[PATCH] fsck -A goes into infinite loop and kills all processes

[Bernhard Fischer]

On Fri, Feb 01, 2008 at 11:46:11AM +0000, Roy Marples wrote:
>
>On Thu, 2008-01-31 at 20:32 +0000, Roy Marples wrote:
>> On Thu, 2008-01-31 at 18:17 +0000, Roy Marples wrote:
>> > Ah yes, it's because there is no helper installed for ext4dev. Ensuring
>> > that all helpers exist fixes the error. So it's not handling exec errors
>> > correctly. I'll see if I can fix fsck now.
>> 
>> The execute function does not return when pid is <1, so it stores -1 as
>> a pid to kill, which kills quite a bit really :)
>
>Attached is a new patch that free's any malloced memory correctly. BTW,

Yes, but this should be dependant on whether ENABLE_FEATURE_CLEAN_UP is
requested.

>this behaviour should match upstream as if the don't find a helper then
>they free the instance and return even if noexecute is set. As the spawn
>function does that as well, we just need to check if pid is less than 1.