Bug 30382

Summary:	The bad sectors on the DVD freezes the system completely.
Product:	IO/Storage	Reporter:	Ilya (dik_again)
Component:	Serial ATA	Assignee:	Jeff Garzik (jgarzik)
Status:	RESOLVED OBSOLETE
Severity:	normal	CC:	alan, tj
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	2.6.36-ARCH	Subsystem:
Regression:	No	Bisected commit-id:
Attachments:	dmesg, lspci. Full dmesg output after reboot and triggering the problem

Description Ilya 2011-03-02 18:31:16 UTC

if I read from a drive containing scratched disk. the system freezes completely (for a 10-20 sec) at those scratches.

If I use this scratched disk on another drive (another drive is on the IDE controller of the motherboard) everything seems OK.

In Attachment is the relevant information from the dmesg and lspci.

Comment 1 Ilya 2011-03-02 18:32:34 UTC

Created attachment 49922 [details]
dmesg, lspci.

Comment 2 Tejun Heo 2011-03-03 05:47:55 UTC

Please attach full output of dmesg after triggering the problem, preferably with CONFIG_PRINTK_TIME turned on.  Thanks.

Comment 3 Ilya 2011-03-03 08:16:15 UTC

I haven't compiled kernel with a PRINTK_TIME enabled. But I will do it nearly. Here is the dmesg output after reboot and the doing:

[dikiy@linuxhost ~]$ dd if=/dev/sr0 of=/dev/null bs=2048
dd: чтение «/dev/sr0»: Ошибка ввода/вывода
475384+0 записей считано
475384+0 записей написано
 скопировано 973586432 байта (974 MB), 302,28 c, 3,2 MB/c
[dikiy@linuxhost ~]$

Comment 4 Ilya 2011-03-03 08:18:01 UTC

Created attachment 49972 [details]
Full dmesg output after reboot and triggering the problem

Comment 5 Ilya 2011-03-03 08:43:48 UTC

Before the system freezes for a long time (15-20 sec) and the dd exits with an error, there are a lot of "little" freezes (approx. 0.1 sec) when the disc spin down and try to reread the scratch.

Comment 6 Tejun Heo 2011-03-04 11:29:31 UTC

Hmm... I'm a bit lost.  It's a SATA device using an independent port so it shouldn't really affect other devices.  Can you please describe the 'hang' in more detail?  What hangs?  Does the kernel complain about anything when it comes back from it?

Thanks.

Comment 7 Ilya 2011-03-04 21:09:51 UTC

The 'hang' is 'absolute'. The mouse freezes too. The kernel doesn't complain anything else I posted in the attachment.

Comment 8 Ilya 2011-03-04 21:13:11 UTC

Oh. There is 'hang' at the spin up of the disc too.

Comment 9 Ilya 2011-03-05 08:34:12 UTC

It looks, like the whole system wait for an operation-results on the disc.

Comment 10 Alan 2013-12-23 11:37:56 UTC

If this is still seen in modern kernels please update