< Previous by Date	Date Index	Next by Date >
< Previous in Thread	Thread Index	Next in Thread >

RE: [reSIProcate] AbstractFifo bug?

From: "Micky Kaufmann" <micky@xxxxxxxxxxx>
Date: Sun, 22 May 2005 11:15:45 +0200

I’m working on windows, visual studio 7.1 (Maybe the problem is with the Mutex being recursive?)

Still what troubles me in the AbstractFifo is the fact that the while condition is ‘ while (mFifo.empty()) ’ but the assert is not on mFifo. So it seems that the while condition should be ‘ while (mSize!=0) ’.

The bug appears when a client tries to reach a non registered contact - the server tries to skip remaining stages of a transaction and terminate it but it fails...

Here’s the stack trace:

> repro.exe!_NMSG_WRITE(int rterrnum=10) Line 195 C

repro.exe!abort() Line 44 + 0x7 C

repro.exe!_assert(const char * expr=0x00a652b0, const char * filename=0x00a65298, unsigned int lineno=57) Line 306 C

repro.exe!resip::AbstractFifo::getNext(int ms=100) Line 57 + 0x1a C++

repro.exe!resip::TimeLimitFifo<resip::Message>::getNext(int ms=100) Line 143 + 0xc C++

repro.exe!repro::Proxy::thread() Line 67 + 0xd C++

repro.exe!threadWrapper(void * threadParm=0x000e06dc) Line 34 + 0xd C++

kernel32.dll!7c80b50b()

ntdll.dll!7c96e0d4()

kernel32.dll!7c8399f3()

And here is the debug info produces by repro:

Assertion failed: mSize != 0, file .\os\AbstractFifo.cxx, line 57

From: Derek MacDonald [mailto:derek@xxxxxxxx]
Sent: ה 19 מאי 2005 20:20
To: Micky Kaufmann; resiprocate-devel@xxxxxxxxxxxxxxxxxxx
Subject: RE: [reSIProcate] AbstractFifo bug?

Can you get a stack trace this? AbstractFifo is used a lot in resip, but it does have several subclasses. The size is incremeneted in the add methods of the subclasses, which lock before adding.

Also, run resiprocate/test/testFifo.cxx on your configuration.

--Derek

PS -- What is your platform/compiler?

CONFIDENTIALITY NOTICE

This email and any files transmitted with it contains proprietary information and, unless expressly stated otherwise, all contents and attachments are confidential. This email is intended for the addressee(s) only and access by anyone else is unauthorized. If you are not an addressee, any disclosure, distribution, printing or copying of the contents of this email or its attachments, or any action taken in reliance on it, is unauthorized and may be unlawful. If you are not an addressee, please inform the sender immediately and then delete this email and any copies of it. Thank you for your co-operation.

From: resiprocate-devel-bounces@xxxxxxxxxxxxxxxxxxx [mailto:resiprocate-devel-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Micky Kaufmann
Sent: Thursday, May 19, 2005 6:31 AM
To: resiprocate-devel@xxxxxxxxxxxxxxxxxxx
Subject: [reSIProcate] AbstractFifo bug?

Hi. All

While running repro I go an assert error from the AbstractFifo::getNext method:

void* AbstractFifo::getNext() {

Lock lock(mMutex); (void)lock;

while (mFifo.empty())

{

mCondition.wait(mMutex);

}

void* firstMessage = mFifo.front();

mFifo.pop_front();

assert(mSize != 0);

mSize--;

return firstMessage;

}

As I see it the problem is caused probably because non thread safe adding of elements to the Fifo. Meaning mSize++ is missing somewhere and from what I investigated the only reason for using mSize instead of using std::deque::size is because we want to make sure that mSize < mMaxSize.

Is there another reason?

Is it ok to replace the lines:

assert(mSize != 0);

mSize--;

with

mSize = mFifo.size();

Is it ok to replace the line:

while (mFifo.empty())

with

while (mSize==0) //this.empty()

Any other solution?

Prev by Date: [reSIProcate] How to release ressource used by the resiprocated library
Next by Date: [reSIProcate] Re: ReSIProcateInfo
Previous by thread: RE: [reSIProcate] AbstractFifo bug?
Next by thread: RE: [reSIProcate] AbstractFifo bug?
Index(es):
- Date
- Thread