[FFmpeg-devel] Revert "avcodec/alsdec: check predictor order against block length"

Submitted by Paul B Mahol on June 30, 2017, 3:31 p.m.

Details

Message ID 20170630153148.2342-1-onemda@gmail.com
State New
Headers show

Commit Message

Paul B Mahol June 30, 2017, 3:31 p.m.
This reverts commit 18f94df8af04f2c02a25a7dec512289feff6517f.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
---
 libavcodec/alsdec.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

Comments

Derek Buitenhuis June 30, 2017, 3:36 p.m.
On 6/30/2017 4:31 PM, Paul B Mahol wrote:
> This reverts commit 18f94df8af04f2c02a25a7dec512289feff6517f.
> 
> Signed-off-by: Paul B Mahol <onemda@gmail.com>
> ---
>  libavcodec/alsdec.c | 6 +-----

http://ffmpeg.org/pipermail/ffmpeg-devel/2016-November/202513.html

- Derek
Thilo Borgmann June 30, 2017, 3:39 p.m.
Hi,

Am 30.06.17 um 17:31 schrieb Paul B Mahol:
> This reverts commit 18f94df8af04f2c02a25a7dec512289feff6517f.
> 
> Signed-off-by: Paul B Mahol <onemda@gmail.com>
> ---
>  libavcodec/alsdec.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)

this has not yet been reverted for a reason and I veto to apply this.

I can make no promises but this is on my todo-list. For the time being, having
this not applied is the smaller evil, so to speak.

-Thilo
Paul B Mahol June 30, 2017, 3:42 p.m.
On 6/30/17, Thilo Borgmann <thilo.borgmann@mail.de> wrote:
> Hi,
>
> Am 30.06.17 um 17:31 schrieb Paul B Mahol:
>> This reverts commit 18f94df8af04f2c02a25a7dec512289feff6517f.
>>
>> Signed-off-by: Paul B Mahol <onemda@gmail.com>
>> ---
>>  libavcodec/alsdec.c | 6 +-----
>>  1 file changed, 1 insertion(+), 5 deletions(-)
>
> this has not yet been reverted for a reason and I veto to apply this.
>
> I can make no promises but this is on my todo-list. For the time being,
> having
> this not applied is the smaller evil, so to speak.

Do you have sample(s) which are causing crash?
Thilo Borgmann June 30, 2017, 3:44 p.m.
Am 30.06.17 um 17:42 schrieb Paul B Mahol:
> On 6/30/17, Thilo Borgmann <thilo.borgmann@mail.de> wrote:
>> Hi,
>>
>> Am 30.06.17 um 17:31 schrieb Paul B Mahol:
>>> This reverts commit 18f94df8af04f2c02a25a7dec512289feff6517f.
>>>
>>> Signed-off-by: Paul B Mahol <onemda@gmail.com>
>>> ---
>>>  libavcodec/alsdec.c | 6 +-----
>>>  1 file changed, 1 insertion(+), 5 deletions(-)
>>
>> this has not yet been reverted for a reason and I veto to apply this.
>>
>> I can make no promises but this is on my todo-list. For the time being,
>> having
>> this not applied is the smaller evil, so to speak.
> 
> Do you have sample(s) which are causing crash?

Yes but I cannot share, unfortunately.

-Thilo
Derek Buitenhuis June 30, 2017, 3:46 p.m.
On 6/30/2017 4:44 PM, Thilo Borgmann wrote:
> Yes but I cannot share, unfortunately.

There are  two backtraces in the mail thread from before that could be helpful.

- Derek
Paul B Mahol June 30, 2017, 3:58 p.m.
On 6/30/17, Derek Buitenhuis <derek.buitenhuis@gmail.com> wrote:
> On 6/30/2017 4:44 PM, Thilo Borgmann wrote:
>> Yes but I cannot share, unfortunately.
>
> There are  two backtraces in the mail thread from before that could be
> helpful.

Thanks, but I do not see important variables.
Paul B Mahol June 30, 2017, 3:58 p.m.
On 6/30/17, Thilo Borgmann <thilo.borgmann@mail.de> wrote:
> Am 30.06.17 um 17:42 schrieb Paul B Mahol:
>> On 6/30/17, Thilo Borgmann <thilo.borgmann@mail.de> wrote:
>>> Hi,
>>>
>>> Am 30.06.17 um 17:31 schrieb Paul B Mahol:
>>>> This reverts commit 18f94df8af04f2c02a25a7dec512289feff6517f.
>>>>
>>>> Signed-off-by: Paul B Mahol <onemda@gmail.com>
>>>> ---
>>>>  libavcodec/alsdec.c | 6 +-----
>>>>  1 file changed, 1 insertion(+), 5 deletions(-)
>>>
>>> this has not yet been reverted for a reason and I veto to apply this.
>>>
>>> I can make no promises but this is on my todo-list. For the time being,
>>> having
>>> this not applied is the smaller evil, so to speak.
>>
>> Do you have sample(s) which are causing crash?
>
> Yes but I cannot share, unfortunately.
>

Thanks. Unfortunately bug will remain open for next few decades I see.
Derek Buitenhuis June 30, 2017, 4:04 p.m.
On 6/30/2017 4:58 PM, Paul B Mahol wrote:
> Thanks, but I do not see important variables.

There are line numbers and dates, though, pointing at:

931:    *raw_samples++ -= y >> 20;

Better than nothing.

- Derek

Patch hide | download patch | download mbox

diff --git a/libavcodec/alsdec.c b/libavcodec/alsdec.c
index d95e30d..209feb8 100644
--- a/libavcodec/alsdec.c
+++ b/libavcodec/alsdec.c
@@ -705,11 +705,7 @@  static int read_var_block_data(ALSDecContext *ctx, ALSBlockData *bd)
         } else {
             *bd->opt_order = sconf->max_order;
         }
-        if (*bd->opt_order > bd->block_length) {
-            *bd->opt_order = bd->block_length;
-            av_log(avctx, AV_LOG_ERROR, "Predictor order too large.\n");
-            return AVERROR_INVALIDDATA;
-        }
+
         opt_order = *bd->opt_order;
 
         if (opt_order) {