Bug 55355
| Summary: | 2.4.9-6 emu10k1 driver hangs entire machine | ||
|---|---|---|---|
| Product: | [Retired] Red Hat Linux | Reporter: | James Ralston <ralston> |
| Component: | kernel | Assignee: | Arjan van de Ven <arjanv> |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | Brock Organ <borgan> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 7.1 | CC: | ed |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | i686 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2002-07-16 19:42:49 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
This is very similar to what I'm experiencing. FWIW, I was able to run 2.4.9-0.5 from Rawhide without hangs, although it would garble the sounds slightly... Ok, a bit more info. The problem is definitely related to the emu10k driver, as I was able to run for extended periods of time without a hang if I disabled all window-manager sounds (which forms the vast majority of my sound usage). I then noticed that my sound card and my NIC were sharing an interrupt. Thinking that this might be related to the problem, I moved the sound card to give it its own IRQ (damn crappy Dell bios!). The result: a hard hang within five minutes of enabling window manager sounds. FWIW, the sounds are also still garbled... Ed Ok, some more info... I went and grabbed the 2001-11-01 snapshot of the emu10k1 driver: ftp://opensource.creative.com/pub/snapshots/emu10k1-20011101.tar.gz I've been running this on both my home machine (custom-built) and my work machine (a Dell box) since November 1. The good news: I haven't experienced any hard system hangs using this snapshot driver. The bad news: I still occasionally get garbled sounds. I'm not certain, but the incidents of garbling seem to appear less frequently the longer I'm using the machine. (Right after I start up an X session, I get garbled sounds fairly frequently, but after an hour or so, I hardly seem to get any.) I experienced neither hangs nor garbling under 2.4.3. Since I submitted this bug report, I've upgraded all of my machines to Red Hat 7.2, and I still get garbled sounds, static, etc. I don't remember now if I had any lockups using the stock emu10k driver that came with 7.2; as soon as I determined that the static/garbling problem was still there, I went and grabbed the 2001-11-19 snapshot from Creative's web site and installed those modules. Has there been any progress on this? Honestly, for my home machine, I'll just replace the sound card. But the Soundblaster Live cards are popular cards; there needs to be some better resolution to this problem than "replace your sound card". (For example, at work, we have hundreds of Dell PCs with Soundblaster Live cards in 'em.) The 2.4.9-21 kernel we shipped has an "experimental" emu10k1 driver in it, called emu10k1_new. Several people say that this fixed quite a few problems for them. Yup, the _new module seems to have fixed it (modulo the occasional burst of noise when a sound first starts playing)... I haven't had any hangs or lockups with the emu10k1_new module, and I've been banging on it for about a month now. I still get occasional bursts of static, noise, pops, and so forth, though. I suspect the crappy Soundblaster hardware is to blame, and that the old (kernel 2.2.* and 2.4.3) drivers were doing something to hack around the sound card's brokenness. I'd suggest phasing in the experimental "emu10k1_new" driver sooner rather than later, if you can. The evidence that the existing emu10k1 driver is busted is fairly convincing... Confirmed: I have similar problems HardHat:audio> cat /etc/modules.conf alias sound-slot-0 emu10k1 HardHat:audio> uname -a Linux HardHat 2.4.9-21 #1 Thu Jan 17 13:35:37 EST 2002 i686 unknown Since upgrading to 2.5.9 I've had huge audio problems. 100% of the time I get very bad glitches when sound is used after a quiet period. Sometimes audio plays back fine after the glitch, sometimes I get terrible noise instead of audio. After extensive use, I've concluded that the emu10k1 module in Red Hat 7.3 works around the Soundblaster Live problems. After several months of use, I've yet to encounter a single hang, or even any audio distortion at all. If anyone monitoring this bug believes otherwise, then speak up; if I don't see any feedback before the end of the week, I'm going to close this bug with CURRENTRELEASE. |
From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.4) Gecko/20010917 Description of problem: I have a Creative Labs Soundblaster Live card in my system, which worked under kernel-2.4.3-12. I'm using enlightenment as my WM and have sounds enabled, so when I'm running X, my computer is playing sounds frequently. When I upgraded to kernel-2.4.9-6, whenever a sound would stop playing, I would sometimes hear a short a burst of (what appeared to be) static. Occasionally, my entire system would completely hang, wedging itself abruptly and completely. (The keyboard would be completely unresponsive, it wouldn't respond to IP packets, etc.) The hangs seemed to happen right after my system finished playing a sound. I might have been able to get away without a hang for an hour or two, but over a significant period of time (e.g. 4-5 hours), I would always see a hang. A few days ago, I had unconfigured my sound card, so that neither the soundcore nor the emu10k1 drivers were loaded. I haven't had one single hang since that time, and I spent about 6 hours performing a stress test (simultaneously recompiling the kernel, ripping CDs, and running a very active gnutella client). I strongly suspect that the kernel-2.4.9-6 emu10k1 driver and/or soundcore driver is buggy. Or, at the minimum, it doesn't do as good of a job at keeping the SB Live hardware from messing up things: http://www.viahardware.com/scfaq.shtm I intend to replace my SB Live with a Turtle Beach Santa Cruz; after I've done that, I'll see if I still get static/hangs. Still, if there's any way to tweak the existing emu10k1 driver, it would probably be a good idea. (I never ran into problems under kernel-2.4.3-12.) I don't know how easy it will be to reproduce this problem, as it might depend on the specific hardware I have in my system. Your best shot to reproduce is probably to use a system with a SB Live card, and make sure your window manager plays sounds for various actions. If you can get it to generate the static bursts, you'll probably see a hang.