lkml.org 
[lkml]   [1996]   [Mar]   [29]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: loop device broken in 1.3.7[89] (by plug/unplug_device ?)


On Thu, 28 Mar 1996, Gerd Knorr wrote:
>
> The loop-device is broken in 1.3.79. I have a script which creates a
> initrd for loading some scsi modules using the nice loop-device. Running
> this script hangs the linux box in 1.3.78 and 79. I can't do anything
> useful, seems disk I/O does'nt work any more. The EIP Alt-ScrollLock
> reports belongs to "hard_idle".
>
> Tried to compile 1.3.79 using drivers/block/ll_rw_blk.c from 1.3.77, now
> it works :-)
> Seems there is something wrong with the plug/unplug_device stuff...

No, I think it's a bug in the loop device driver. Maybe it doesn't
understand large requests or that the request queue is a linked list and
not just one single entry, and the new plug code just changes the way the
queue normally looks enough to confuse loop (and md).

The plug code is not really new: it's just used in a new way that results
in a better layout of the request queue. And it looks like loop and md
made some assumptions about how requests can look which aren't true any
more.

Linus


\
 
 \ /
  Last update: 2005-03-22 13:36    [W:0.059 / U:0.420 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site