[parsec-users] Seg Fault in DEDUP - PARSEC

Yungang Bao ybao at CS.Princeton.EDU
Tue May 1 11:44:41 EDT 2012


Hi Akshatha,

Actually, the value of "buf =0xde654dc9dd3533aa" is absolutely erroneous. I suspect that there are some bugs in the lock-free queue such that the dequeued buf is wrong.

------------------------------------------------
#15 0x00000000004049a9 in Calc_SHA1Sig (
buf=0xde654dc9dd3533aa <Address 0xde654dc9dd3533aa out of bounds>, num_bytes=-1761803567, digest=0x7f8b24315f00 "") at util.c:15
#16 0x00000000004065cd in th_sub_ChunkProcess (chunk=...,
---Type <return> to continue, or q <return> to quit---
thread_id=2) at encoder.c:300
------------------------------------------------



Yungang

----- Original Message -----
From: "Akshatha mulki Bhat" <fdm546 at my.utsa.edu>
To: "PARSEC Users" <parsec-users at lists.cs.princeton.edu>
Sent: Tuesday, May 1, 2012 11:30:14 AM
Subject: Re: [parsec-users] Seg Fault in DEDUP - PARSEC


Hi Yungang, 


That was strange for me too..I am trying to investigate it myself.. 
I just replaced the queues in dedup to use lock free queue from the c++ concurrent building blocks.. 
I am now compiling dedup using g++ to incorporate the lock free queues, and in the process i am facing lot of seg faults.. 

Regards 
Akshatha 




On Tue, May 1, 2012 at 9:52 AM, Yungang Bao < ybao at cs.princeton.edu > wrote: 


Hi Akshatha, 

Since I don't know what you have changed, I have no idea what happened. But I'm curious of the value of the parameter “num_bytes = -1761803567” passed into Calc_SHA1Sig (). Why it is negative? 

Best Regards, 
Yungang 



----- Original Message ----- 
From: "Akshatha mulki Bhat" < fdm546 at my.utsa.edu > 
To: "PARSEC Users" < parsec-users at lists.cs.princeton.edu > 
Sent: Saturday, April 28, 2012 6:17:51 PM 
Subject: [parsec-users] Seg Fault in DEDUP - PARSEC 


Hi, 


I am using running some lock-free implementations for Dedup Data structures and after I made those changes I get a Segmentation fault for SHA1_Update function in the util.c file. Any idea why this set fault started popping up ? 
Here is a excerpt of the segfault extracted using gdb's core file: 



#0 0x00007f8b2f6b9b90 in ?? () 
from /lib/x86_64-linux-gnu/libcrypto.so.1.0.0 
#1 0x00007f8b24315bd0 in ?? () 
#2 0x00007f8b2fb91d60 in ?? () 
#3 0x00000000000033aa in ?? () 
#4 0x00007f8b2b4d0b50 in ?? () 
#5 0x00007f8b2b4d0b50 in ?? () 
#6 0x00007f8b2ea46b34 in mcount () 
from /lib/x86_64-linux-gnu/libc.so.6 
#7 0x00007f8b2b4d0b60 in ?? () 
#8 0x00007f8b2ea46b34 in mcount () 
from /lib/x86_64-linux-gnu/libc.so.6 
#9 0x00007f8b2b4d0b10 in ?? () 
#10 0x00007f8b04283ac0 in ?? () 
#11 0xde654dc97432366a in ?? () 
#12 0x00007f8b2b4d0b90 in ?? () 
#13 0x0000000000000011 in ?? () 
#14 0x00007f8b2f6b8690 in SHA1_Update () 
from /lib/x86_64-linux-gnu/libcrypto.so.1.0.0 
#15 0x00000000004049a9 in Calc_SHA1Sig ( 
buf=0xde654dc9dd3533aa <Address 0xde654dc9dd3533aa out of bounds>, num_bytes=-1761803567, digest=0x7f8b24315f00 "") at util.c:15 
#16 0x00000000004065cd in th_sub_ChunkProcess (chunk=..., 
---Type <return> to continue, or q <return> to quit--- 
thread_id=2) at encoder.c:300 
#17 0x0000000000406a5f in ChunkProcess (targs=0x7fff77e76770) 
at encoder.c:498 
#18 0x00007f8b2ed08efc in start_thread () 
from /lib/x86_64-linux-gnu/libpthread.so.0 
#19 0x00007f8b2ea4359d in clone () 
from /lib/x86_64-linux-gnu/libc.so.6 
#20 0x0000000000000000 in ?? () 




Please help me with this.. 


Thanks 
Akshatha 
_______________________________________________ 
parsec-users mailing list 
parsec-users at lists.cs.princeton.edu 
https://lists.cs.princeton.edu/mailman/listinfo/parsec-users 
_______________________________________________ 
parsec-users mailing list 
parsec-users at lists.cs.princeton.edu 
https://lists.cs.princeton.edu/mailman/listinfo/parsec-users 


_______________________________________________
parsec-users mailing list
parsec-users at lists.cs.princeton.edu
https://lists.cs.princeton.edu/mailman/listinfo/parsec-users


More information about the parsec-users mailing list