Should this scare me?

Started by Ian Hardingabout 23 years ago3 messagesgeneral
Jump to latest
#1Ian Harding
ianh@tpchd.org

psql in free(): warning: modified (chunk-) pointer.

I am used to psql dumping core from time to time, although since upgrading to 7.3.1 it has not happened.

Then this. Is it anything to fear?

NetBSD devil.tpchd.org 1.5.2 NetBSD 1.5.2 (planning) #0: Mon Jul 29 12:37:56 PDT 2002 iharding@devil:/usr/src/sys/arch/i386/compile/planning i386

PostgreSQL 7.3.1 on i386--netbsdelf, compiled by GCC egcs-1.1.2

#2Tom Lane
tgl@sss.pgh.pa.us
In reply to: Ian Harding (#1)
Re: Should this scare me?

"Ian Harding" <ianh@tpchd.org> writes:

psql in free(): warning: modified (chunk-) pointer.
I am used to psql dumping core from time to time, although since upgrading to 7.3.1 it has not happened.
Then this. Is it anything to fear?

It sure sounds like free() has detected something clobbering memory it
shouldn't --- at a guess, something overrunning a malloc'd buffer by a
few bytes. I'd recommend trying to trace the cause if you can. Is it
readily reproducible?

regards, tom lane

#3Ian Harding
ianh@tpchd.org
In reply to: Tom Lane (#2)
Re: Should this scare me?

It just happened for the first time yesterday... I will let you know if a pattern emerges... How would I trace it?

Ian Harding
Programmer/Analyst II
Tacoma-Pierce County Health Department
iharding@tpchd.org
(253) 798-3549

"If any man can convince me and bring home to me that I do not think or act aright, gladly will I change; for I search after truth, by which man never yet was harmed. But he is harmed who abideth on still in his deception and ignorance"

-- Marcus Aurelius

Tom Lane <tgl@sss.pgh.pa.us> 02/19/03 07:59PM >>>

"Ian Harding" <ianh@tpchd.org> writes:

psql in free(): warning: modified (chunk-) pointer.
I am used to psql dumping core from time to time, although since upgrading to 7.3.1 it has not happened.
Then this. Is it anything to fear?

It sure sounds like free() has detected something clobbering memory it
shouldn't --- at a guess, something overrunning a malloc'd buffer by a
few bytes. I'd recommend trying to trace the cause if you can. Is it
readily reproducible?

regards, tom lane