Linux Hangs!

nikhiljosh · 07-10-2001, 06:21 AM

I have been using Linux since RHL 5.2 and haven't faced any major problems since then.
However with PCQ Linux 7.1 , which was distributed by PCQuest in it's July 2001 issue,I am facing
severe problems as described below.

I have Cyrix MII 300 Processor,64 MB SDRam,700 MB Hard Disk Space reserved for Linux,Yamaha Sound Card,Sis 6215c PCI graphic adapter (No prob with this thing though)

Now whenever I start playing mp3 using either mpg123 in character mode or XMMS in GNOME and work on something else like just managing files or editing a text file ,etc.. the LINUX simply hangs!
Even if I press <alt><F2> for another terminal nothing happens.
Also while multitasking the quality of sound is pathetic.

I reinstalled RHL 6.2 and I did not find any of the problems.Linux is as stable as it is known for.

I tried not to do *any* other jobs while playing mp3 and still 7.1 hangs !
Please help me as I am very frustrated by this as I love mp3's as much as I love Linux.

P.S.

While booting LINUX gives following messages which are not displayed by 6.2

hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hda: dma_intr: error=0x84 { DriveStatusError BadCRC }
hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hda: dma_intr: error=0x84 { DriveStatusError BadCRC }
hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hda: dma_intr: error=0x84 { DriveStatusError BadCRC }
hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hda: dma_intr: error=0x84 { DriveStatusError BadCRC }

Could This be the problem ?

jharris · 07-10-2001, 07:54 AM

I'm not sure about the DMA errors. Have you got DMA enabled by default? If so try turning it off and seeing if that helps (I think something like hdparm -d 0 /dev/hda /dev/hdb /dev/etc... should do it) and see if its causing you the problems. I'm just thinking that if its getting CRC errors then it might be retrying a lot hence chewing CPU time. Check out your process list with top and see whats using your processor cycles.

What user you running things as? If its root then try running your progs as a non privileged user, that way you can find out if its hanging hardware wise, or wether its a severe software crash. A non-root software crash shouldn't bring the system down as the kernel will (hopefully!) beat it into submission before hand.

HTH

Jamie...

isajera · 07-10-2001, 05:16 PM

i've heard a lot of horror stories about the linux cd's that come with mags of any kind. it's probably best to avoid them altogether. even if you get past this one prob, you'll probably encounter another one right away. i'd recommend going back to 5.2 or 6.2 if you can.

nikhiljosh · 07-10-2001, 05:22 PM

No, in *character mode* no process is using considerable cpu time.
mpg123 takes about 29% to 40 % of the cpu.(Quite normal I guess)
vmstat gives around 45 MB free RAM.

I also tried using non previleged user but the result is same.
I started a mp3 and after some time the system just froze!

About the bad crc
I installed the system using "Check For Bad Blocks Option".
This didn't help.
Anyways this problem did'nt crop up in RHL 6.2

Guess will have to wait for 7.2

DMR · 07-10-2001, 06:02 PM

Quote:

Originally posted by nikhiljosh
About the bad crc.
I installed the system using "Check For Bad Blocks Option". This didn't help. Anyways this problem did'nt crop up in RHL 6.2. Guess will have to wait for 7.2

It's a UDMA problem, either drive or cable related. Apparently, the older kernels didn't attempt UDMA mode, which would explain why you don't see the errors under RH6.2's kernel. Basically, the newer kernel in 7.1 is just telling you something that's probably been going on for a while. From the linux-kernel mailing list FAQ:

Quote:

In UDMA modes, each transfer is checksum'ed for integrity (like Ultra2 SCSI, and more robust than normal SCSI's parity checking). When a transfer fails this test, it is retried, and this
warning is reported. Seeing these warnings occasionally is not unusual or even a bad thing - they just inflate your logs a little. If this really bothers you, you can comment out the warning
in the driver. Seeing lots of these warnings (multiple per second) is almost certainly a sign that your IDE hardware is broken. For reference, all IDE must:
1. have a cable length of 18" or less
2. have both ends plugged in (no stub)
3. be 80-conductor cable if you're using a mode > udma33.
IDE modes are generally also generated from the system clock, so if you're overclocking (for instance, running AGP at 75 MHz), you're violating IDE specs, and should not expect correct behavior. Similarly, it's possible for your controller's driver to get timing parameters wrong, but this is certainly not the first explanation to adopt.

As suggested before, try turning off UDMA to see if the errors vanish. Also, try a new/shorter cable.

HTH

nikhiljosh · 07-10-2001, 09:57 PM

It worked !
I set the DMA mode off and everything was fine.
Just one question though -
How to permanently set the DMA off ?

DMR · 07-11-2001, 08:35 PM

I think you can muck with DMA using hdparm, but I haven't tried it. The Ultra-DMA mini HOWTO has a little more on the subject.

cinnix · 07-12-2001, 12:33 AM

There are a bunch of preconfigured hdparm settings in /etc/rc.sysinit. It looks to me like it is a 'probe and set'' deal.

jharris · 07-12-2001, 03:32 AM

Quote:

Originally posted by nikhiljosh
How to permanently set the DMA off ?

There's an option when you build the kernel to 'Use DMA by default' so it thats turned on then you'll need to turn it off then rebuild. Or just edit the startup scripts to add the hdparm line, this is probably the easier method unless you want and excuse to recompile.

BTW - mpg123 taking 29-40% of CPU time?!?! On my server (IDT WinChip-2/233) it only takes 9-12%... Did you build it from source or use the RPM (if there is one). The reason I asked is if your CPU supports MMX instructions then there is an option to build mpg123 so that it uses MMX. I've never tried it without but I'll assume that's why my CPU utilisation is so much lower, especially on a slower machine (although I don't know how the WinChip-2 compares to the MII on things like FPU etc...)

Cheers

Jamie...

saroj · 09-19-2001, 07:36 AM

Hi all,
I have install PCQ linux-7.1 in my PC. But Iam not
able to do telnet or ftp to other m/c. I have enable wu-ftpd, gss-ftpd, telnet, tftp through setup and /etc/xinetd.d/* . Other m/c able to do ftp & telnet to my m/c. I am able to do telnet,ftp to my m/c(localhosts). In linux 6.1, I am able to
do all these services i.e from linux-7.1 to linux- 6.1 but not reverse.
Anyone have fased this problem ??
Pl. tell me how to overcome this problem..

Thanks in advance,
Saroj