Experimental Physics and
Industrial Control System

<[email protected]> · Fri, 28 Jul 2017 17:07:00 +0000

Hi Michael,

I put a breakpoint in camessage where "misssaligned protocol" is initially detected and if I step a few lines to delay that thread before continuing, I always just get the "force disconnect message" in the IOC and no assert; if I immediately continue from the breakpoint I always get an assert failure instead, though I've noticed that sometimes the assert backtrace is to read_reply() ... event_read() rather than to send_err() ... camessage() as shown below. Let me know if there is anything that you would like me to try/examine

Regards,

Freddie

-----Original Message-----
From: Michael Davidsaver [mailto:[email protected]] 
Sent: 27 July 2017 11:01
To: Akeroyd, Freddie (STFC,RAL,ISIS); [email protected]
Subject: Re: IOC error message on camonitor disconnect

This looks like it is related to:

https://bugs.launchpad.net/epics-base/+bug/541330

Although I suspect there is more than one bug involved.

On 07/26/2017 07:58 PM, [email protected] wrote:
> Apologies, forgot to include the stack trace:
> 
>  
> 
>                 dbCore.dll!cas_commit_msg(client * pClient, unsigned int
> size) Line 357               C
> 
>                dbCore.dll!vsend_err(const caHdrLargeArray * curp, int 
> status, client * client, const char * pformat, char * args) Line
> 245                C
> 
>                dbCore.dll!send_err(const caHdrLargeArray * curp, int 
> status, client * client, const char * pformat, ...) Line 264  C
> 
>                dbCore.dll!camessage(client * client) Line 2571 C
> 
>                 dbCore.dll!camsgtask(void * pParm) Line 129     C
> 
>                Com.dll!epicsWin32ThreadEntry(void * lpParameter) Line
> 501    C
> 
>  
> 
> Regards,
> 
>  
> 
> Freddie
> 
>  
> 
> *From:*[email protected]
> [mailto:[email protected]] *On Behalf Of 
> *[email protected]
> *Sent:* 26 July 2017 18:25
> *To:* [email protected]
> *Subject:* IOC error message on camonitor disconnect
> 
>  
> 
> Hi,
> 
>  
> 
> I am seeing this using EPICS base 3.15.5 running on windows x64. I 
> start softIOC.exe and then repeatedly connect to a PV with camonitor 
> and then CTRL-C the camonitor, each time I see either:
> 
>  
> 
> epics> CAS: request from 127.0.0.1:20272 => CAS: Missaligned protocol
> rejected
> 
> CAS: Request from 127.0.0.1:20272 => cmmd=26209 cid=0x150008 
> type=14592
> count=0 postsize=24885
> 
> CAS: Request from 127.0.0.1:20272 =>   available=0x0    N=1
> paddr=0000000000000000
> 
> CAS: forcing disconnect from 127.0.0.1:20272
> 
> epics>
> 
>  
> 
> or
> 
>  
> 
> A call to 'assert(size <= ntohs ( pMsg->m_postsize ))'
> 
>     by thread 'CAS-client' failed in 
> ../../../src/ioc/rsrv/caserverio.c
> line 357.
> 
> Dumping a stack trace of thread 'CAS-client':
> 
> [  000007FEEE71D565]
> 
> [  000007FEEE71D206]
> 
> [  000007FEEE714EF8]
> 
> [  000007FEED1A6AD5]
> 
> [  000007FEED1AC21B]
> 
> [  000007FEED1AC29B]
> 
> [  000007FEED1AAF1C]
> 
> [  000007FEED1AAAE5]
> 
> [  000007FEEE7177F1]
> 
> [  000007FECEB3B0A8]
> 
> [  000007FECEB3AD11]
> 
> [  00000000771459CD]
> 
> [  000000007727A561]
> 
> EPICS Release EPICS R3.15.5.
> 
> Local time is 2017-07-26 16:38:50.636232257 GMT Daylight Time
> 
> Please E-mail this message to the author or to [email protected] 
> <mailto:[email protected]>
> 
> Calling epicsThreadSuspendSelf()
> 
> Thread CAS-client (000000000281A040) suspended
> 
>  
> 
> It looks similar to the message reported in 
> http://www.aps.anl.gov/epics/tech-talk/2016/msg01828.php
> 
>  
> 
> If I break into the program at the point of the assert I see the 
> following details:
> 
>  
> 
> size = 56
> 
>  
> 
> *pMsg
> 
>    m_cmmd=256
> 
>    m_postsize=6144
> 
>    m_dataType=5120
> 
>    m_count = 256
> 
>    m_cid = 16777216
> 
>   m_available = 16777216
> 
>  
> 
> We only recently updated to 3.15.5 and I didn't remember seeing such 
> messages before, so I connected to a machine still running 3.14 and 
> confirmed that I don't get any errors reported for the same test on 
> 3.14.12.6.
> 
>  
> 
> Regards,
> 
>  
> 
> Freddie
> 
>  
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System