Page: Previous  1, 2, 3

aus9
Status: Assistant
Joined: 21 Sep 2008
Posts: 358
Location: Australia
Reply Quote
theExplorer

I have intel sound, use Alsa and flashplayer from adobe

I have similar issues and sometimes flash site freezes my PC but I think it relates to me using non-free flash and the flash sitte not being able to handle a quirk where it detects my flashplayer version .............already reported here

techpatterns.com/forums/about1678.html

I guess I will have to try out an older kernel to see if that is the answer.
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 1142
Reply Quote
It seems to me that the flashplayer is trying to directly access your sound card when pulseaudio is not detected. I switched to pulseaudio so that I could change the default output device on the fly (laptop headphone out, usb optical audio out, usb headset from logitech for skype, etc).

You may want to force ALSA to use dmix for all outputs or try your hand at setting up PulseAudio here: forums.debian.net/viewtopic.php?t=12497

EDIT: Also note that adobe has been pushing newer versions of flash and it doesn't surprise me that things may not work so well when the audio and video configuration of your system doesn't match what the current big distros implement (ubuntu, opensuse, etc).
Back to top
TheExplorer
Status: Interested
Joined: 21 Sep 2010
Posts: 35
Reply Quote
Sh**, that crappy flashplayer really hangs Firefox. Fortunately, only Firefox is affected and it's easy to SIGTERM it (the system does not hang).
Switched to 2.6.35.8 liquorix and this happened to me, so I apologize for calling your 2.6.36 branch unstable :)


P.S. By the way, damentz, do you plan to update the 2.6.35 branch? It's been updated to 2.6.35.9 already.
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 1142
Reply Quote
I will eventually, maybe sometime this week.
Back to top
TheExplorer
Status: Interested
Joined: 21 Sep 2010
Posts: 35
Reply Quote
By the way, I took the two latest vanilla kernels 2.6.35.9 and 2.6.36.2 and applied only the BFQ I/O Scheduler patch. They compiled without a single issue with disabled unused hardware I don't use like infrared support, bluetooth, firewire, unused file systems and a big bunch of other options. Successfully switched on the userspace cgroup feature. The kernel is small, very fast and responsive.

That should be the thing for you to think over whether we need such a big number of patches that can easily break some kernel API. I personally, as well as many other people out there, really doubt in the need of BFS cpu scheduler.

The kernel shouldn't be the only stumbling block to optimize the system performance. The user has to be familiar with the OS to optimize it's features first and then, if really needed, use the optimized kernel. But, hey, the kernel can be optimized simply by its own options without patching! Patching itself is a very interesting thing... It can be sometimes dangerous and can prejudice the system stability and security.
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 1142
Reply Quote
Your experience with the vanilla kernel totally depends on your workflow. If you don't exploit the weaknesses of the kernel, then you will not know why there are so many patches in the Zen-Kernel sources.

For users that play games, watch videos, and multi task at the same time, a properly tuned process scheduler is necessary to make all these tasks appear to run simultaneously. Humans, according to Con Kolivas, can visually detect 6ms of jitter. This means that a process scheduler that is unfairly distributing CPU bandwidth can cause jitters or inconsistencies on their screen that ruin their experience on the computer. I cherry-picked patches and/or patch sets that adjust the behavior to make problems like these more tolerable.

Maybe you're really big on virtual machines and you're constantly hitting the limit of your physical ram consumption. Everytime you want to load a new application, everything stalls. Everytime you try to run an rsync backup, your computer becomes useless. I put in patches to help mediate this problem too.

What if you copy or write lots of files and you notice a new version of the kernel radically reduces your throughput when writing to disk. I added patches to correct this bug too.

In short, if your workflow doesn't consist of actions that exploit weaknesses in the kernel, then you'll find that switching to Liquorix based kernels will be fruitless.
Back to top
Display posts from previous:   
Page: Previous  1, 2, 3
All times are GMT - 8 Hours