#freedesktop on 2024-10-01 — irc logs at oftc.irclog.whitequark.org

2024-07-16 04:52 ChanServ changed the topic of #freedesktop to: https://www.freedesktop.org infrastructure and online services || for questions about freedesktop.org projects, please see each project's contact || for discussions about specifications, please use https://gitlab.freedesktop.org/xdg or xdg@lists.freedesktop.org

00:01 nephyrin has joined #freedesktop

01:34 scrumplex_ has joined #freedesktop

01:36 karenw has quit [Ping timeout: 480 seconds]

01:37 alpernebbi has quit [Ping timeout: 480 seconds]

01:37 scrumplex has quit [Ping timeout: 480 seconds]

01:39 digetx has joined #freedesktop

01:45 digetx is now known as Guest5068

01:46 digetx has joined #freedesktop

01:48 alpernebbi has joined #freedesktop

01:48 Guest5068 has quit [Ping timeout: 480 seconds]

01:54 digetx has quit [Ping timeout: 480 seconds]

02:12 digetx has joined #freedesktop

02:22 dcunit3d has joined #freedesktop

02:34 digetx is now known as Guest5071

02:34 digetx has joined #freedesktop

02:37 Guest5071 has quit [Ping timeout: 480 seconds]

03:04 ximion has quit [Remote host closed the connection]

03:53 lsd|2 has joined #freedesktop

04:29 eluks has quit [Remote host closed the connection]

04:32 eluks has joined #freedesktop

05:07 swatish2 has joined #freedesktop

05:27 bmodem has joined #freedesktop

05:28 swatish2 has quit [Ping timeout: 480 seconds]

05:39 swatish2 has joined #freedesktop

05:56 sima has joined #freedesktop

06:04 jsa has joined #freedesktop

06:21 tzimmermann has joined #freedesktop

06:28 jsa1 has joined #freedesktop

06:33 jsa has quit [Ping timeout: 480 seconds]

06:34 lsd|2 has quit [Quit: KVIrc 5.2.2 Quasar http://www.kvirc.net/]

06:36 i-garrison has quit []

06:39 i-garrison has joined #freedesktop

06:48 AbleBacon has quit [Read error: Connection reset by peer]

07:02 jsa1 has quit [Ping timeout: 480 seconds]

07:10 smpl has joined #freedesktop

07:10 Haaninjo has joined #freedesktop

07:14 swatish21 has joined #freedesktop

07:14 jsa has joined #freedesktop

07:15 <MrCooper> DavidHeidelberg: what difference does F40/41 make?

07:17 <DavidHeidelberg> newer compiler?

07:17 <MrCooper> what's the problem with the old one?

07:20 <DavidHeidelberg> Nothing, just the color is little bit off. (Ref to Snatch movie).

07:20 swatish2 has quit [Ping timeout: 480 seconds]

07:20 <DavidHeidelberg> Usually the more recent are better at catching the bugs.. (and produce more bogus warnings....)

07:21 jsa1 has joined #freedesktop

07:23 jsa has quit [Ping timeout: 480 seconds]

07:29 <MrCooper> I'm still none the wiser as to why LTO should be disabled now and enabled again with F40/41

07:35 <DavidHeidelberg> Because it's pre-merge and won't likely catch anything above non-LTO. What I want is MORE lto testing, but at night, when we have the time and resources.

07:36 <DavidHeidelberg> Also what is point testing LTO build specifically on F38?

07:36 <DavidHeidelberg> If it was last rhel release or something, it would make more sense in my eyes

07:39 <DavidHeidelberg> But.. Then, even with RHEL, we cannot guarantee it'll not break runtime due to some unknown circumstance, so why?

07:41 <DavidHeidelberg> I don't think LTO is currently reasonable for pre-merge. For release, yes. For nightly, why not.

07:44 <DavidHeidelberg> I wanted LTO and enabled it for debian-build-testing before, but I don't have feeling it gave some extra value except longer link time. Now when we can do it in nightly, it feels perfect.

07:52 ximion has joined #freedesktop

07:53 <bentiss> heh, fwiw, m3l-31 has one job running, and load average is 101.22 over 5 minutes, and guess what: /usr/libexec/gcc/x86_64-redhat-linux/13/lto1 -quiet -dumpbase src/gallium/targets...

07:54 <mupuf> bentiss: wow

07:54 <bentiss> TBH, I don't complain, because now the job is DoS itself :)

07:55 <bentiss> again, was https://gitlab.freedesktop.org/mesa/mesa/-/jobs/64437031

07:56 <bentiss> so to me, it all comes down to: is 18 minutes OK for this job to run, or should you try to disable LTO to speed it up?

07:56 tpalli has joined #freedesktop

07:57 lynxeye has joined #freedesktop

07:57 <mupuf> I would say OK for nightly runs... IF IT DIDN'T EXCEED ITS ALLOCATED CPU COUNT!

07:58 <mupuf> if the thing takes 19 minutes AND slows down all the other jobs running on the runner... no way jose!

07:58 <tpalli> I'm having some trouble with mesa ci, a660_gl 1/4 job failing which is quite unexpected since I'm not touching anything a660 would be using

07:58 <bentiss> mupuf: but the container only sees 8 CPUs, and the load is restricted to these 8 CPUs, so it doesn't DoS the system at all

07:59 <mupuf> tpalli: valentine just said she had merged an MR for it

07:59 <mupuf> https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31456

07:59 <tpalli> mupuf ok cool, will try again then

07:59 <mupuf> tpalli: crossing fingers for you :)

08:00 <mupuf> bentiss: hmm, then yeah, I guess it is ok for nightly

08:00 <tpalli> mupuf thanks! :)

08:02 <MrCooper> DavidHeidelberg: as I wrote on gitlab, the point is to ensure that Mesa can build with LTO

08:03 <DavidHeidelberg> One a day ensurance is not enough?

08:04 <MrCooper> the same argument applies to anything, why bother with pre-merge CI?

08:05 <MrCooper> anyway, I'm not blocking this, just saying the rationale seems somewhat arbitrary

08:05 <bentiss> do you have any tests when you reduce LTO so it's less agressive, in terms of job total time?

08:07 <bentiss> because if you reduce the number of parallel LTOs and the job takes the same time or is longer, then we can make a better decision

08:07 <bentiss> and AFAICT, as long as all 24 slots are not used with those LTO jobs, we are fine

08:12 <mupuf> bentiss: there are other things running in parallel, so I would vote against pre-merge

08:47 pixelcluster_ has quit []

08:47 pixelcluster has joined #freedesktop

08:53 Haaninjo has quit [Quit: Ex-Chat]

09:02 psykose has joined #freedesktop

09:03 psykose has quit [Remote host closed the connection]

09:07 <bilboed> regarding the web server, might be worth setting up mod_evasive ? https://medium.com/@iyaadluq/how-to-setup-rate-limiting-on-apache-server-using-mod-evasive-2961ddeb466a

09:14 ximion has quit [Remote host closed the connection]

09:15 swatish2 has joined #freedesktop

09:20 swatish21 has quit [Ping timeout: 480 seconds]

09:20 jsa1 has quit [Ping timeout: 480 seconds]

09:59 gchini has quit [Remote host closed the connection]

09:59 gchini has joined #freedesktop

10:00 Shibe is now known as ShibaInu

10:00 ShibaInu is now known as Shibe

10:11 <valentine> mupuf: my pronouns are he/him btw :P

10:12 <mupuf> valentine: oh, sorry about this! In France, Valentine is generally given to women, I should have checked!

10:13 <valentine> no worries, it's both a masculine and a feminine given name

10:14 <mupuf> I see :) Where are you from?

10:15 <valentine> I'm european, and it's the anglicized equivalent of my name

10:16 gchini has quit [Quit: Leaving]

10:16 gchini has joined #freedesktop

10:16 <mupuf> valentine: ack, thanks and sorry for the misgendering

10:17 <valentine> no worries at all

10:26 sentriz has quit [Ping timeout: 480 seconds]

10:42 sentriz has joined #freedesktop

10:58 swatish2 has quit [Ping timeout: 480 seconds]

11:09 jsa has joined #freedesktop

11:16 swatish2 has joined #freedesktop

11:35 swatish2 has quit [Ping timeout: 480 seconds]

11:49 guludo has joined #freedesktop

12:36 swatish2 has joined #freedesktop

12:40 jsa has quit [Ping timeout: 480 seconds]

12:51 jsa has joined #freedesktop

12:52 codegirl has joined #freedesktop

12:55 swatish2 has quit [Ping timeout: 480 seconds]

12:57 lsd|2 has joined #freedesktop

13:12 ___nick___ has joined #freedesktop

13:36 bmodem has quit [Ping timeout: 480 seconds]

13:42 AbleBacon has joined #freedesktop

14:17 JanC is now known as Guest5103

14:17 JanC has joined #freedesktop

14:21 Guest5103 has quit [Ping timeout: 480 seconds]

14:28 jsa has quit [Ping timeout: 480 seconds]

14:29 JanC is now known as Guest5104

14:29 JanC has joined #freedesktop

14:32 jsa has joined #freedesktop

14:33 Guest5104 has quit [Ping timeout: 480 seconds]

14:46 marc2377 has joined #freedesktop

14:47 swatish2 has joined #freedesktop

14:48 <marc2377> Hey guys. Which is faster / more modern? fbdev or vesa? (My hardware doesn't seem to support uvesa because it's VBE version 1)

14:48 <marc2377> I'm compiling and about to test drm-openchrome (and its DDX driver), but want a fallback to my Via Unichrome Pro chip

14:48 JanC is now known as Guest5108

14:48 JanC has joined #freedesktop

14:53 Guest5108 has quit [Ping timeout: 480 seconds]

15:15 ___nick___ has quit []

15:17 swatish2 has quit [Ping timeout: 480 seconds]

15:17 ___nick___ has joined #freedesktop

15:18 ___nick___ has quit []

15:20 ___nick___ has joined #freedesktop

15:49 krushia has joined #freedesktop

15:54 swatish2 has joined #freedesktop

16:03 swatish2 has quit [Ping timeout: 480 seconds]

16:46 The_Loko has joined #freedesktop

16:52 <eric_engestrom> bentiss: I had the same idea as daniels (the "placeholder job that checks for the container and creates a dynamic pipeline that contains the real container job if needed") but when I talked to mupuf about it, he pointed out each the build job would then have to wait on the whole child pipeline instead of being able to wait on only the relevant container job :(

16:56 ___nick___ has quit [Ping timeout: 480 seconds]

16:56 <eric_engestrom> perhaps the trade-off is worth it though

16:58 <daniels> I think it's worth it, given the frequency of needing a container build vs. not

16:58 <eric_engestrom> all the pipelines that _don't_ touch the containers will be faster, and the pipelines that _do_ touch the containers are slow anyway (and should not be allowed in merge pipelines, but that's a different issue), so making them slower shouldn't be too much of an issue

16:58 <daniels> ++

16:59 tzimmermann has quit [Quit: Leaving]

17:02 sentriz has quit [Ping timeout: 480 seconds]

17:05 sentriz has joined #freedesktop

17:05 lynxeye has quit [Quit: Leaving.]

17:06 ___nick___ has joined #freedesktop

17:06 * bentiss eagerly waits for teh ci-templates patches :)

17:07 <eric_engestrom> regular ci I could give it a shot, but ci-templates I always get lost

17:07 <bentiss> heh

17:12 <mupuf> I wonder how this could become a ci-template feature... but definitely +1 for making the dynamic pipeline IF we can make it maintainable

17:15 cisco87 has joined #freedesktop

17:29 jsa has quit [Ping timeout: 480 seconds]

18:09 scrumplex has joined #freedesktop

18:15 scrumplex_ has quit [Ping timeout: 480 seconds]

18:28 eluks has quit [Remote host closed the connection]

18:32 ximion has joined #freedesktop

18:35 eluks has joined #freedesktop

18:35 eluks has quit [Remote host closed the connection]

18:35 eluks has joined #freedesktop

18:39 swatish2 has joined #freedesktop

18:50 Haaninjo has joined #freedesktop

19:36 vx has quit [Quit: G-Line: User has been permanently banned from this network.]

19:37 vx has joined #freedesktop

20:05 ___nick___ has quit [Remote host closed the connection]

20:15 proc1 has quit [Ping timeout: 480 seconds]

20:19 swatish2 has quit [Ping timeout: 480 seconds]

21:04 The_Loko has quit [Quit: Leaving]

21:06 sima has quit [Ping timeout: 480 seconds]

21:35 alanc has quit [Remote host closed the connection]

21:35 alanc has joined #freedesktop

22:01 smpl has quit [Ping timeout: 480 seconds]

22:05 Haaninjo has quit [Quit: Ex-Chat]

22:37 guludo has quit [Quit: WeeChat 4.4.2]

22:41 vx has quit [Quit: G-Line: User has been permanently banned from this network.]

22:41 vx has joined #freedesktop

23:40 marc2377-2 has joined #freedesktop

23:46 marc2377 has quit [Ping timeout: 480 seconds]