There's no need > >> to report that during boot, just clutters the log. You'll find all those > >> entries in /sys/kernel/debug/dri/<number>/pstate. > >> Nvidia xserver setting have clock from 135-405 and from 135 - 745 Mhz of clock (see attachment) Created attachment 140333 [details]
Nvidia xserver clocks -printscreen
(In reply to atirage21 from comment #4) > (In reply to Karol Herbst from comment #3) > > can you print the output of "nvapeek 0x20340 0x8" please? > > Please send me exatly guide how, because command "nvapeek 0x20340 0x8" not > found. nvapeek 0x20340 0x8 00020340: 00000360 00000000 (In reply to Roy from comment #1) > To debug this, I'd like you to attach to this bug report: > - A better description of what you're trying to do. I suspect you're trying > to set a pstate, but please be clear about the steps that you follow to get > the result you claim to get. > - A copy of your video BIOS (/sys/kernel/debug/dri/<number>/vbios.rom > - A peek of the strap register (nvapeek 0x101000) > - A full unfiltered dmesg from kernel 4.17 after failure like you did for > the 3.19 kernel, preferably booted with the nouveau.debug="PMU=debug" > parameter. > > (In reply to atirage21 from comment #0) > > Maybe it is connected on your changes from 2016: > > https://cgit.freedesktop.org/nouveau/xf86-video-nouveau/commit/ > > ?id=a24ded627e42d453546c553d609edc073f59cd57 > > This patch... no, this whole repository is unrelated to your problems. nvapeek 0x101000 00101000: df40ac9a Created attachment 140413 [details]
Dmesg mesage Kernel 4.15 with nvidia propretiary driver
Linux mint 19 mate 64bit
Karol Herbst: Hi, do you still have your GK107 Nvidia GPU? I just recently found out, that there is a strange thing to control the GPU voltage there which I never saw on any other board and I would like to know if you are up for doing some testing. Greetings Karol Attila: Hi, yes, i am. So , i tried todya oibaf witk last mesa 18.3 devel, but is same. 4.496952] nouveau 0000:01:00.0: bios: version 80.07.40.00.02 [ 4.497677] nouveau 0000:01:00.0: mxm: BIOS version 3.0 [ 4.501493] nouveau 0000:01:00.0: volt: Type mismatch between the voltage table type and the GPIO table. Fallback to GPIO mode. [ 4.502201] nouveau 0000:01:00.0: fb: 2048 MiB DDR3 Nothing new ? Created attachment 142724 [details]
dmesg message from kernel 4.15 from december 2018 (with last mesa and nouveau driver from OIBAF)
It is possible solve this issue with gk107 on base new publicated documentation from nvidia for gk 104 ? Also i saw error in envytools in row for gk107: NVE7 [GK107]: ?:?:2:1 https://github.com/mharsch/envytools/commit/b70216fcf3d268995d7f9aa275db3a359c6a80bd Created attachment 145371 [details]
Valgrind-mmt log from glxgears with driver nvidia 390.116ubuntu0.18.04.1
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/xorg/driver/xf86-video-nouveau/issues/441. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.