<html><body><div style="color:#000; background-color:#fff; font-family:HelveticaNeue, Helvetica Neue, Helvetica, Arial, Lucida Grande, sans-serif;font-size:16px"><div dir="ltr" id="yui_3_16_0_1_1429210732892_23686"><span>Hi Suresh ,</span></div><div dir="ltr" id="yui_3_16_0_1_1429210732892_23687"><span><br></span></div><div dir="ltr" id="yui_3_16_0_1_1429210732892_23687">Can you reproduce this ?</div><div dir="ltr" id="yui_3_16_0_1_1429210732892_23688"><span id="yui_3_16_0_1_1429210732892_23690">What are you using for transcoding ?</span></div><div dir="ltr" id="yui_3_16_0_1_1429210732892_23688"><span id="yui_3_16_0_1_1429210732892_23843">Perhaps the bug is in your transcoding software.</span></div><div dir="ltr" id="yui_3_16_0_1_1429210732892_23688"><span id="yui_3_16_0_1_1429210732892_23844">You can use wireshark to extract the contents of your RTP packets in a separate file.</span></div><div dir="ltr" id="yui_3_16_0_1_1429210732892_23688"><br></div><div dir="ltr" id="yui_3_16_0_1_1429210732892_23651"><span>Regards,</span></div><div dir="ltr" id="yui_3_16_0_1_1429210732892_23652"><span>Dragos </span></div><br> <div style="font-family: HelveticaNeue, Helvetica Neue, Helvetica, Arial, Lucida Grande, sans-serif; font-size: 16px;" id="yui_3_16_0_1_1429210732892_23625"> <div style="font-family: HelveticaNeue, Helvetica Neue, Helvetica, Arial, Lucida Grande, sans-serif; font-size: 16px;" id="yui_3_16_0_1_1429210732892_23624"> <div dir="ltr" id="yui_3_16_0_1_1429210732892_23623"> <hr size="1"> <font size="2" face="Arial" id="yui_3_16_0_1_1429210732892_23626"> <b><span style="font-weight:bold;">From:</span></b> Suresh Thiriveedi <sthiriveedi@gmail.com><br> <b><span style="font-weight: bold;">To:</span></b> Jean-Marc Valin <jmvalin@jmvalin.ca> <br><b><span style="font-weight: bold;">Cc:</span></b> opus@xiph.org <br> <b><span style="font-weight: bold;">Sent:</span></b> Thursday, April 16, 2015 2:32 PM<br> <b><span style="font-weight: bold;">Subject:</span></b> Re: [opus] Availability of the 1.1.1 stable version<br> </font> </div> <div class="y_msg_container" id="yui_3_16_0_1_1429210732892_23656"><br><div id="yiv0654384112"><div id="yui_3_16_0_1_1429210732892_23655"><div dir="ltr" id="yui_3_16_0_1_1429210732892_23654"><div class="yiv0654384112gmail_default" style="color:#000000;" id="yui_3_16_0_1_1429210732892_23653">This is observed on a live call between webRTC browser client and another legacy client. Our server is there in between and transcoding from opus to another codec and this is observed while decoding the opus.</div><div class="yiv0654384112gmail_default" style="color:#000000;" id="yui_3_16_0_1_1429210732892_23657"><br clear="none"></div><div class="yiv0654384112gmail_default" style="color:#000000;" id="yui_3_16_0_1_1429210732892_23896">Anyway, I'll try to capture/dump the packets in the server before feeding to the opus_decode and share with you. But this will not have the first 8 bytes (length+enc range) to directly feed to the sample binary. Please let me know if this is fine.</div><div class="yiv0654384112gmail_default" style="color:#000000;" id="yui_3_16_0_1_1429210732892_23897"><br clear="none"></div><div class="yiv0654384112gmail_default" style="color:#000000;" id="yui_3_16_0_1_1429210732892_23898">Thanks</div><div class="yiv0654384112gmail_default" style="color:#000000;" id="yui_3_16_0_1_1429210732892_23899">Suresh</div></div><div class="yiv0654384112gmail_extra" id="yui_3_16_0_1_1429210732892_23666"><br clear="none"><div class="qtdSeparateBR"><br><br></div><div class="yiv0654384112yqt1762627306" id="yiv0654384112yqtfd09456"><div class="yiv0654384112gmail_quote" id="yui_3_16_0_1_1429210732892_23665">On 16 April 2015 at 17:36, Jean-Marc Valin <span dir="ltr"><<a rel="nofollow" shape="rect" ymailto="mailto:jmvalin@jmvalin.ca" target="_blank" href="mailto:jmvalin@jmvalin.ca">jmvalin@jmvalin.ca</a>></span> wrote:<br clear="none"><blockquote class="yiv0654384112gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;" id="yui_3_16_0_1_1429210732892_23664">Please provide the input file that produces this with opus_demo.<br clear="none">
<span class="yiv0654384112" id="yui_3_16_0_1_1429210732892_23900"><br clear="none">
On 16/04/15 03:24 AM, Suresh Thiriveedi wrote:<br clear="none">
> Hi Jean-Marc,<br clear="none">
><br clear="none">
> Could you please update if you got a chance to look into. As I<br clear="none">
> mentioned, I don't see the same issue in 1.1.1, but I don't see any<br clear="none">
> difference in 1.1.1 other than optimization based on the architecture.<br clear="none">
> This optimization could have fixed some stack overflow issue in some<br clear="none">
> specific cases?<br clear="none">
><br clear="none">
><br clear="none">
> Thanks<br clear="none">
> Suresh<br clear="none">
><br clear="none">
> On 13 April 2015 at 12:39, Suresh Thiriveedi <<a rel="nofollow" shape="rect" ymailto="mailto:sthiriveedi@gmail.com" target="_blank" href="mailto:sthiriveedi@gmail.com">sthiriveedi@gmail.com</a><br clear="none">
</span><span class="yiv0654384112">> <mailto:<a rel="nofollow" shape="rect" ymailto="mailto:sthiriveedi@gmail.com" target="_blank" href="mailto:sthiriveedi@gmail.com">sthiriveedi@gmail.com</a>>> wrote:<br clear="none">
><br clear="none">
> Hi Jean-Marc,<br clear="none">
><br clear="none">
> Thanks for your response. Please find the details as below.<br clear="none">
><br clear="none">
</span>> *_Backtrace we got for this crash:_*<br clear="none">
<div id="yui_3_16_0_1_1429210732892_23663"><div class="yiv0654384112h5" id="yui_3_16_0_1_1429210732892_23662">><br clear="none">
> #0 0x0000000000800c54 in opus_decode_frame (st=0x38906b8f99d09c5,<br clear="none">
><br clear="none">
> data=0xf0aa10b4ef1008ae <Address 0xf0aa10b4ef1008ae out of<br clear="none">
> bounds>, len=-188613428, pcm=0x6e80016085efd57,<br clear="none">
><br clear="none">
> frame_size=44037315, decode_fec=58716895) at src/opus_decoder.c:384<br clear="none">
><br clear="none">
><br clear="none">
> #1 0x00000000008009c0 in opus_decode_frame (st=0x712357d0,<br clear="none">
><br clear="none">
> data=0x7effff9ab72d<br clear="none">
> "~▒`\\▒K\005▒▒y▒w+g~▒S2\025▒\036T▒\002x▒▒h!▒▒▒\220\233\066s▒\030#gb<br clear="none">
> \rn▒rF\005Q▒\213;▒`\207$O▒(m\222=9▒▒/h▒▒t▒▒E묳w▒\237\"\206z\005<br clear="none">
> \213»u@e", len=88, pcm=0x7effff9a6a80, frame_size=640, decode_fec=0)<br clear="none">
> at src/opus_decoder.c:319<br clear="none">
><br clear="none">
><br clear="none">
> #2 0x0000000000801be1 in opus_decode_native (st=0x712357d0,<br clear="none">
><br clear="none">
> data=0x7effff9ab72d<br clear="none">
> "~▒`\\▒K\005▒▒y▒w+g~▒S2\025▒\036T▒\002x▒▒h!▒▒▒\220\233\066s▒\030#gb<br clear="none">
> \rn▒rF\005Q▒\213;▒`\207$O▒(m\222=9▒▒/h▒▒t▒▒E묳w▒\237\"\206z\005<br clear="none">
> \213»u@e", len=89, pcm=0x7effff9a6a80, frame_size=640, decode_fec=0,<br clear="none">
> self_delimited=0,<br clear="none">
><br clear="none">
> packet_offset=0x0, soft_clip=1) at src/opus_decoder.c:681<br clear="none">
><br clear="none">
><br clear="none">
> #3 0x000000000080226c in opus_decode (st=0x712357d0,<br clear="none">
><br clear="none">
> data=0x7effff9ab72c<br clear="none">
> "▒~▒`\\▒K\005▒▒y▒w+g~▒S2\025▒\036T▒\002x▒▒h!▒▒▒\220\233\066s▒\030#gb<br clear="none">
> \rn▒rF\005Q▒\213;▒`\207$O▒(m\222=9▒▒/h▒▒t▒▒E묳w▒\237\"\206z\005<br clear="none">
> \213»u@e", len=89, pcm=0x71245a60, frame_size=640, decode_fec=0) at<br clear="none">
> src/opus_decoder.c:867<br clear="none">
><br clear="none">
><br clear="none">
> #4 0x00000000004fd6b5 in kn_opus_decode (decHandle=0x712357d0,<br clear="none">
> decProp=0x1675698, src=0x16756d0, dest=0x71245a60,<br clear="none">
><br clear="none">
> dstLen=0x1673210) at MSTranscodeOPUS.c:100<br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
</div></div>> *_And the code flow what we have observed for this specific incident:_*<br clear="none">
> *_<br clear="none">
> _*<br clear="none">
> *_Called this as mode is CELT_ONLY,_*<br clear="none">
<span class="yiv0654384112">><br clear="none">
> if (data!=NULL && st->prev_mode > 0 && (<br clear="none">
> (mode == MODE_CELT_ONLY && st->prev_mode != MODE_CELT_ONLY &&<br clear="none">
> !st->prev_redundancy)<br clear="none">
> || (mode != MODE_CELT_ONLY && st->prev_mode == MODE_CELT_ONLY) )<br clear="none">
> )<br clear="none">
> {<br clear="none">
</span>> _transition = 1_;<br clear="none">
<span class="yiv0654384112">> /* Decide where to allocate the stack memory for pcm_transition */<br clear="none">
> if (mode == MODE_CELT_ONLY)<br clear="none">
> pcm_transition_celt_size = F5*st->channels;<br clear="none">
> else<br clear="none">
> pcm_transition_silk_size = F5*st->channels;<br clear="none">
> }<br clear="none">
><br clear="none">
</span>> *_So transition is made as 1 called this,_*<br clear="none">
<span class="yiv0654384112">><br clear="none">
> if (transition && mode == MODE_CELT_ONLY)<br clear="none">
> {<br clear="none">
> pcm_transition = pcm_transition_celt;<br clear="none">
> opus_decode_frame(st, NULL, 0, pcm_transition, IMIN(F5,<br clear="none">
> audiosize), 0);<br clear="none">
> }<br clear="none">
><br clear="none">
</span>> *_In "opus_decode_frame" again, as data is passed as NULL, goes to<br clear="none">
> else part_*<br clear="none">
<span class="yiv0654384112">><br clear="none">
> if (data != NULL)<br clear="none">
> {<br clear="none">
> audiosize = st->frame_size;<br clear="none">
> mode = st->mode;<br clear="none">
> ec_dec_init(&dec,(unsigned char*)data,len);<br clear="none">
> } else {<br clear="none">
> audiosize = frame_size;<br clear="none">
> mode = st->prev_mode;<br clear="none">
><br clear="none">
</span>> *_As the mode is made as prev mode now, which was a silk, this goes<br clear="none">
> inside,_*<br clear="none">
<span class="yiv0654384112">><br clear="none">
> /* SILK processing */<br clear="none">
> if (mode != MODE_CELT_ONLY)<br clear="none">
> {<br clear="none">
><br clear="none">
</span>> *_Then in this function called this_*,<br clear="none">
<span class="yiv0654384112">><br clear="none">
> silk_ret = silk_Decode( silk_dec, &st->DecControl,<br clear="none">
> lost_flag, first_frame, &dec,<br clear="none">
> pcm_ptr, &silk_frame_size );<br clear="none">
><br clear="none">
><br clear="none">
</span>> *_And finally, somehow, the "silk_frame_size" is a negative value (<br clear="none">
<span class="yiv0654384112">> say -1376272 in our case), then in the same function called the<br clear="none">
</span>> below and this crashes here._*<br clear="none">
<span class="yiv0654384112">><br clear="none">
> pcm_ptr += silk_frame_size * st->channels;<br clear="none">
><br clear="none">
><br clear="none">
> Please help.<br clear="none">
><br clear="none">
> Thanks<br clear="none">
> Suresh<br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
> On 12 April 2015 at 21:23, Jean-Marc Valin <<a rel="nofollow" shape="rect" ymailto="mailto:jmvalin@jmvalin.ca" target="_blank" href="mailto:jmvalin@jmvalin.ca">jmvalin@jmvalin.ca</a><br clear="none">
</span><span class="yiv0654384112">> <mailto:<a rel="nofollow" shape="rect" ymailto="mailto:jmvalin@jmvalin.ca" target="_blank" href="mailto:jmvalin@jmvalin.ca">jmvalin@jmvalin.ca</a>>> wrote:<br clear="none">
><br clear="none">
> Do you have any file that demonstrates the problem with either<br clear="none">
> opus_demo<br clear="none">
> or opusdec?<br clear="none">
><br clear="none">
> Jean-Marc<br clear="none">
><br clear="none">
> On 09/04/15 04:01 AM, Suresh Thiriveedi wrote:<br clear="none">
> > Hi,<br clear="none">
> ><br clear="none">
> > I'm curious to know when would be the 1.1.1 stable version<br clear="none">
> available.<br clear="none">
> ><br clear="none">
> > In 1.1, we are facing crash when opus library is trying to<br clear="none">
> decode the<br clear="none">
> > CELT-only, full band and 20 ms. So we tried with 1.1.1 beta<br clear="none">
> and it looks<br clear="none">
> > to be fine. Is there any open issue regarding this in 1.1 version?<br clear="none">
> ><br clear="none">
> > Thanks<br clear="none">
> > Suresh<br clear="none">
> ><br clear="none">
> ><br clear="none">
> > _______________________________________________<br clear="none">
> > opus mailing list<br clear="none">
</span>> > <a rel="nofollow" shape="rect" ymailto="mailto:opus@xiph.org" target="_blank" href="mailto:opus@xiph.org">opus@xiph.org</a> <mailto:<a rel="nofollow" shape="rect" ymailto="mailto:opus@xiph.org" target="_blank" href="mailto:opus@xiph.org">opus@xiph.org</a>><br clear="none">
> > <a rel="nofollow" shape="rect" target="_blank" href="http://lists.xiph.org/mailman/listinfo/opus">http://lists.xiph.org/mailman/listinfo/opus</a><br clear="none">
> ><br clear="none">
><br clear="none">
><br clear="none">
><br clear="none">
</blockquote></div><br clear="none"></div></div></div></div><br><div class="yqt1762627306" id="yqtfd53447">_______________________________________________<br clear="none">opus mailing list<br clear="none"><a shape="rect" ymailto="mailto:opus@xiph.org" href="mailto:opus@xiph.org">opus@xiph.org</a><br clear="none"><a shape="rect" href="http://lists.xiph.org/mailman/listinfo/opus" target="_blank">http://lists.xiph.org/mailman/listinfo/opus</a><br clear="none"></div><br><br></div> </div> </div> </div></body></html>