Bug#717805: Patch for dealing with newer initramfs images with microcode headers
On Tue, Oct 29, 2013 at 06:58:11PM +0100, Olivier Berger wrote:
> Hi.
>
> On Thu, Sep 26, 2013 at 06:19:38AM +0100, Ben Hutchings wrote:
> > On Tue, 2013-09-24 at 16:11 +0100, Brett Parker wrote:
> > > Hi,
> > >
> > > Here's a patch to lsinitramfs to deal with initramfs images that start
> > > with the microcode archive and then a real archive afterwards.
> >
> > Thanks, Brett.
> >
>
> What do you think of the proposed version I'm attaching ?
>
> I don't know about the +8 offset, but I do hope I have addressed the rest of the comments.
>
> Not tested on various compressions, but basically helps solve this bug here, AFAICT.
>
> I'm adding the new file as well, as it may help others without having to replay the patch.
>
> Na warranty whatsoever.
Instead of complicated decoding logic, wouldn't it be easier if the
microcode part (and any other part that may be concatenated) was
required to be compressed in the same way as the main initramfs cpio?
All the relevant compression methods allow compressed files to be
concatenated and result in the concatenation of the input files when
decompressed.
Decompression or listing would then only have to consist of one
zcat/xzcat/... of the archive and running cpio repeatedly until the
stream is exhausted instead of just once.
--
Andreas Bombe
Reply to: