dev 1143 then wdt_reset udp server high rate packages (sdk 1.4)

[willemwouters]

Hi All.

I'm getting this error for a while now, but never find anything online about it.

When im using the ESP (SDK 1.4), and let it connect to my mobile phone (as AP).
Then running an UDP client on my mobile, and UDP server on ESP.

And I'm sending at a very high rate (while true loop) udp packages from mobile to esp.
But after a while(3min) it first prints "dev 1143" then after 8 seconds wdt_reset occurs.

Why is "dev 1143" and who is printing it?

[2015-10-09 19:38:31] dev 1143
[2015-10-09 19:38:39]
[2015-10-09 19:38:39] ets Jan 8 2013,rst cause:4, boot mode:(3,7)
[2015-10-09 19:38:39]
[2015-10-09 19:38:39] wdt reset
[2015-10-09 19:38:39] load 0x40100000, len 30688, room 16
[2015-10-09 19:38:39] tail 0
[2015-10-09 19:38:39] chksum 0xac
[2015-10-09 19:38:39] load 0x3ffe8000, len 5028, room 8
[2015-10-09 19:38:39] tail 12
[2015-10-09 19:38:39] chksum 0xf9
[2015-10-09 19:38:39] ho 0 tail 12 room 4
[2015-10-09 19:38:39] load 0x3ffe93a8, len 852, room 12
[2015-10-09 19:38:39] tail 8
[2015-10-09 19:38:39] chksum 0xea
[2015-10-09 19:38:39] csum 0xea


Thanks,

Willem

[Lieven]

Don't know if this helps. I got the same error dev 1143 using TCP connection/disconnect loop ( 2 sec interval )

dumping stack gives :
3FFFFD70:4000050C CODE _xtos_set_exception_handler +0xB8
3FFFFD80:40104311 CODE wDev_ProcessFiq +0x2E5
3FFFFD90:40104311 CODE wDev_ProcessFiq +0x2E5
3FFFFDE0:4020615A CODE ets_vsnprintf +0x1D6
3FFFFE00:40002514 CODE ets_printf +0x48
3FFFFE90:4000050C CODE _xtos_set_exception_handler +0xB8
3FFFFEB0:40000F68 CODE ets_post +0x144
3FFFFEC0:40000F58 CODE ets_post +0x134
3FFFFF20:40213CDA CODE esf_buf_recycle +0xD6
3FFFFF30:40211A88 CODE ppRecycleRxPkt +0x54
3FFFFF40:4021A92E CODE sta_input +0xD6
3FFFFF80:40211D42 CODE pp_tx_idle_timeout +0x36
3FFFFF90:402117E3 CODE ppPeocessRxPktHdr +0x277
3FFFFFB0:40000E19 CODE ets_run +0x15
3FFFFFC0:40001100 CODE ets_isr_unmask +0x158
3FFFFFD0:40100276 CODE call_user_start_local +0xE
3FFFFFE0:4010000D CODE call_user_start +0x9
3FFFFFF0:4000044C CODE UartDev +0x263C

This is not necessarily the exact trace. But disassembly shows :
40104302: fe9221 l32r a2, 40103d4c <trc_NeedRTS+0x22c>
40104305: fe9231 l32r a3, 40103d50 <trc_NeedRTS+0x230>
40104308: 77a442 movi a4, 0x477
4010430b: ef9d01 l32r a0, 40100180 <_UserExceptionVector_1+0x24>
4010430e: 0000c0 callx0 a0
40104311: ffff06 j 40104311 <wDev_ProcessFiq+0x2e5>

So a while(1); at 40104311

[ESP_Faye]

Hi,

Here is an example of TCP connection: http://bbs.espressif.com/viewtopic.php?f=31&t=232

If your problem is still unsolved, please provide your test code, we will help debug it.

Thanks for your interest in ESP8266 !

[willemwouters]

The example code doesnt really help at all.
I'm using udp, and using the lwip stack.

But @Lieven

I think your piece of disassembled is the exception handler of the crash.

Somewhere inside your trace I saw a esf_pbuf_recycle.... That sound like a possible suspect.
Going to try some with custom lwip stack and other pbuf settings in lwip...

Thanks

[ESP_Faye]

Hi willemwouters,

If your problem is still unsolved, please provide your test code, we will help debug it.

Thanks for your interest in ESP8266 !

[willemwouters]

Hi,

The problem can be reproduces by os_intr_lock() and os_intr_unlock()
But not holding the lock too long, and then in combination with lots of udp multicast packages.

I was driving a string of ws2812b led via bitbang. But after i switched to i2s implementation with dma all problems are gone.
Even increasing the packages more no problems found.

So for me its solved, i'm will keep my hands of the intr_lock in combination with wifi packages.

Thanks

[tomeko]

I'm seeing very similar restart randomly (after 5-20 minutes) when using ESP8266 as RTP streaming receiver. In my application RTP packets are coming to ESP8266 at 50pkts/s rate then buffered in kind of jitter buffer and passed to I2S "fake PWM" as in popular mp3 decoder project. Sometimes log line is different: "mac 674", but usually it's the same - obscure "dev 1143". Source code I'm using and PC application I'm using for testing is available at http://tomeko.net/projects/rtp_pager/

[ESP_Faye]

Hi,

Please do NOT disable the hardware interrupt for a long time, it should NOT be longer than 50us.

And you can have a try with our latest ESP8266_NONOS_SDK_V1.5.3 which has fixed the "mac 674" problem.

If your problem is still unsolved, please feel free to let us know.

Thanks for your interest in ESP8266 !

[milesstone]

I'm seeing very similar restart randomly (after 5-20 minutes) when using ESP8266 as RTP streaming receiver. In my GBWhatsApp application RTP packets are coming to ESP8266 at 50pkts/s rate then buffered in kind of jitter buffer and passed to I2S "fake PWM" as in popular mp3 decoder project. Sometimes log line is different: "mac 674", but usually it's the same - obscure "dev 1143". Source code I'm using and PC application I'm using for testing is available at http://tomeko.net/projects/rtp_pager/

Thank you so much for sharing this info which is really useful for me to solve my issue.

[Nielson]

I'm seeing very similar restart randomly (after 5-20 minutes) when using ESP8266 as RTP streaming receiver. In my GBWhatsApp application RTP packets are coming to ESP8266 at 50pkts/s rate then buffered in kind of jitter buffer and passed to I2S "fake PWM" as in popular mp3 decoder project. Sometimes log line is different: "mac 674", but usually it's the same - obscure "dev 1143". Source code I'm using and PC application I'm using for testing is available at http://tomeko.net/projects/rtp_pager/

Thank you so much for sharing this info which is really useful for me to solve my issue.

I think if you visit the official to solve this issue you will also solve it on official site as well.