Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 9 Jun 2022 10:11:59 +0200
From:   Morten Linderud <morten@linderud.pw>
To:     Jarkko Sakkinen <jarkko@kernel.org>
Cc:     Peter Huewe <peterhuewe@gmx.de>, Jason Gunthorpe <jgg@ziepe.ca>,
        linux-integrity@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] tpm/eventlog: Don't abort tpm_read_log on faulty ACPI
 address
Message-ID: <20220609081159.rjj7a3x63y7ag4ty@terminus>
References: <20220608123109.678343-1-morten@linderud.pw>
 <YqF7NktlAJg26Bvd@iki.fi>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <YqF7NktlAJg26Bvd@iki.fi>
Precedence: bulk

On Thu, Jun 09, 2022 at 07:46:46AM +0300, Jarkko Sakkinen wrote:
> On Wed, Jun 08, 2022 at 02:31:08PM +0200, Morten Linderud wrote:
> > tpm_read_log_acpi() should return -ENODEV when no eventlog from the ACPI
> > table is found. If the firmware vendor includes an invalid log address
> > we are unable to map from the ACPI memory and the function returns -EIO
> > which would abort discovery of the eventlog.
> > 
> > This change ensure we always return -ENODEV in tpm_read_log_acpi() and
> > fallback to the EFI configuration table.
> 
> Please do not use "we" in commit messages. Or start a sentence
> with "this patch", "this commit" or "this change". It is always
> best just to go down to the roots and use imperative form.
> 
> E.g. you could rephrase the last paragraph as
> 
> "Change the return value from -EIO to -ENODEV when acpi_os_map_iomem()
> fails to map the event log."

ack

> > The following hardware was used to test this issue:
> >     Framework Laptop (Pre-production)
> >     BIOS: INSYDE Corp, Revision: 3.2
> >     TPM Device: NTC, Firmware Revision: 7.2
> > 
> > Dump of the faulty ACPI TPM2 table:
> >     [000h 0000   4]                    Signature : "TPM2"    [Trusted Platform Module hardware interface Table]
> >     [004h 0004   4]                 Table Length : 0000004C
> >     [008h 0008   1]                     Revision : 04
> >     [009h 0009   1]                     Checksum : 2B
> >     [00Ah 0010   6]                       Oem ID : "INSYDE"
> >     [010h 0016   8]                 Oem Table ID : "TGL-ULT"
> >     [018h 0024   4]                 Oem Revision : 00000002
> >     [01Ch 0028   4]              Asl Compiler ID : "ACPI"
> >     [020h 0032   4]        Asl Compiler Revision : 00040000
> > 
> >     [024h 0036   2]               Platform Class : 0000
> >     [026h 0038   2]                     Reserved : 0000
> >     [028h 0040   8]              Control Address : 0000000000000000
> >     [030h 0048   4]                 Start Method : 06 [Memory Mapped I/O]
> > 
> >     [034h 0052  12]            Method Parameters : 00 00 00 00 00 00 00 00 00 00 00 00
> >     [040h 0064   4]           Minimum Log Length : 00010000
> >     [044h 0068   8]                  Log Address : 000000004053D000
> > 
> > Signed-off-by: Morten Linderud <morten@linderud.pw>
> > 
> > ---
> > 
> > v2: Tweak commit message and opt to return -ENODEV instead of loosening up the
> >     if condition in tpm_read_log()
> > 
> > ---
> >  drivers/char/tpm/eventlog/acpi.c | 6 +++++-
> >  1 file changed, 5 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/char/tpm/eventlog/acpi.c b/drivers/char/tpm/eventlog/acpi.c
> > index 1b18ce5ebab1..2b15d6eebd69 100644
> > --- a/drivers/char/tpm/eventlog/acpi.c
> > +++ b/drivers/char/tpm/eventlog/acpi.c
> > @@ -136,8 +136,12 @@ int tpm_read_log_acpi(struct tpm_chip *chip)
> >  
> >  	ret = -EIO;
> >  	virt = acpi_os_map_iomem(start, len);
> > -	if (!virt)
> > +	if (!virt) {
> > +		dev_warn(&chip->dev, "%s: Failed to map ACPI memory\n", __func__);
> > +		/* try EFI log next */
> > +		ret = -ENODEV;
> >  		goto err;
> > +	}
> 
> It is wrong to try out EFI, if this fails. TPM2 ACPI table was already
> detected.

The next branch tries out EFI if the eventlog it found is empty, as it created
an empty file. This branch would produce no eventlog if we fail to map the
memory. I don't understand why there would be a difference between these two
branches?

This seems like an oversight after 3dcd15665aca80197333500a4be3900948afccc1

> >  
> >  	memcpy_fromio(log->bios_event_log, virt, len);
> >  
> > -- 
> > 2.36.1
> 
> What you are using this for? Without any actual bug report, this 
> is an obvious NAK.

I have hardware with faulty ACPI values which prevents me from getting a
eventlog. I can surely make a bugreport if it helps the case, but that seems
like an arbiterary hurdle when I have already spent the time tracking down the
issue and proposed a fix.

-- 
Morten Linderud
PGP: 9C02FF419FECBE16