From LeonidS at riscogroup.com  Tue Jun  4 07:24:07 2024
From: LeonidS at riscogroup.com (Leonid Shigris)
Date: Tue, 4 Jun 2024 07:24:07 +0000
Subject: [opus] opus library issues
In-Reply-To: <AM9PR08MB7135698D82161C76C4D20D08D8F72@AM9PR08MB7135.eurprd08.prod.outlook.com>
References: <AM9PR08MB7135698D82161C76C4D20D08D8F72@AM9PR08MB7135.eurprd08.prod.outlook.com>
Message-ID: <DB9PR08MB7130873AEF65F4518F6618CFD8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>

Hello,


I integrated the latest library 1.5.2 into my system.

To compile the library have to allow VLA in IAR.

Running FreeRTOS on MCU NXP IMXRT1060.

I encoding the file using opus_demo.exe.

And for decoding I took the part of opus_demo.c and implemented on my MCU.

I experience a system crash from time to time by running the decoding.

I don?t see any memory overflow not in heap, not thread overflow.

Is there a situation that silk frame corruption causes to system crash?



Pls advise,

Kind Regards,
Leonid Shigris
RT Engineer

Email: LeonidS at riscogroup.com<mailto:LeonidS at riscogroup.com>


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xiph.org/pipermail/opus/attachments/20240604/ccbcbdd3/attachment.htm>

From tterribe at xiph.org  Tue Jun  4 10:16:38 2024
From: tterribe at xiph.org (Timothy B. Terriberry)
Date: Tue, 4 Jun 2024 03:16:38 -0700
Subject: [opus] opus library issues
In-Reply-To: <DB9PR08MB7130873AEF65F4518F6618CFD8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>
References: <AM9PR08MB7135698D82161C76C4D20D08D8F72@AM9PR08MB7135.eurprd08.prod.outlook.com>
 <DB9PR08MB7130873AEF65F4518F6618CFD8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>
Message-ID: <5b2a5b21-9f73-69a0-9089-4cc03763aa6d@xiph.org>

Leonid Shigris wrote:
> Is there a situation thatsilk frame corruption causes to system crash?

The library is able to decode corrupt frames without crashing.

From LeonidS at riscogroup.com  Tue Jun  4 10:27:53 2024
From: LeonidS at riscogroup.com (Leonid Shigris)
Date: Tue, 4 Jun 2024 10:27:53 +0000
Subject: [opus] opus library issues
In-Reply-To: <5b2a5b21-9f73-69a0-9089-4cc03763aa6d@xiph.org>
References: <AM9PR08MB7135698D82161C76C4D20D08D8F72@AM9PR08MB7135.eurprd08.prod.outlook.com>
 <DB9PR08MB7130873AEF65F4518F6618CFD8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>
 <5b2a5b21-9f73-69a0-9089-4cc03763aa6d@xiph.org>
Message-ID: <DB9PR08MB7130ED5E4406F40BDB6A66E6D8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>

Hi Timothy,

Because I don't see any memory overflow, remains only any memory corruption and for sure it happens inside the library.

Has anyone managed to run this library on embedded systems such as STM32, NXP IMXRT1060? By using opus_demo code (decoding part)?


Pls advise,

Kind Regards,
Leonid Shigris
RT Engineer

Email: LeonidS at riscogroup.com


-----Original Message-----
From: opus <opus-bounces at xiph.org> On Behalf Of Timothy B. Terriberry
Sent: Tuesday, 4 June 2024 13:17
To: opus at xiph.org
Subject: Re: [opus] opus library issues

[You don't often get email from tterribe at xiph.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]

Leonid Shigris wrote:
> Is there a situation thatsilk frame corruption causes to system crash?

The library is able to decode corrupt frames without crashing.
_______________________________________________
opus mailing list
opus at xiph.org
http://lists.xiph.org/mailman/listinfo/opus

From gmaxwell at gmail.com  Tue Jun  4 10:31:20 2024
From: gmaxwell at gmail.com (Greg Maxwell)
Date: Tue, 4 Jun 2024 10:31:20 +0000
Subject: [opus] opus library issues
In-Reply-To: <DB9PR08MB7130ED5E4406F40BDB6A66E6D8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>
References: <AM9PR08MB7135698D82161C76C4D20D08D8F72@AM9PR08MB7135.eurprd08.prod.outlook.com>
 <DB9PR08MB7130873AEF65F4518F6618CFD8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>
 <5b2a5b21-9f73-69a0-9089-4cc03763aa6d@xiph.org>
 <DB9PR08MB7130ED5E4406F40BDB6A66E6D8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>
Message-ID: <CAAS2fgTCo0L-HVLiJWUfC8qbYCM9tnJb2NFSfZZUCi4ecjH90w@mail.gmail.com>

Many people have run opus on small embedded devices.  It may be the
case that your VLA support isn't providing enough stack.  You may want
to try using the pseudostack mode which will instead put that memory
on the heap and give you better control over making sure you have
enough available.

On Tue, Jun 4, 2024 at 10:28?AM Leonid Shigris <LeonidS at riscogroup.com> wrote:
>
> Hi Timothy,
>
> Because I don't see any memory overflow, remains only any memory corruption and for sure it happens inside the library.
>
> Has anyone managed to run this library on embedded systems such as STM32, NXP IMXRT1060? By using opus_demo code (decoding part)?
>
>
> Pls advise,
>
> Kind Regards,
> Leonid Shigris
> RT Engineer
>
> Email: LeonidS at riscogroup.com
>
>
> -----Original Message-----
> From: opus <opus-bounces at xiph.org> On Behalf Of Timothy B. Terriberry
> Sent: Tuesday, 4 June 2024 13:17
> To: opus at xiph.org
> Subject: Re: [opus] opus library issues
>
> [You don't often get email from tterribe at xiph.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
>
> Leonid Shigris wrote:
> > Is there a situation thatsilk frame corruption causes to system crash?
>
> The library is able to decode corrupt frames without crashing.
> _______________________________________________
> opus mailing list
> opus at xiph.org
> http://lists.xiph.org/mailman/listinfo/opus
> _______________________________________________
> opus mailing list
> opus at xiph.org
> http://lists.xiph.org/mailman/listinfo/opus

From LeonidS at riscogroup.com  Tue Jun  4 10:40:28 2024
From: LeonidS at riscogroup.com (Leonid Shigris)
Date: Tue, 4 Jun 2024 10:40:28 +0000
Subject: [opus] opus library issues
In-Reply-To: <CAAS2fgTCo0L-HVLiJWUfC8qbYCM9tnJb2NFSfZZUCi4ecjH90w@mail.gmail.com>
References: <AM9PR08MB7135698D82161C76C4D20D08D8F72@AM9PR08MB7135.eurprd08.prod.outlook.com>
 <DB9PR08MB7130873AEF65F4518F6618CFD8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>
 <5b2a5b21-9f73-69a0-9089-4cc03763aa6d@xiph.org>
 <DB9PR08MB7130ED5E4406F40BDB6A66E6D8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>
 <CAAS2fgTCo0L-HVLiJWUfC8qbYCM9tnJb2NFSfZZUCi4ecjH90w@mail.gmail.com>
Message-ID: <DB9PR08MB7130A7A6121E0B97EB37ACF9D8F82@DB9PR08MB7130.eurprd08.prod.outlook.com>

Hi Greg,

Attached my opus config.

As you suggested not to use VLA, what changes are required in configurations?



Pls advise,

Kind Regards, 
Leonid Shigris
RT Engineer

Email: LeonidS at riscogroup.com

-----Original Message-----
From: Greg Maxwell <gmaxwell at gmail.com> 
Sent: Tuesday, 4 June 2024 13:31
To: Leonid Shigris <LeonidS at riscogroup.com>
Cc: Timothy B. Terriberry <tterribe at xiph.org>; opus at xiph.org
Subject: Re: [opus] opus library issues

[You don't often get email from gmaxwell at gmail.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]

Many people have run opus on small embedded devices.  It may be the case that your VLA support isn't providing enough stack.  You may want to try using the pseudostack mode which will instead put that memory on the heap and give you better control over making sure you have enough available.

On Tue, Jun 4, 2024 at 10:28?AM Leonid Shigris <LeonidS at riscogroup.com> wrote:
>
> Hi Timothy,
>
> Because I don't see any memory overflow, remains only any memory corruption and for sure it happens inside the library.
>
> Has anyone managed to run this library on embedded systems such as STM32, NXP IMXRT1060? By using opus_demo code (decoding part)?
>
>
> Pls advise,
>
> Kind Regards,
> Leonid Shigris
> RT Engineer
>
> Email: LeonidS at riscogroup.com
>
>
> -----Original Message-----
> From: opus <opus-bounces at xiph.org> On Behalf Of Timothy B. Terriberry
> Sent: Tuesday, 4 June 2024 13:17
> To: opus at xiph.org
> Subject: Re: [opus] opus library issues
>
> [You don't often get email from tterribe at xiph.org. Learn why this is 
> important at https://aka.ms/LearnAboutSenderIdentification ]
>
> Leonid Shigris wrote:
> > Is there a situation thatsilk frame corruption causes to system crash?
>
> The library is able to decode corrupt frames without crashing.
> _______________________________________________
> opus mailing list
> opus at xiph.org
> https://eur02.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists
> .xiph.org%2Fmailman%2Flistinfo%2Fopus&data=05%7C02%7CLeonidS%40riscogr
> oup.com%7C865f927dc4184eb1c1f708dc8481825b%7Ce76b724084814694bfc77d157
> 2a40711%7C0%7C0%7C638530938991071671%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiM
> C4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C60000%7C%7
> C%7C&sdata=LIX9b0PyWoo5GtwmdYRf2BRqO3rb4ilbeihLsAZe3Ng%3D&reserved=0
> _______________________________________________
> opus mailing list
> opus at xiph.org
> https://eur02.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists
> .xiph.org%2Fmailman%2Flistinfo%2Fopus&data=05%7C02%7CLeonidS%40riscogr
> oup.com%7C865f927dc4184eb1c1f708dc8481825b%7Ce76b724084814694bfc77d157
> 2a40711%7C0%7C0%7C638530938991082353%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiM
> C4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C60000%7C%7
> C%7C&sdata=mgvz4LvaLKJo2aUdgsLzHuQ4mYlGooPLV6Ue8hyVXjc%3D&reserved=0
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: opus_config.h
URL: <http://lists.xiph.org/pipermail/opus/attachments/20240604/d4708c5a/attachment-0002.h>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: stack_alloc.h
URL: <http://lists.xiph.org/pipermail/opus/attachments/20240604/d4708c5a/attachment-0003.h>