lkml.org 
[lkml]   [1998]   [Jan]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Linux-2.1.83 - fixed irq handling..

On Fri, 30 Jan 1998, Linus Torvalds wrote:

> Ok,
> I just put 2.1.83 on ftp.kernel.org, and this one should fix a nasty
> problem with interrupts in 2.1.82 where a bottom half handler that tried
> to disable interrupts was only able to disable them on the local CPU
> rather than globally. Oops (it is still true that if you try to disable
> interrupts from within a hardware interrupt handler it will refuse to try
> to disable them globally due to deadlock reasons, but the fact that the
> test for that also triggered in a bottom half routine was a big bug).
>
> The fix for this actually also simplifies the SMP IRQ handling further,
> and I'm reasonably confident that it should be correct (modulo any bad
> drivers that try to disable interrupt globally from within an interrupt: I
> hope and believe that is not an issue for us).

Each time the semantic of a kernel internal service is changed, this
makes bad any kernel module that has'nt been fortunate enough not to
conflict with the new semantic.
On the other hand, not all hackers have a SMP machine or are able to
understand instantaneously all the subtleties of IRQ handling with SMP.
Make things a bit stable and, IMO, the modulo will vanish quickly.

> As the patch is so small, I'll just include an uuencoded version here too,
>
> Linus

Gerard.


\
 
 \ /
  Last update: 2005-03-22 13:41    [W:0.040 / U:0.068 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site