Re: [patch, 2.5] vmalloc.c error path fixes

[Andrew Morton]

Marcus Alanen wrote:
> 
> I think there are some problems in vmalloc.c.  The two first parts
> of the diff fix a spinlock being held if an error occurs in
> map_vm_area, and the last part fixes the error path of __vmalloc.

Certainly agree on the first chunk.  Listen to the old guy: "Thou
shalt not have more than one return statement per function".

Second chunk looks good too, but perhaps there's another way of doing
it.  The 2.5.31 code just tossed it all at vfree.  Perhaps hch can
comment?