[SOLVED] Requests for -current (20151216)

atelszewski · 02-06-2016, 10:08 AM

Hi,

What is the policy of updating xf86-video-intel?
Maybe it could/should be updated to the latest git?

--
Best regards,
Andrzej Telszewski

gmgf · 02-06-2016, 10:49 AM

new libsndfile-1.0.26

Version 1.0.26 (November 22 2015) Fix for CVE-2014-9496, CVE-2014-9756 and CVE-2015-7805. Add ALAC/CAF support. Minor bug fixes and improvements.

http://www.mega-nerd.com/libsndfile/...-1.0.26.tar.gz

bassmadrigal · 02-06-2016, 04:12 PM

Quote:

Originally Posted by atelszewski

What is the policy of updating xf86-video-intel?
Maybe it could/should be updated to the latest git?

I don't think there is any policy. If there's bugfixes or decent improvements, then it makes sense to try different versions. Looking at their commit log, it looks like there's been quite a bit of work on SNA. Currently, the default is UXA because there were issues with SNA earlier in -current.

Pat and team may already be looking at this, but it probably wouldn't hurt if you could grab the latest git and see if there's improvements that makes it worth upgrading to and then reporting it here.

atelszewski · 02-06-2016, 04:36 PM

Hi,

Quote:

but it probably wouldn't hurt if you could grab the latest git and see if there's improvements that makes it worth upgrading to and then reporting it here.

I'm already doing this for quite some time. My version as of now is xf86-video-intel-git_20160117_371219c-x86_64-1, unfortunately I don't know which commit it was exactly.

When it comes to the improvements, well, there is something going on. I'm running on a Broadwell system with HD 5500 graphics, with SNA enabled. I used to see graphics related errors in dmesg, especially when using Qt Creator, but I don't see those errors any more. It is hard to tell if the situation has improved due to using newer driver or rather the whole kernel, mesa, and friends being updated. My experience tells me that it's due to the overall stack improvement rather than solely the driver update.

I'm about to update to the latest commit now, just for fun as it looks like the graphics stack works quite good.

EDIT:
OK, I'm running the latest commit as of today, we'll see how it works.
Anyway, I think that any SNA related updates are worth updating the driver as UXA is no go for me. It's extremely slow when web-browsing, SNA on the other side works fine.

--
Best regards,
Andrzej Telszewski

atelszewski · 02-06-2016, 05:52 PM

Hi,

Maybe libvdpau-va-gl could be added?
It is the only missing bit to enable VDPAU on my system, looks like all the other libva* stuff is already present.
I don't have much knowledge about VDPAU/VA-API, so I can't say if that's the good choice.

EDIT:
OK, I can see that adding libvdpau-va-gl and using it with mplayer lowers system load when watching a movie. VDPAU gives me 6[percent] load, other drivers from 12 to 30. So it looks like this thing is worth considering

--
Best regards,
Andrzej Telszewski

atelszewski · 02-06-2016, 06:21 PM

Hi,

Quote:

but I don't see those errors any more.

You do know this situation when something just happens at the same moment when you say it is not happening any more:

Code:

[ 2623.457955] ------------[ cut here ]------------
[ 2623.457985] WARNING: CPU: 3 PID: 2788 at drivers/gpu/drm/i915/intel_uncore.c:619 hsw_unclaimed_reg_debug+0x69/0x90 [i915]()
[ 2623.457987] Unclaimed register detected before writing to register 0xa068
[ 2623.457988] Modules linked in: rfcomm bnep xt_conntrack ipv6 xt_tcpudp xt_limit iptable_mangle iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_filter ip_tables x_tables joydev btusb btrtl btbcm btintel bluetooth hid_generic usbhid fuse snd_hda_codec_realtek snd_hda_codec_hdmi snd_hda_codec_generic coretemp hwmon iwlmvm intel_rapl iosf_mbi mac80211 x86_pkg_temp_thermal intel_powerclamp kvm_intel kvm iwlwifi irqbypass crct10dif_pclmul cfg80211 crc32_pclmul crc32c_intel psmouse rfkill snd_hda_intel serio_raw snd_hda_codec snd_hda_core mei_me snd_hwdep snd_pcm e1000e snd_timer ptp snd soundcore i2c_dev mei lpc_ich shpchp pps_core ehci_pci i2c_i801 ehci_hcd thermal fan battery dw_dmac dw_dmac_core i2c_hid fjes elan_i2c evdev hid sdhci_acpi sdhci mmc_core 8250_dw i2c_designware_platform
[ 2623.458054]  tpm_tis i2c_designware_core tpm acpi_als kfifo_buf processor acpi_pad industrialio loop i915 video button drm_kms_helper drm intel_gtt agpgart i2c_algo_bit fb_sys_fops syscopyarea sysfillrect sysimgblt i2c_core ext4 mbcache jbd2
[ 2623.458074] CPU: 3 PID: 2788 Comm: kworker/u8:0 Not tainted 4.4.1 #1
[ 2623.458075] Hardware name:                  /NUC5i3RYB, BIOS RYBDWi35.86A.0350.2015.0812.1722 08/12/2015
[ 2623.458085] Workqueue: i915 gen6_pm_rps_work [i915]
[ 2623.458087]  ffffffffc0263a10 ffff880098227c40 ffffffff8130218f ffff880098227c88
[ 2623.458091]  ffff880098227c78 ffffffff8108c132 ffff88023c3b0000 000000000000a068
[ 2623.458094]  000000000000a068 ffff88023c3b0080 00000000000027ac ffff880098227cd8
[ 2623.458097] Call Trace:
[ 2623.458102]  [<ffffffff8130218f>] dump_stack+0x44/0x55
[ 2623.458107]  [<ffffffff8108c132>] warn_slowpath_common+0x82/0xc0
[ 2623.458110]  [<ffffffff8108c1bc>] warn_slowpath_fmt+0x4c/0x50
[ 2623.458114]  [<ffffffff810b92a5>] ? update_curr+0xc5/0x130
[ 2623.458129]  [<ffffffffc01e00a9>] hsw_unclaimed_reg_debug+0x69/0x90 [i915]
[ 2623.458142]  [<ffffffffc01e3048>] gen8_write32+0x68/0x1c0 [i915]
[ 2623.458153]  [<ffffffffc0198d80>] gen6_set_rps_thresholds+0xf0/0x320 [i915]
[ 2623.458163]  [<ffffffffc01a36cb>] gen6_set_rps+0x6b/0x260 [i915]
[ 2623.458174]  [<ffffffffc01a3c8c>] intel_set_rps+0x1c/0x30 [i915]
[ 2623.458182]  [<ffffffffc0192b8c>] gen6_pm_rps_work+0x11c/0x320 [i915]
[ 2623.458185]  [<ffffffff810a2db7>] process_one_work+0x147/0x460
[ 2623.458188]  [<ffffffff810a3118>] worker_thread+0x48/0x490
[ 2623.458190]  [<ffffffff810a30d0>] ? process_one_work+0x460/0x460
[ 2623.458192]  [<ffffffff810a30d0>] ? process_one_work+0x460/0x460
[ 2623.458195]  [<ffffffff810a8339>] kthread+0xc9/0xe0
[ 2623.458198]  [<ffffffff810a8270>] ? kthread_worker_fn+0x160/0x160
[ 2623.458202]  [<ffffffff8166a7df>] ret_from_fork+0x3f/0x70
[ 2623.458205]  [<ffffffff810a8270>] ? kthread_worker_fn+0x160/0x160
[ 2623.458207] ---[ end trace 170f7b3cb176c8b2 ]---

I'm a bit tired of this and don't know where does it come from. Anyway, nothing has hanged, so it's not bad

--
Best regards,
Andrzej Telszewski

gmgf · 02-07-2016, 09:33 AM

new fuse-2.9.5 fix:

https://github.com/libfuse/libfuse/releases

gmgf · 02-07-2016, 09:56 AM

stunnel-5.30 is available:

https://www.stunnel.org/sdf_ChangeLog.html

many version are available since 5.23

https://www.stunnel.org/versions.html

https://www.stunnel.org/downloads.html

gmgf · 02-07-2016, 11:22 PM

again new python-setuptools-20.0:

https://pypi.python.org/pypi/setuptools

gmgf · 02-08-2016, 01:42 AM

new libzip-1.1.1:

http://www.nih.at/libzip/

http://www.nih.at/libzip/NEWS.html

Build fixes for Linux

Fix some warnings reported by PVS-Studio

gmgf · 02-08-2016, 04:31 AM

mcabber 1.0.1 released.

loudmouth 1.5.2 released.

http://www.lilotux.net/~mikael/mcabber/

ReaperX7 · 02-08-2016, 06:26 AM

Quote:

Originally Posted by atelszewski

Hi,

I'm already doing this for quite some time. My version as of now is xf86-video-intel-git_20160117_371219c-x86_64-1, unfortunately I don't know which commit it was exactly.

When it comes to the improvements, well, there is something going on. I'm running on a Broadwell system with HD 5500 graphics, with SNA enabled. I used to see graphics related errors in dmesg, especially when using Qt Creator, but I don't see those errors any more. It is hard to tell if the situation has improved due to using newer driver or rather the whole kernel, mesa, and friends being updated. My experience tells me that it's due to the overall stack improvement rather than solely the driver update.

I'm about to update to the latest commit now, just for fun as it looks like the graphics stack works quite good.

EDIT:
OK, I'm running the latest commit as of today, we'll see how it works.
Anyway, I think that any SNA related updates are worth updating the driver as UXA is no go for me. It's extremely slow when web-browsing, SNA on the other side works fine.

--
Best regards,
Andrzej Telszewski

SNA is a mixed bag. It works for some systems, but others it causes problems so you have to experiment with SNA and UXA to see what works best for your system.

I've used SNA without a problem but then again the system I used it on was a Sandybridge system anyways.

mats_b_tegner · 02-08-2016, 01:03 PM

cURL 7.47.1
https://curl.haxx.se/changes.html

atelszewski · 02-08-2016, 03:29 PM

Hi,

For the non-root cifs mounts could the steps and additions provided here be followed?
I wouldn't request it, but the cifs-utils package requires to be rebuilt for the changes to take effect.

--
Best regards,
Andrzej Telszewski

USUARIONUEVO · 02-08-2016, 07:19 PM

X.Org Server 1.18.1 Has Fixes For GLAMOR, Modesetting & DRI3 Presentation

Arround 50 fixes

http://ftp.x.org/archive//individual...1.18.1.tar.bz2