Arm-dev

arm-dev@lists.centos.org

686 discussions

introduction
by Jiri Kastner 27 Apr '16

27 Apr '16

hi, my name is jiri kastner and i would like to join centos@arm building. i maintain few fedora packages and realtime is my 'daily bread'. best regards jiri kastner

2 2

Clustering Project for Pi2's using CentOS Userland
by Ed Anderson 25 Apr '16

25 Apr '16

Folks, I'm working on building a Pi2 Cluster and I've been reluctant to fallback to Debian (Jessie). I'd prefer CentOS because I know it better and can get things done - what I'm missing are some of the devel rpms needed for compiling things. I'm not usually building an end-to-end system from source, I usually just get the O/S and build my tools locally but I can't seem to find things like libxslt-devel for Arm7. While I'm a noob to Arm7/Pi2, I'm not a noob to clustering or building stuff - but I'm having trouble finding some of the things I need. Can you point me to a thread, someplace where I can RTFM so I can get my stuff built on the Arm7 platform. My overall intent, after getting my cluster running is to package setup in an easy to deploy image with scripting to setup nodes... and then put it out there for the community to play with and improve. I'm using ClusterLabs s/w with crmsh (because I don't want to build or waste space on Ruby to support pcs). Thoughts? Any existing end-to-end cluster setups for the Pi2? Any help on finding the right pieces for my build would be greatly appreciated! Thanks! -- Digital Mechanic

3 3

Re: [Arm-dev] kernel modules vs builtin
by Jim Perrin 25 Apr '16

25 Apr '16

Sorry for the delayed response, I've been traveling for the last couple weeks. On 04/22/2016 05:15 AM, Sakar Arora wrote: > Hi Jim > Should we use built in kernel support for devices/file systems (like > USB, XFS, SquashFS etc.) needed for CentOS installation, or go with the > modular approach of having the modules placed in the initrd.img? You should not need to change filesystem bits like XFS or SquashFS. That already works out of the box on other gear and shouldn't require modification. > Are all the modules in initrd.img copied to the > initramfs-4.2.0-0.27.el7.1.aarch64.img, created at the completion of > CentOS installation? > The initramfs is created when the kernel package is installed, as part of the post-installation scripts. The installer initramfs is a bit custom, and generated via the lorax command run. The short answer is 'no, not all modules are copied'. > > Right now I am facing an issue, where, if I go with the modular > approach, USB device driver does not get probed correctly and am not > able to install CentOS from USB. This sounds a bit like a driver issue without knowing the specifics. USB already has a fair bit of support compiled in so that it does autodetect devices. This is fairly critical for keyboard/mouse support as well as USB storage. If it's not working, I'd suspect something in how the chipset is communicating to the kernel. > > On the other hand, if I compile the USB driver built in to the kernel, > USB device driver gets probed fine and I am able to proceed with CentOS > installation. Is the module missing from the installer, or present and not working as expected? -- Jim Perrin The CentOS Project | http://www.centos.org twitter: @BitIntegrity | GPG Key: FA09AD77

1 0

Re: [Arm-dev] [arm-dev] Gigabyte MP30-AR0
by Marcin Juszkiewicz 25 Apr '16

25 Apr '16

W dniu 25.04.2016 o 06:51, Ronald Maas pisze: > I quickly started the Centos 7.2 and OpenSuse Tumbleweed installers > from USB drive with acpi=off, but that made no difference. I found > the apm-mustang.dtb file in the linux stable source. Is there a way > to specify which DTB file to use before booting the Linux kernel? "devicetree file.dtb" in grub

1 0

Re: [Arm-dev] [arm-dev] Gigabyte MP30-AR0
by Marcin Juszkiewicz 25 Apr '16

25 Apr '16

W dniu 24.04.2016 o 17:46, Ronald Maas pisze: > On Sun, Apr 24, 2016 at 2:09 AM, Gordan Bobic <gordan(a)redsleeve.org> > wrote: >> Given that we are seeing kernel panics with different SATA, SAS, >> AMD and Nvidia cards, the only thing I can conclude is that there >> is a major bug somewhere in the PCIe driver. The main difference >> between my kernel 4.4.x kernels and the CentOS 4.2.0 is that the >> latter uses 64KB memory pages, so that is probably a good place to >> start looking. > You would expect 64 KB pages to give more issues compared to 4 KB as > the latter has been the norm for decades. But I see if I can buy some > other more PCIe graphics cards on Craigslist for cheap and test again > with the various distros. > Otherwise hope these issues will be ironed out over time when > 96Boards EE boards finally become widely available. I have used Sil SATA controller, Matrox g550, few NVidia cards and two Radeon ones (HD5450 and R240) with APM Mustang (also X-Gene1 like Gigabyte). All with 64KB kernels. Radeon sits in this machine since my "AArch64 desktop week" which was September 2015. Used that card since 3.19 kernel and went through all versions used in Fedora/rawhide. I rather suspect Gigabyte firmware having some issues. Try booting with "acpi=off" and proper DTB to make sure that it is not ACPI related?

2 2

Re: [Arm-dev] [arm-dev] Gigabyte MP30-AR0
by Gordan Bobic 24 Apr '16

24 Apr '16

On 24/04/16 01:18, Ronald Maas wrote: > No luck with any PCIe adapter on Centos 7.2 > 1) SIIG SATA crashes Tianocore when I connect it to a hard drive > http://pastebin.com/yDu3mcCK I haven't tried a Marvell based add-in card myself as I don't have one lying around, but I have tried Silicon Image 3132 SATA and 3ware 9650SE SAS. They both crash only when I load the kernel driver, they all made it past the Tianocore UEFI boot (chain-loaded from u-boot - are you booting directly into Tianocore?) It is worth noting that I managed to get it to not crash when using the CentOS supplied 4.2.0 kernel, but with a mainline 4.4.6 it panics the kernel due to, IIRC, an unaligned access error. Here is the kernel I used: http://ftp.redsleeve.org/pub/misc/aarch64/kernel-4.4.6-1.el7.nosrc.rpm There is 4.4.8 in the same folder, but I haven't tried that with PCIe cards yet as the machine moved from my test bench into production. > 2) Radeon 290X caused kernel panic on boot when using Marcin's > xorg.conf and kernel patch. I tried compiling the radeon driver as > module and as built-in, but that made no difference > http://pastebin.com/be4ucvfk > 3) nVidia GT70 is not recognized by X even when using Marcin xorg.conf > + kernel patch.Compiling nouveau driver as module or built-in made no > difference. Both X -configure and startx returned an error every time > I tried http://pastebin.com/q5ANbtp5 http://pastebin.com/XigAJ6zk Does the kernel nouveau driver load? > In all 3 cases lspci recognized the PCIe adapter http://pastebin.com/EYAErr57 Given that we are seeing kernel panics with different SATA, SAS, AMD and Nvidia cards, the only thing I can conclude is that there is a major bug somewhere in the PCIe driver. The main difference between my kernel 4.4.x kernels and the CentOS 4.2.0 is that the latter uses 64KB memory pages, so that is probably a good place to start looking. Gordan

1 0

Testing some packages from epel-pass-1, patch munin-node for raspberry 2/3
by Mark Verlinde 21 Apr '16

21 Apr '16

Hello, Were conducting a little experiment porting a e-smith/SME based server application to armv7hfp. At this stage we dont let us held back by questioning what feasible user cases could be [1] and make gratefully use of (unsigned) packages in found in /repodir/epel-pass-1/. Just to give a little feedback; we are testing, on a functional level, packages of: shorewall, amavisd-new (including ClamAV), openvpn, collectd and some additional perl- and pythonlibrarys (see attached for detailed list) collectd-rrdtools depends on, go figure, rrdtool which is found in /repodir/c7-pass-1/ but has not made it to the published repos (jet?). Tools I personally use for monitoring/testing are : htop, iotop and munin-node for the raspberry pi munin-node needed a little patch: http://munin-monitoring.org/ticket/1462 (direct link to patch) http://munin-monitoring.org/attachment/ticket/1462/munin-v2.0.6-raspberry_p… <http://munin-monitoring.org/attachment/ticket/1462/munin-v2.0.6-raspberry_p…> Tnx for your hard work!. grt Mark Verlinde 1 <ps>it would take the fun part away, nevertheless <ps> you could argue the memory footprint of ClamAV (ca 350MB) pushed a raspberry pi to it’s hardware limits.

1 0

error message while running lorax command
by Lijun Pan 21 Apr '16

21 Apr '16

Hi, Do any of you know how to fix the following errors while running lorax? " Can't stat exclude path "/install", No such file or directory - ignoring. setfiles get context on / failed: 'Operation not supported' " Thanks, Lijun ...... 2016-04-20 17:10:17,651: cleaning unneeded files cleaning unneeded files 2016-04-20 17:10:17,794: running runtime-cleanup.tmpl running runtime-cleanup.tmpl 2016-04-20 17:10:33,820: creating the runtime image creating the runtime image 2016-04-20 17:10:33,821: no BCJ filter for arch aarch64 no BCJ filter for arch aarch64 mke2fs 1.42.9 (28-Dec-2013) Discarding device blocks: done Filesystem label=Anaconda OS type: Linux Block size=1024 (log=0) Fragment size=1024 (log=0) Stride=0 blocks, Stripe width=0 blocks 131072 inodes, 2097152 blocks 0 blocks (0.00%) reserved for the super user First data block=1 Maximum filesystem blocks=35651584 256 block groups 8192 blocks per group, 8192 fragments per group 512 inodes per group Superblock backups stored on blocks: 8193, 24577, 40961, 57345, 73729, 204801, 221185, 401409, 663553, 1024001, 1990657 Allocating group tables: done Writing inode tables: done Creating journal (32768 blocks): done Writing superblocks and filesystem accounting information: done [87241.832495] EXT4-fs (loop0): mounted filesystem with ordered data mode. Opts: (null) [87263.542522] EXT4-fs (loop0): mounted filesystem with ordered data mode. Opts: (null) Can't stat exclude path "/install", No such file or directory - ignoring. setfiles get context on / failed: 'Operation not supported' Traceback (most recent call last): File "/sbin/lorax", line 323, in <module> main(sys.argv) File "/sbin/lorax", line 225, in main remove_temp=True) File "/usr/lib/python2.7/site-packages/pylorax/__init__.py", line 297, in run compression=compression, compressargs=compressargs) File "/usr/lib/python2.7/site-packages/pylorax/treebuilder.py", line 161, in create_runtime "Anaconda", size=size) File "/usr/lib/python2.7/site-packages/pylorax/imgutils.py", line 107, in mkrootfsimg runcmd(cmd, root=root) File "/usr/lib/python2.7/site-packages/pylorax/executils.py", line 414, in runcmd return execWithRedirect(cmd[0], cmd[1:], **kwargs) File "/usr/lib/python2.7/site-packages/pylorax/executils.py", line 179, in execWithRedirect raise subprocess.CalledProcessError(ret, [command]+argv) subprocess.CalledProcessError: Command '['setfiles', '-e', '/proc', '-e', '/sys', '-e', '/dev', '-e', '/install', '/etc/selinux/targeted/contexts/files/file_contexts', '/']' returned non-zero exit status 255

1 0

Re: [Arm-dev] Gigabyte MP30-AR0
by Gordan Bobic 19 Apr '16

19 Apr '16

On 2016-04-19 03:45, Ronald Maas wrote: > I tried to rebuild the Debian 4.4.6 kernel with radeon driver, a > slightly modified version of your xorg.conf, and applying the kernel > patch as described in your post. But the result is a kernel panic. > > Kernel config file: http://pastebin.com/3usRTzex > > Kernel messages: http://pastebin.com/XssgxCFn > > Note when I inserted a SIIG PCIe SATA card or an nVidia GT 7800 video > card, I got similar kernel panics, which is a bit concerning. > > At this moment it is not possible for me to determine if this is > caused by a hardware problem in the motherboard, a firmware issue, > some X-Gene driver that needs time to mature, or maybe I made a > mistake when building the kernel. Hope it is the latter rather then > the first. > > Appreciate any thoughts how to progress this. Thanks anyway The kernel I am using is here: http://ftp.redsleeve.org/pub/misc/aarch64/kernel-4.4.7-1.el7.nosrc.rpm (add tarball from kernel.org) This is near identical to the 4.4.6 version I was using before, but with fixed up SELinux configuration. I haven't re-tested PCIe add-in cards, but back when I was running 4.4.6 I, too, was having issues with immediate kernel panics when the driver for a SATA or SAS card loads (tested with SIL3132 SATA card, and 3ware SAS card). I never managed to get to the bottom of it, or re-test with 4.4.7. I can confirm that CentOS shipped 4.2.0 kernel works, and the main difference is that that comes configured with 64KB pages. So it could be that this masks the bug that makes the kernel panic as soon as a PCIe card driver is loaded. I haven't had a chance to file a bug report regarding this on kernel.org yet. Gordan

1 0

Work flow for fixing packages for Pi3
by Lalatendu Mohanty 19 Apr '16

19 Apr '16

Hi, I have some questions about the work flow we are following for fixing stuff for CentOS ARM . For example I am trying to make docker run on Pi3. So I tried to rebuild it on Pi3 using centos-7-armhfp.cfg [1] and failed because of a dependency error i.e. no Package found for go-md2man>= 1.0.4. I am sure, to build docker for Pi3 I need to build the dependency chain. But the question is how do I know if a certain is dependency chain is already fixed in CentOS ARM project and I can reuse the packages. Basically I do not want to redo the stuff which is already fixed and available for reuse. Is there a repository for these? The second question is if I get successful at rebuild a package with some changes, how do contribute it back so that others can reuse the stuff? I guess the answer is I should send the patch to arm-dev mailing list. Not sure if we have a wiki page for this kind of information. If we have please point me to that. [1] http://armv7.dev.centos.org/repodir/centos-7-armhfp.cfg Thanks, Lala

2 1

← Newer
1
...
45
46
47
48
49
50
51
...
69
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

Arm-dev