linux-kernel - Re: question about setting TPM_CHIP_FLAG_IRQ in tpm_tis_core

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 12 Nov 2019 13:17:34 -0700
From:   Jerry Snitselaar <jsnitsel@...hat.com>
To:     Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>
Cc:     Stefan Berger <stefanb@...ux.ibm.com>,
        linux-integrity@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: question about setting TPM_CHIP_FLAG_IRQ in tpm_tis_core_init

On Tue Nov 12 19, Jarkko Sakkinen wrote:
>On Mon, Nov 11, 2019 at 08:36:37PM -0700, Jerry Snitselaar wrote:
>> Question about 1ea32c83c699 ("tpm_tis_core: Set TPM_CHIP_FLAG_IRQ
>> before probing for interrupts").  Doesn't tpm_tis_send set this flag,
>> and setting it here in tpm_tis_core_init short circuits what
>> tpm_tis_send was doing before? There is a bug report of an interrupt
>> storm from a tpm on a t490s laptop with the Fedora 31 kernel (5.3),
>> and I'm wondering if this change could cause that. Before they got the
>> warning about interrupts not working, and using polling instead.
>
>Looks like it. Stefan?
>
>/Jarkko
>

Stefan is right about the condition check at the beginning of tpm_tis_send.

	if (!(chip->flags & TPM_CHIP_FLAG_IRQ) || priv->irq_tested)
		return tpm_tis_send_main(chip, buf, len);

Before his change it would've gone straight to calling
tpm_tis_send_main instead of jumping down and doing the irq test, due
to the flag not being set. With his change it should now skip this
tpm_tis_send_main call when tpm_tis_gen_interrupt is called, and then
after that time through tpm_tis_send priv->irq_tested will be set, and
the flag should be set as to whether or not irqs were working.

I should hopefully have access to a t490s in a few days so I can look at it,
and try to figure out what is happening.