Re: [PATCH v2 1/2] efi+tpm: Don't access event->count when it isn't mapped.

[Jarkko Sakkinen]

On Wed, Sep 25, 2019 at 09:41:33AM -0700, Jerry Snitselaar wrote:
> On Wed Sep 25 19, Jerry Snitselaar wrote:
> > On Wed Sep 25 19, Jarkko Sakkinen wrote:
> > > On Wed, Sep 25, 2019 at 12:25:05PM +0200, Ard Biesheuvel wrote:
> > > > On Wed, 25 Sep 2019 at 12:16, Jarkko Sakkinen
> > > > <jarkko.sakkinen@xxxxxxxxxxxxxxx> wrote:
> > > > > 
> > > > > From: Peter Jones <pjones@xxxxxxxxxx>
> > > > > 
> > > > > Some machines generate a lot of event log entries.  When we're
> > > > > iterating over them, the code removes the old mapping and adds a
> > > > > new one, so once we cross the page boundary we're unmapping the page
> > > > > with the count on it.  Hilarity ensues.
> > > > > 
> > > > > This patch keeps the info from the header in local variables so we don't
> > > > > need to access that page again or keep track of if it's mapped.
> > > > > 
> > > > > Fixes: 44038bc514a2 ("tpm: Abstract crypto agile event size calculations")
> > > > > Cc: linux-efi@xxxxxxxxxxxxxxx
> > > > > Cc: linux-integrity@xxxxxxxxxxxxxxx
> > > > > Cc: stable@xxxxxxxxxxxxxxx
> > > > > Signed-off-by: Peter Jones <pjones@xxxxxxxxxx>
> > > > > Tested-by: Lyude Paul <lyude@xxxxxxxxxx>
> > > > > Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx>
> > > > > Acked-by: Matthew Garrett <mjg59@xxxxxxxxxx>
> > > > > Acked-by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
> > > > > Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx>
> > > > 
> > > > Thanks Jarkko.
> > > > 
> > > > Shall I take these through the EFI tree?
> > > 
> > > Would be great, if you could because I already sent one PR with fixes for
> > > v5.4-rc1 yesterday.
> > > 
> > > /Jarkko
> > 
> > My patch collides with this, so I will submit a v3 that applies on top of
> > these once I've run a test with all 3 applied on this t480s.
> 
> Tested with Peter's patches, and that was the root cause on this 480s.
> 
> I think there should still be a check for tbl_size to make sure we
> aren't sticking -1 into efi_tpm_final_log_size though, which will be
> the case right now if it fails to parse an event.

You could sent a follow up patch for that I think. The current
ones are kind of already "went through the process" and do right
things but I do agree that a sanity check would make sense just
in case.

/Jarkko