Date: Wed, 19 Jun 2002 10:21:07 -0700 From: Alfred Perlstein <bright@mu.org> To: Don Lewis <dl-freebsd@catspoiler.org> Cc: jhb@freebsd.org, hackers@freebsd.org Subject: Re: Suggested fixes for uidinfo "would sleep" messages Message-ID: <20020619172107.GF85935@elvis.mu.org> In-Reply-To: <200206190810.g5J8AKM1065115@gw.catspoiler.org> References: <20020619061332.GA85935@elvis.mu.org> <200206190810.g5J8AKM1065115@gw.catspoiler.org>
next in thread | previous in thread | raw e-mail | index | archive | help
re execve() issues...
* Don Lewis <dl-freebsd@catspoiler.org> [020619 01:10] wrote:
> On 18 Jun, Alfred Perlstein wrote:
> > Thanks a ton for taking care of it, your patch is actually cleaner
> > than what I had started on, I'll be committing it shortly.
>
> While you're working in this area, take a look at execve(). If the
> fdcheckstd() test fails, we leak ucred and uidinfo structures, and also
> leave the proc locked. The fix is pretty straightforward.
I'm a bit confused actually, it looks like just unlocking the proc
and then moving the 'exec_fail_dealloc' label higher would fix it,
except I'm not sure about the:
/*
* Handle deferred decrement of ref counts.
*/
if (textvp != NULL)
vrele(textvp);
#ifdef KTRACE
if (tracevp != NULL)
vrele(tracevp);
#endif
pargs_drop(oldargs);
part... should that be before or after exec_fail_dealloc?
Any ideas?
--
-Alfred Perlstein [alfred@freebsd.org]
'Instead of asking why a piece of software is using "1970s technology,"
start asking why software is ignoring 30 years of accumulated wisdom.'
Tax deductible donations for FreeBSD: http://www.freebsdfoundation.org/
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020619172107.GF85935>