Re: [PATCH 0/3] efivarfs: bug fixes

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, 2024-12-09 at 10:20 +0100, Ard Biesheuvel wrote:
> On Sun, 8 Dec 2024 at 19:34, James Bottomley
> <James.Bottomley@xxxxxxxxxxxxxxxxxxxxx> wrote:
> > 
> > Patch 1 is stand alone, but 2 depends on 3
> > 
> > Regards,
> > 
> > James
> > 
> > ---
> > 
> > James Bottomley (3):
> >   efivarfs: fix error on non-existent file
> >   efivarfs: fix memory leak on variable removal
> >   efivarfs: fix incorrect variable creation
> > 
> 
> Thanks James,
> 
> I've queued these up now.

Thanks, but I need to redo 3/3: there's a bug where if the variable is
created to do a write which fails, it remains on the list even though
the entry is freed.

It also begs the question: why does this list of variables exist?  All
it does is cause management complexity and overhead and its only
function seems to be to free the entries when the filesystem is
unmounted, which could much more easily be accomplished by implementing
a superblock evict_inode() callback that kfree's i_private, which would
mean the entry was freed when the inode was and thus wouldn't have to
be explicitly freed anywhere.

Regards,

James





[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [ECOS]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux