All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Christian König" <ckoenig.leichtzumerken@gmail.com>
To: Lionel Landwerlin <lionel.g.landwerlin@intel.com>,
	dri-devel@lists.freedesktop.org, david1.zhou@amd.com
Subject: Re: [PATCH] dma-buf: fix stack corruption in dma_fence_chain_release
Date: Mon, 5 Aug 2019 15:02:51 +0200	[thread overview]
Message-ID: <af470fba-fb46-e0c4-caa0-220d6083ffc3@gmail.com> (raw)
In-Reply-To: <1ad7ef20-0207-7f80-ee12-ec98c1d5dfcb@intel.com>

Not even remotely :)I tested this with my own crafted code inside the 
kernel.

It's probably quite some hassle to actually trigger this problem from 
userspace and I only found it because I created a very very long 
sequence chain by accident.

Christian.

Am 05.08.19 um 14:03 schrieb Lionel Landwerlin:
> By any change, did you run into this with a CTS test whose name ends 
> with ".chain" ? :)
>
> -Lionel
>
> On 05/08/2019 10:36, Christian König wrote:
>> We can't free up the chain using recursion or we run into a stack 
>> overflow.
>>
>> Manually free up the dangling chain nodes to avoid recursion.
>>
>> Signed-off-by: Christian König <christian.koenig@amd.com>
>> ---
>>   drivers/dma-buf/dma-fence-chain.c | 24 +++++++++++++++++++++++-
>>   1 file changed, 23 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/dma-buf/dma-fence-chain.c 
>> b/drivers/dma-buf/dma-fence-chain.c
>> index b5089f64be2a..44a741677d25 100644
>> --- a/drivers/dma-buf/dma-fence-chain.c
>> +++ b/drivers/dma-buf/dma-fence-chain.c
>> @@ -178,8 +178,30 @@ static bool dma_fence_chain_signaled(struct 
>> dma_fence *fence)
>>   static void dma_fence_chain_release(struct dma_fence *fence)
>>   {
>>       struct dma_fence_chain *chain = to_dma_fence_chain(fence);
>> +    struct dma_fence *prev;
>> +
>> +    /* Manually unlink the chain as much as possible to avoid recursion
>> +     * and potential stack overflow.
>> +     */
>> +    while ((prev = rcu_dereference_protected(chain->prev, true))) {
>> +        struct dma_fence_chain *prev_chain;
>> +
>> +        if (kref_read(&prev->refcount) > 1)
>> +               break;
>> +
>> +        prev_chain = to_dma_fence_chain(prev);
>> +        if (!prev_chain)
>> +            break;
>> +
>> +        /* No need for atomic operations since we hold the last
>> +         * reference to prev_chain.
>> +         */
>> +        chain->prev = prev_chain->prev;
>> +        RCU_INIT_POINTER(prev_chain->prev, NULL);
>> +        dma_fence_put(prev);
>> +    }
>> +    dma_fence_put(prev);
>>   -    dma_fence_put(rcu_dereference_protected(chain->prev, true));
>>       dma_fence_put(chain->fence);
>>       dma_fence_free(fence);
>>   }
>
>

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

  reply	other threads:[~2019-08-05 13:02 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20190805073657.1389-1-christian.koenig@amd.com>
2019-08-05  8:23 ` [PATCH] dma-buf: fix stack corruption in dma_fence_chain_release Lionel Landwerlin
2019-08-05 12:03 ` Lionel Landwerlin
2019-08-05 13:02   ` Christian König [this message]
2019-08-05 13:32     ` Lionel Landwerlin
2019-08-05 16:30       ` Daniel Vetter

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=af470fba-fb46-e0c4-caa0-220d6083ffc3@gmail.com \
    --to=ckoenig.leichtzumerken@gmail.com \
    --cc=christian.koenig@amd.com \
    --cc=david1.zhou@amd.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=lionel.g.landwerlin@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.