linux-kernel - Re: [PATCH] tpm/tpm_tis: Fix variable reset during IRQ probing

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]

Message-ID: <X/+xr/L+m2k5fObZ@kernel.org>
Date:   Thu, 14 Jan 2021 04:51:27 +0200
From:   Jarkko Sakkinen <jarkko@...nel.org>
To:     Tianjia Zhang <tianjia.zhang@...ux.alibaba.com>
Cc:     Peter Huewe <peterhuewe@....de>, Jason Gunthorpe <jgg@...pe.ca>,
        linux-integrity@...r.kernel.org, linux-kernel@...r.kernel.org,
        Jia Zhang <zhang.jia@...ux.alibaba.com>
Subject: Re: [PATCH] tpm/tpm_tis: Fix variable reset during IRQ probing

On Wed, Jan 13, 2021 at 08:00:21PM +0800, Tianjia Zhang wrote:
> In tpm_tis_core_init(), tpm2_probe() will be called first, this
> function will eventually call tpm_tis_send(), and then
> tpm_tis_probe_irq_single() will detect whether the interrupt is
> normal, mainly the installation interrupted, set `priv->irq_tested`
> to false. The logic will eventually be executed to tpm_tis_send()
> to trigger an interrupt.
> 
> There is currently such a scenario, which will cause the IRQ probe
> code to never be executed, so that the TPM device is in polling
> mode: after setting irq_tested to false, an interrupt occurs
> between entering the ttpm_tis_send() function, and the interrupt
> will be first set irq_tested to true will cause the IRQ probe code
> to never be executed.

Can you describe the scenario more detail?

> It seems that this interrupt comes from tpm2_probe(). Although the
> interrupt has not been installed when tpm2_probe() is called, the
> interrupt of tpm2_probe() is only received after IRQ detection.
> 
> This patch solves this issue by introducing a new variable, which
> is only used in interrupts, and irq_tested only marks whether the
> interrupt test has been completed.
> 
> Signed-off-by: Tianjia Zhang <tianjia.zhang@...ux.alibaba.com>
> ---

I'm not sure I understand this patch. TPM should be in polling
mode. This is also assumption before calling tpm_get_timeouts():

/* Before doing irq testing issue a command to the TPM in polling mode
 * to make sure it works. May as well use that command to set the
 * proper timeouts for the driver.
 */
if (tpm_get_timeouts(chip)) {
        dev_err(dev, "Could not get TPM timeouts and durations\n");
        rc = -ENODEV;
        goto out_err;
}

/Jarkko