- Arch-general - lists.archlinux.org

[arch-general] mce after linux-3.11.5-1 on NP900X3C
by Rasmus Liland 29 Nov '14

29 Nov '14

Dear @arch-general readers, I'm experiencing machine check exceptions since every kernel after package linux-3.11.5-1 (Oct 14 2013). I hope some nice people will be able to assist me or perhaps point me in a direction of something fruitful. First here goes some kernel panic output I was able to snap on May 25th (I also made it an attachment): > [19367.116180] Disabling lock debugging due to kernel taint > [19367.116196] mce: [Hardware Error]: CPU 1: Machine Check Exception: 5 Bank 4: b200000000100402 > [19367.116202] mce: [Hardware Error]: RIP !INEXACT! 33:<00007f8b4934c8b7> > [19367.116205] mce: [Hardware Error]: TSC 2824672b8e7 > [19367.116211] mce: [Hardware Error]: PROCESSOR 0:306a9 TIME 14010118857 SOCKET 0 APIC 1 microcode 12 > [19367.116213] mce: [Hardware Error]: Run the above through 'mcelog --ascii' > [19367.116216] mce: [Hardware Error]: Some CPUs didn't answer in synchronization > [19367.116218] mce: [Hardware Error]: Machine check: Invalid > [19367.116220] Kernel panic - not syncing: Fatal machine check on current CPU > [19368.211815] Shutting down cpus with NMI > [19368.222834] Kernel Offset: 0x0 from 0xffffffff81000000 0000 (relocation range: 0xffffffff80000000-0xffffffff9fffffff) > [19368.222942] drm_kms_helper: panic occurred, switching back to text console > [19368.245774] Rebooting in 30 seconds > [19398.323579] ACPI RECOVERY or RESET_REG. Assuming I did a complete system update about 5.4 hours earlier, this implies kernel output for linux kernel version 3.14.4. This has been the case for every kernel after version 3.11.5; from what I can trace in the `/var/cache/pacman/pkg' directory this also calls for _at least_ version 3.15.8; I've removed all other archived versions as they used valuable space. In dmesg output, my good version 3.11.5-1 is called: > Linux version 3.11.5-1-ARCH (tobias@T-POWA-LX) (gcc version 4.8.1 20130725 (prerelease) (GCC)) #1 SMP PREEMPT Mon Oct 14 08:31:43 CEST 2013 The hardware of my system is perhaps relevant. I've got a Samsung NP900X3C-A01SE (Jun.2012) laptop: > Intel Core i5-3317U CPU @ 1.70GHz stepping 9 microcode 0x12 > 4GB RAM (3743M/3888M available, 2048M available to graphics) > SanDisk SSD U100 128GB, 10.01.04, max UDMA/133 > Intel Centrino Advanced-N 6235 AGN, REV=0xB0 Presuming this is a good starting point, if more specific hardware information is needed in this thread in the future I can add the output of e.g. lscpi -vvnn. --Rasmus -- Rasmus Liland, jrl(a)jrl.dyndns.dk

5 21

[arch-general] Persistent black screen of death issues with radeon graphics
by Robbie Smith 28 Nov '14

28 Nov '14

I am at a loss as to what the hell is going on with my laptop. It's a HP Pavillion with the dual AMD graphics (7520G + 7500M), and I will never again buy a laptop containing components from either manufacturer. More often than not when I boot, when the kernel tries to do modesetting the screen will just turn off. I can SSH in, so it's clearly booting up, but the only hint that anything unusual is happening in the logs is this error message: > [ 28.879834] radeon 0000:01:00.0: No connectors reported connected with modes > [ 28.879842] [drm] Cannot find any crtc or sizes - going 1024x768 > [ 28.881975] [drm] fb mappable at 0xE0474000 > [ 28.881978] [drm] vram apper at 0xE0000000 > [ 28.881980] [drm] size 3145728 > [ 28.881982] [drm] fb depth is 24 > [ 28.881985] [drm] pitch is 4096 > [ 28.882332] radeon 0000:01:00.0: fb1: radeondrmfb frame buffer device > [ 28.997034] [drm] Initialized radeon 2.40.0 20080528 for 0000:01:00.0 on minor 1 I feel like I've tried almost everything: - Upgrading and downgrading kernels (I'm currently using 3.17.1) - Various combinations of kernel boot parameters (fbcon=map:0, fbcon=map:1, radeon.modeset=1, radeon.modeset=0, etc) - Adding lines to /etc/X11/xorg.conf.d/20-radeon.conf that tell it which PCI device to map to. - Uninstalling and reinstalling the drivers I used to use catalyst because that was the only driver that worked, but as it doesn't support KMS, I couldn't use GNOME anymore, so switching back to that isn't an option. Searching for the problem doesn't seem to find anything relevant; there's a few forum posts by Ubuntu users but mostly tumbleweed. The bug tracker on freedesktop.org is awkward to navigate so I have no idea if what I'm experiencing has occurred to anyone else. Even just knowing where to go to report a bug or a more appropriate mailing list that could point me in the right direction would be helpful at this stage. The most frustrating bit is that this occurs almost at random. I could use the laptop all day at the library, no issues; turn it off (because suspend is broken on radeon), come home and no screen for what feels like a thousand reboots.

4 5

[arch-general] kdebase-workspace and kdeplasma-addons-frameworks are in conflict
by Damjan Georgievski 25 Nov '14

25 Nov '14

I have KDE Plasma 5 installed, and it needed kdebase-workspace removed. Today when I try to update the system I get a unresolvable conflict. $ sudo pacman -Su :: Starting full system upgrade... resolving dependencies... looking for inter-conflicts... :: kdebase-workspace and kdeplasma-addons-frameworks are in conflict. Remove kdeplasma-addons-frameworks? [y/N] error: unresolvable package conflicts detected error: failed to prepare transaction (conflicting dependencies) :: kdebase-workspace and kdeplasma-addons-frameworks are in conflict The general question is, how can I see what package is pulling kdebase-workspace? -- damjan

6 8

[arch-general] wget and sourceforge.net
by Jude DaShiell 23 Nov '14

23 Nov '14

I found using --max-redirect=1000 and --trust-server-names on a wget command line made downloads from sourceforge.net go much more smoothly. That way, you have to get over 1000 redirects for you download to fail. jude <jdashiel(a)shellworld.net> Twitter: @jdashiel

1 0

Re: [arch-general] Official releases from upstream
by Eli Schwartz 23 Nov '14

23 Nov '14

On Sun, Nov 23, 2014 at 3:35 AM, Ralf Mardorf <ralf.mardorf(a)rocketmail.com> wrote: > Hi, > > while for virtualbox Arch Linux does follow upstream, even while there > is a critical known USB issue, for Claws Mail, where AFAIK isn't a > critical issue, it doesn't follow upstream. > I will go out on a limb here and say if it is flagged out-of-date that implies it is meant to be updated. Or the maintainer would simply re-mark it as up to date. I am sure there could be all kinds of reasons why any specific package could be out of date for a while, without implying that other packages should be left in the same unfortunate condition to match. ;) I try to understand this, but I can't understand it. Is there latitude > for the maintainers to decide what is an official release from upstream > and what isn't? > Obviously the maintainer hasn't gotten around to updating it yet, which is why the package was left with a big question mark over its head for a month. (For a given value of question mark, anyway.) If the maintainer went "nanana I think it's up to date anyway", then maybe that would indicate it isn't following upstream, and we should all start panicking.

1 0

[arch-general] Information: libcups lib32-libcups
by Ralf Mardorf 23 Nov '14

23 Nov '14

Hi, libcups is in extra https://www.archlinux.org/packages/extra/x86_64/libcups/ and it conflicts with the upgrade of lib32-libcups from multilib https://www.archlinux.org/packages/multilib/x86_64/lib32-libcups/ . [rocketmouse@archlinux ~]$ pacman -Qo /usr/lib/cups/backend/usb /usr/lib/cups/backend/usb is owned by libcups 2.0.1-1 [rocketmouse@archlinux ~]$ date ; sudo pacman -Syu Sun Nov 23 15:15:01 CET 2014 [snip] Packages (10): lib32-gettext-0.19.3-1 lib32-glib2-2.42.1-1 lib32-gmp-6.0.0-1 lib32-gnutls-3.3.10-1 lib32-libcups-2.0.1-1 lib32-libgpg-error-1.17-1 lib32-libpng-1.6.15-1 lib32-libtasn1-4.2-1 lib32-nettle-2.7.1-1 lib32-p11-kit-0.22.1-1 [snip] error: failed to commit transaction (conflicting files) lib32-libcups: /usr/lib/cups/backend/usb exists in filesystem Errors occurred, no packages were upgraded. [rocketmouse@archlinux ~]$ sudo rm /usr/lib/cups/backend/usb [rocketmouse@archlinux ~]$ sudo pacman -Syu There are no news about it on https://www.archlinux.org/ and also not on https://www.archlinux.de/ . After removing the existing file, the upgrade does work. I wanted to filed a bug report against lib32-libcups, but somebody already did that: https://bugs.archlinux.org/task/42882?string=lib32-libcups&project=5&search… I don't use cups at the moment, so I'm not sure if it's wise to run rm as I did, if somebody wants to print. Regards, Ralf

2 3

[arch-general] Official releases from upstream
by Ralf Mardorf 23 Nov '14

23 Nov '14

Hi, while for virtualbox Arch Linux does follow upstream, even while there is a critical known USB issue, for Claws Mail, where AFAIK isn't a critical issue, it doesn't follow upstream. On Sun, 23 Nov 2014 09:07:22 +0000 noreply(a)thewildbeast.co.uk wrote: > The latest release is v. 3.11.1, what can you observe with that? "Stable release 3.11.1 (October 27, 2014; 23 days ago)" - https://en.wikipedia.org/wiki/Claws_Mail "Latest Release: 3.11.1 (News), on 27 October 2014" - http://www.claws-mail.org/news.php $ pacman -Si claws-mail | grep Ver Version : 3.11.0-2 ;) I try to understand this, but I can't understand it. Is there latitude for the maintainers to decide what is an official release from upstream and what isn't? Regards, Ralf PS: The Claws package isn't a problem for me. $ pacman -Q claws-mail-git claws-mail-git 3.11.1.r26.g00cade7-1 But the virtualbox package is really an annoyance.

4 5

[arch-general] CMake silently disregards the C preprocessor flags environment variable
by Neven Sajko 23 Nov '14

23 Nov '14

http://www.cmake.org/Bug/view.php?id=12928 Did you know about this? And this is an error in design, it has been like this since the beginning of CMake. It seems like a pretty shameful error for a build system. Although patching for every package using Cmake is probably out of the question, I really feel like this is an issue something should be done about.

1 1

[arch-general] Package search problem on the website
by David Raphaël 21 Nov '14

21 Nov '14

Hi, I was looking for a specific package at https://www.archlinux.org/packages/ and I noticed something weird under the filter repository: there is no entry called 'any'. Thus, if we select one entry, we cannot undo it for a next search. I had to modify the parameters manually in the address bar. Who should I contact to correct this problem? And also, is there a specific reason why there is no category filter like on AUR? Cheers, Raphaël

2 2

[arch-general] New version of "top"?
by Vladimir Nikšić 21 Nov '14

21 Nov '14

I noticed that after a last "$pacman -Syu" that I got a pretty different version of the utility top in the package procps-ng. What exactly happened, how can I get the old-style one back? Thanks. :)

5 7