Date: Mon, 11 Feb 2002 23:13:02 -0500 (EST) From: John Baldwin <jhb@FreeBSD.org> To: Terry Lambert <tlambert2@mindspring.com> Cc: Julian Elischer <julian@elischer.org>, bde@freebsd.org, current@freebsd.org, Alfred Perlstein <bright@mu.org> Subject: Re: ucred holding patch, BDE version Message-ID: <XFMail.020211231302.jhb@FreeBSD.org> In-Reply-To: <3C6886C4.B2B08C5B@mindspring.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 12-Feb-02 Terry Lambert wrote: > John Baldwin wrote: >> Yes, calling free() without Giant is about as good as calling fdrop() >> without >> Giant Alfred. :) > > Alfred would be right, for per processor memory pools. 8-). > >> >> And on the way into the system it does: >> >> lock process >> >> crhold() (which includes mutex ops) >> >> unlock process >> > >> > This isn't needed, at least afaik. >> >> Not strictly for the comparison as Julian and Terry pointed out since the >> race >> can occur anyway (i.e., you don't need the lock to see if p_ucred changed), >> however, if you are actually doing a crhold(), you want to make sure p_ucred >> isn't stale, so you need the proc lock. > > No. If you _depend_ on the frequency of change being low, > you can do this with only atomic reference counts. See the > pseudo code in my other posting, in direct response to you. Yes, the broken code with the race condition that can corrupt random kernel structures long enough to trigger a panic or break a condition test in a branch or loop. I saw that, yes. > -- Terry -- John Baldwin <jhb@FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.020211231302.jhb>