Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 266396 - Kernel (any ver. of gentoo-sources) with "SCSI CDROM support" and "ATA_PIIX" fails in writing DVD-R.
Summary: Kernel (any ver. of gentoo-sources) with "SCSI CDROM support" and "ATA_PIIX" ...
Status: RESOLVED WONTFIX
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Core system (show other bugs)
Hardware: AMD64 Linux
: High critical
Assignee: Gentoo Kernel Bug Wranglers and Kernel Maintainers
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-04-16 16:41 UTC by Zorzo Luca
Modified: 2010-01-01 22:50 UTC (History)
1 user (show)

See Also:
Package list:
Runtime testing required: ---


Attachments
Info and errors. (info and errors,26.47 KB, text/plain)
2009-04-16 16:43 UTC, Zorzo Luca
Details
Full Dmesg and Configs (dmesg and configs.tar.gz,35.81 KB, text/plain)
2009-04-16 21:19 UTC, Zorzo Luca
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Zorzo Luca 2009-04-16 16:41:59 UTC
My gentoo fails in writing blank dvd-r (DVD-R, no cd-r or dvd+rw; never tried with dvd+r): every used app (k3b, growiso, wodim, nero linux, imgburn with wine) can only burn lead-in and then hangs.
When everything hangs i've 100% cpu usage caused by "scsi_eh_0".
Every app that was using the burner will hangs too (as hal polling) and i can't kill them normally (only with AltGr+Stamp+E/I).
When a blank dvd is inserted sys log reports this:

cdrom: This disc doesn't have any tracks I recognize!
sr 0:0:1:0: [sr1] Unhandled sense code
sr 0:0:1:0: [sr1] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE,SUGGEST_OK
sr 0:0:1:0: [sr1] Sense Key : Blank Check [current]
sr 0:0:1:0: [sr1] Add. Sense: No additional sense information
end_request: I/O error, dev sr1, sector 0
Buffer I/O error on device sr1, logical block 0
sr 0:0:1:0: [sr1] Unhandled sense code
sr 0:0:1:0: [sr1] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE,SUGGEST_OK
sr 0:0:1:0: [sr1] Sense Key : Blank Check [current]
sr 0:0:1:0: [sr1] Add. Sense: No additional sense information
end_request: I/O error, dev sr1, sector 0
Buffer I/O error on device sr1, logical block 0

I've found that this happens only if i have "SCSI CDROM support" and so devices are /dev/sr*.
If i turn it off and uses old ide drivers (PIIX_IDE instead of ATA_PIIX) everything is ok.
There are scsi problems only with dvd-r, different media types write ok, and in other operating systems it's all good.

P.S. i'm running stable amd64. Using a sabayon x86 liveDVD with kernel 2.6.27 (the same that i have) i can burn correctly.

Another thing: it's ages that i have this problem, but it was 50% ok/50% fail. Now i've 100% fails. The only thing that is changes is gcc that was 4.1.2 and now is 4.3.2-r3. I'm compiling kernels (2.6.26-r4, 2.6.27-r7, 2.6.27-r8,2.6.28-r*,2.6.29-r1) with this new one.



Reproducible: Always

Steps to Reproduce:
1. Compile kernel with "SCSI CDROM support";
2. Insert a blank dvd-r;
3. Read syslog's errors;
4. Try to burn something and watch it fail.

Actual Results:  
System hangs with 100% cpu usage, dvd-r fails to write, after this sometimes i can't shutdown the system.


See the attachment.
Comment 1 Zorzo Luca 2009-04-16 16:43:55 UTC
Created attachment 188592 [details]
Info and errors.
Comment 2 Zorzo Luca 2009-04-16 21:19:12 UTC
Created attachment 188615 [details]
Full Dmesg and Configs

With old driver and no scsi it works really good. This is only a workaround, of course (i hope...),and as every workaround there are little problems:

[    0.970309] piix 0000:00:1f.1: IDE controller (0x8086:0x266f rev 0x03)
[    0.970321] PIIX_IDE 0000:00:1f.1: PCI INT A -> GSI 18 (level, low) -> IRQ 18
[    0.970339] piix 0000:00:1f.1: not 100% native mode: will probe irqs later
[    0.970345]     ide0: BM-DMA at 0xf000-0xf007
[    0.970352]     ide1: BM-DMA at 0xf008-0xf00f
[    0.970356] Probing IDE interface ide0...
[    1.642439] hda: PLEXTOR DVDR PX-760A, ATAPI CD/DVD-ROM drive
[    2.356438] hdb: PLEXTOR DVDR PX-760A, ATAPI CD/DVD-ROM drive
[    2.407314] hda: host max PIO4 wanted PIO255(auto-tune) selected PIO4
[    2.407445] hda: UDMA/66 mode selected
[    2.407541] hdb: host max PIO4 wanted PIO255(auto-tune) selected PIO4
[    2.407646] hdb: UDMA/66 mode selected
[    2.407738] Probing IDE interface ide1...
[    2.926322] ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
[    2.929665] ide1 at 0x170-0x177,0x376 on irq 15

It takes a lot because kernel has to probe irqs twice (not 100% native mode: will probe irqs later). There is a way to force them manually? With scsi my kernel load time was 4.89 sec, now it's 6.50/6.60 sec .

Here is attached full dmesg and kernel .config ( with and without scsi )
Comment 3 Lars Wendler (Polynomial-C) (RETIRED) gentoo-dev 2009-04-18 19:21:59 UTC
Please choose the correct mime-type when you attach compressed files (binary file (application/octet-stream)).
Comment 4 Mike Pagano gentoo-dev 2009-05-01 17:03:20 UTC
Can you please test with vanilla-sources-2.6.29.2?
Comment 5 Zorzo Luca 2009-05-01 17:32:31 UTC
Can i test it with my main gentoo? Or i should make a test box? (There could be problems?)
Comment 6 George Kadianakis (RETIRED) gentoo-dev 2009-05-08 18:51:24 UTC
(In reply to comment #5)
> Can i test it with my main gentoo? Or i should make a test box? (There could be
> problems?)
> 

Well, just install vanilla-sources-2.6.29.2 like you would any Linux kernel. Don't delete your previous kernels and if this one doesn't work, you can just change your GRUB/LILO config and pop into your last working kernel version!

If you require any more assistance, don't hesitate to ask.
Comment 7 Zorzo Luca 2009-05-08 21:22:20 UTC
I have tested vanilla-sources-2.6.29.2 and the problem is still here: i can burn cds normally, but with dvd-r it can write only the lead-in. The only difference is that cpu was not at 100% only after the lead-in: scsi_eh_0 starts with 1/2 100% cpu spikes every second also during lead in, and then there is 100% cpu until shutdown.
In the end it's the same thing, only a little (very little) worse.
Comment 8 Daniel Drake (RETIRED) gentoo-dev 2009-05-11 23:08:59 UTC
Let's move to reporting this upstream... Can you reproduce this on the latest development kernel, currently v2.6.30-rc4?
Please also examine "dmesg" after the bug appears...are there any messages or errors logged related to this problem?
Comment 9 Zorzo Luca 2009-05-12 08:16:49 UTC
It's really necessary the 2.6.30-rc4 test? This is because every time i try a kernel i've problems with my 30 partitions and 12 lvms. I've already tested vanilla-sources-2.6.29.2, and errors was the same in the the dmsg log attached.
Comment 10 Daniel Drake (RETIRED) gentoo-dev 2009-05-12 13:18:25 UTC
There's no immediately obvious fix that I can find, sorry. If you want to move forward on this then we need to report upstream which really needs you to test the latest code, and later will probably involve testing various other kernels and patches. If you don't have time then that's understandable, but I don't know of any other option for attacking this issue.
Comment 11 Zorzo Luca 2009-05-25 08:55:45 UTC
Sorry, i can't do this. Anyway i think that his bug should stay here (closed, of course), maybe someone else out there has the same problem, and this bug can help.