On Thu Dec 26, 2024 at 6:18 PM EET, Jarkko Sakkinen wrote:
> On Wed Dec 25, 2024 at 9:32 PM EET, Jarkko Sakkinen wrote:
> > The following failure was reported:
> >
> > [ 10.693310][ T1] tpm_tis STM0925:00: 2.0 TPM (device-id 0x3, rev-id 0)
> > [ 10.848132][ T1] ------------[ cut here ]------------
> > [ 10.853559][ T1] WARNING: CPU: 59 PID: 1 at mm/page_alloc.c:4727 __alloc_pages_noprof+0x2ca/0x330
> > [ 10.862827][ T1] Modules linked in:
> > [ 10.866671][ T1] CPU: 59 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.12.0-lp155.2.g52785e2-default #1 openSUSE Tumbleweed (unreleased) 588cd98293a7c9eba9013378d807364c088c9375
> > [ 10.882741][ T1] Hardware name: HPE ProLiant DL320 Gen12/ProLiant DL320 Gen12, BIOS 1.20 10/28/2024
> > [ 10.892170][ T1] RIP: 0010:__alloc_pages_noprof+0x2ca/0x330
> > [ 10.898103][ T1] Code: 24 08 e9 4a fe ff ff e8 34 36 fa ff e9 88 fe ff ff 83 fe 0a 0f 86 b3 fd ff ff 80 3d 01 e7 ce 01 00 75 09 c6 05 f8 e6 ce 01 01 <0f> 0b 45 31 ff e9 e5 fe ff ff f7 c2 00 00 08 00 75 42 89 d9 80 e1
> > [ 10.917750][ T1] RSP: 0000:ffffb7cf40077980 EFLAGS: 00010246
> > [ 10.923777][ T1] RAX: 0000000000000000 RBX: 0000000000040cc0 RCX: 0000000000000000
> > [ 10.931727][ T1] RDX: 0000000000000000 RSI: 000000000000000c RDI: 0000000000040cc0
> >
> > Above shows that ACPI pointed a 16 MiB buffer for the log events because
> > RSI maps to the 'order' parameter of __alloc_pages_noprof(). Address the
> > bug with kvmalloc() and devres_add().
> >
> > Cc: stable@xxxxxxxxxxxxxxx # v2.6.16+
> > Fixes: 55a82ab3181b ("[PATCH] tpm: add bios measurement log")
> > Reported-by: Andy Liang <andy.liang@xxxxxxx>
> > Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219495
> > Suggested-by: Matthew Garrett <mjg59@xxxxxxxxxxxxx>
>
> Oops, needs to be dropped from this.
>
> > Signed-off-by: Jarkko Sakkinen <jarkko@xxxxxxxxxx>
> > ---
> > v6:
> > * A new patch.
> > ---
> > drivers/char/tpm/eventlog/acpi.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/char/tpm/eventlog/acpi.c b/drivers/char/tpm/eventlog/acpi.c
> > index 69533d0bfb51..7cd44a46a0d7 100644
> > --- a/drivers/char/tpm/eventlog/acpi.c
> > +++ b/drivers/char/tpm/eventlog/acpi.c
> > @@ -136,10 +136,12 @@ int tpm_read_log_acpi(struct tpm_chip *chip)
> > }
> >
> > /* malloc EventLog space */
> > - log->bios_event_log = devm_kmalloc(&chip->dev, len, GFP_KERNEL);
> > + log->bios_event_log = kvmalloc(len, GFP_KERNEL);
> > if (!log->bios_event_log)
> > return -ENOMEM;
> >
> > + devres_add(&chip->dev, log->bios_event_log);
> > +
>
> We either need to git revert 441b7152729f ("tpm: Use managed allocation
> for bios event log") OR alternatively use devm_add_action() creating a
> fix that wastes 16 MiB of memory and obfuscates flows more than needed.
>
> I don't necessarily get how come this is "less intrusive"...
Right, so I guess it's not actually very complicated:
ret = devm_add_action_or_reset(&chip->dev, kvfree, &log->bios_event_log)
if (ret) {
log->bios_event_log = NULL;
return ret;
}
Had not used this API since we used it in tpmm_chip_alloc(). I'll tweak
the patch accordingly...
And since 2/2 becomes a feature patch I'll refine it to do a traverse of
the log so we know the actual size of the populated contents and then
move back on using devm_kmalloc() in that patch (i.e. it will render
out also the fix that the first patch makes).
Not too bad, I can live this.
BR, Jarkko
