[cvsnt-dev] Re: bug in lockserver (unix build)

Richard Wirth r.wirth at wirthware.de
Wed Feb 2 08:13:44 GMT 2005

Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.

Hello Tony,

Friday, January 28, 2005, 5:53:41 AM, you wrote:

TH> Richard Wirth wrote:
>> Hello Tony,
>>   there is a bug in cvslockd:
>>   It consumes 99% of cpu time.
>>   in CSocketIO::accept() the select() statement seams to be buggy...

TH> It's working fine here...

TH> top - 04:52:48 up 2 days,  7:42,  1 user,  load average: 0.05, 0.15, 0.42
TH> Tasks:   1 total,   0 running,   1 sleeping,   0 stopped,   0 zombie
TH> Cpu(s):  0.0% us,  0.0% sy,  0.0% ni,  0.0% id,  0.0% wa,  0.0% hi,  0.0% si
TH> Mem:    906796k total,   881588k used,    25208k free,    13560k buffers
TH> Swap:  1004052k total,    91268k used,   912784k free,   374864k cached

TH> 22415 tmh       15   0 12680 1740 2872 S  0.0  0.2   0:00.14 cvslockd

TH> What OS is having problems?

System is linux 2.6.8-24.11 x86_64

So I've found a workaround for the problem:

When sockets are only created for AF_INET or AF_INET6, all is going
well. The problem only occures, when the listener sockets are created
for both AF_INET and AF_INET6. Than accept() returns -1 and errno is
22 (INVALID PARAMETER) for the second listener socket created.

It looks like the system can not listen on both AFs at the same time!
May be it is mapped internaly between the AFs?

Best regards,
 Richard                            mailto:r.wirth at wirthware.de

More information about the cvsnt-dev mailing list