Created attachment 343423 [details] log #1 http://cgit.freedesktop.org/systemd/systemd/patch/?id=8cc3f8c0bcd23bb68166cb197a4c541d7621b19c NEWS: * The udev daemon now adjusts its default number of parallel executed worker processes based on the number of available CPUs instead of the amount of available RAM. This is supposed to provide a more reliable default and limit a too agressive paralellism for setups with 1000s of devices connected. User at Freenode, rei4dan reported this causing ~10 minute uevent startup time with Intel i7 Quad Core HT on DELL XPS L502X laptop. Attaching his logs here.
Created attachment 343425 [details] log #2 in .xz format due to filesize
Created attachment 343427 [details, diff] Patch reverting the upstream commit
Comment on attachment 343427 [details, diff] Patch reverting the upstream commit No, it wasn't this. Testing reverting http://cgit.freedesktop.org/systemd/systemd/commit/?id=f4fbb8be14051c0ea920cebcb011c965d95a7b86 now
(In reply to comment #3) > Comment on attachment 343427 [details, diff] [details, diff] > Patch reverting the upstream commit > > No, it wasn't this. > > Testing reverting > http://cgit.freedesktop.org/systemd/systemd/commit/ > ?id=f4fbb8be14051c0ea920cebcb011c965d95a7b86 > now Not this either. Closing as NEEDINFO for now.
I think I hit this, too. With udev-199 my boot takes one additional minute. I created a bootchart, but don't think that helps much for debugging. For the moment I reverted to udev-198-r6.
(In reply to comment #5) > I think I hit this, too. With udev-199 my boot takes one additional minute. > I created a bootchart, but don't think that helps much for debugging. For > the moment I reverted to udev-198-r6. well we'd need someone with the problem to try bisecting git commits between 198 and 199 and see what caused this remember you can enable extra logging through /etc/udev/udev.conf, and /etc/conf.d/udev, just as long as CONFIG_LOG_BUF_SHIFT= value is high enough, like 20 or such in kernel so it won't just get wiped away /etc/conf.d/bootmisc also has log_dmesg= and /etc/rc.conf even more logging options to enable http://cgit.freedesktop.org/systemd/systemd/ also it could be you two are having different issues, so the patch from Comment #2 might be your issue but not the persons I've filed this bug for: 07:57 < ssuominen> rei4dan: 16:21 < rei4dan> ssuominen, epatch -R "${FILESDIR}/f4fbb8be14051c0ea920cebcb011c965d95a7b86.patch" didn't help 07:57 < ssuominen> rei4dan: but yet you say, 18:37 < rei4dan> ssuominen, hmm.. f4fbb8be14051c0ea920cebcb011c965d95a7b86 is the first bad commit 07:57 < ssuominen> rei4dan: those 2 contradict each other. i'm not sure what to trust anymore. that's how our last conversation ended. hope you can understand we are a bit stuck here... :-/
I believe this bug is a messy duplicate of bug 463604 and is now fixed with =sys-fs/199-r1, as well as in =sys-fs/udev-9999 Franz?
My second bisect attempt led to this commit: a3bd8447be4ea2ce230eb8ae0e815c04d85fa15a is the first bad commit commit a3bd8447be4ea2ce230eb8ae0e815c04d85fa15a Author: Tom Gundersen <teg@jklm.no> Date: Mon Mar 18 15:12:18 2013 +0100 udev: make firmware loading optional and disable by default Distros that whish to support old kernels should set --with-firmware-dirs="/usr/lib/firmware/updates:/usr/lib/firmware" to retain the old behaviour. :100644 100644 7e9cdfdb261a9dd6515c67d6a6ddb26b8d3da80e 37eeb01d583e1f6141c1f7abbe691ac7ac8427f3 M Makefile.am :100644 100644 28d7286c8338f7d7a5fd521c0755f2082448f4eb 14411e3601f1ed779977293fd756bcb09c49982f M configure.ac :040000 040000 d46e177689a0cb64f023ff3acad8753b9efb936f f594c591fd2e7a07088a87bda62ed35d0395cf01 M rules :040000 040000 60d3299063a9e5bb6e1ff6ba3f327a2534663d98 7888b7faa97987834ba2c62a8dc07d4a24419844 M shell-completion :040000 040000 86cb05a395cf2971569a2501fcd5c15017a25cfd c31aefae5c161640c26348948fc83c1f945b215a M src Reverting it works (even on top of v199). Also tried the patchset 3 and it works with USE=firmware-loader, but not without it.
Created attachment 343550 [details] config of v3.9-rc4-138-gde55eb1 AFAICT it has firmware loading enabled so I expected it to work even without USE=firmware-loader for udev. The same issue was with 3.8.4-gentoo.
confirmed: udev-199-r1[firmware-loader] fixes my issue. With firmware-loader disabled I still had a long bootup time. Also did not check with udev-199 and enabled firmware-loader - probably that would have been enough...
Thanks guys. Can you please make another bug about firmware not getting properly loaded with USE="-firmware-loader" and newer kernels like >= 3.8 and attach your configs there? Please, this is getting a bit messy here Thanks! I'll duplicate this to the systemd (udev) bug *** This bug has been marked as a duplicate of bug 463604 ***
