Subject: Re: [PATCH] tpm: Fix kexec crash due to access to ops NULL pointer
 (powerpc)
To:     Stefan Berger <stefanb@linux.ibm.com>,
        Jason Gunthorpe <jgg@ziepe.ca>
Cc:     jarkko@kernel.org, peterhuewe@gmx.de,
        linux-integrity@vger.kernel.org, Korrapati.Likhitha@ibm.com,
        pavrampu@in.ibm.com, linux-kernel@vger.kernel.org,
        linux-security-module@vger.kernel.org, gcwilson@us.ibm.com,
        linuxppc-dev@lists.ozlabs.org
References: <20211212012804.1555661-1-stefanb@linux.ibm.com>
 <1052cd36-1b85-5d36-045f-5c5bf9f0fc4e@linux.ibm.com>
 <d9eafa8f-4006-4bc2-c09b-6b02a14c6ef3@linux.ibm.com>
 <20211221011320.GM6467@ziepe.ca>
 <f80e33d2-948d-f885-b063-245eb37de8c1@linux.ibm.com>
From:   Tyrel Datwyler <tyreld@linux.ibm.com>
Message-ID: <bad84f8a-9e07-0920-5c07-475e30e710e9@linux.ibm.com>
Date:   Mon, 20 Dec 2021 17:37:36 -0800
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.13.0
MIME-Version: 1.0
In-Reply-To: <f80e33d2-948d-f885-b063-245eb37de8c1@linux.ibm.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 12/20/21 5:31 PM, Stefan Berger wrote:
> 
> On 12/20/21 20:13, Jason Gunthorpe wrote:
>> On Mon, Dec 20, 2021 at 08:05:58PM -0500, Stefan Berger wrote:
>>
>>> diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c
>>> index ddaeceb7e109..4cb908349b31 100644
>>> +++ b/drivers/char/tpm/tpm-chip.c
>>> @@ -473,15 +473,8 @@ static void tpm_del_char_device(struct tpm_chip *chip)
>>>          mutex_unlock(&idr_lock);
>>>
>>>          /* Make the driver uncallable. */
>>> -       down_write(&chip->ops_sem);
>>> -       if (chip->flags & TPM_CHIP_FLAG_TPM2) {
>>> -               if (!tpm_chip_start(chip)) {
>>> -                       tpm2_shutdown(chip, TPM2_SU_CLEAR);
>>> -                       tpm_chip_stop(chip);
>>> -               }
>>> -       }
>>> -       chip->ops = NULL;
>>> -       up_write(&chip->ops_sem);
>>> +       if (chip->ops)
>> ops cannot be read without locking
> 
> This here could be an alternative:

I still think code de-duplication is a good thing. Maybe combine the two
approaches. Call tpm_class_shutdown from tpm_del_char_device and add a chip->ops
check in tpm_class_shutdown to ensure we only do the shutdown when chip->ops is
valid.

-Tyrel

> 
> diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c
> index ddaeceb7e109..7772b475ebc0 100644
> --- a/drivers/char/tpm/tpm-chip.c
> +++ b/drivers/char/tpm/tpm-chip.c
> @@ -474,7 +474,7 @@ static void tpm_del_char_device(struct tpm_chip *chip)
> 
>         /* Make the driver uncallable. */
>         down_write(&chip->ops_sem);
> -       if (chip->flags & TPM_CHIP_FLAG_TPM2) {
> +       if (chip->ops && chip->flags & TPM_CHIP_FLAG_TPM2) {
>                 if (!tpm_chip_start(chip)) {
>                         tpm2_shutdown(chip, TPM2_SU_CLEAR);
>                         tpm_chip_stop(chip);
> 
>    Stefan
> 
> 
>>
>> Jason