Date: Fri, 10 Nov 1995 14:38:20 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: bakul@netcom.com (Bakul Shah) Cc: rminnich@sarnoff.com, lm@slovax.engr.sgi.com, hackers@FreeBSD.ORG, waa@aurora.cis.upenn.edu, deraadt@theos.com, chuck@maria.wustl.edu Subject: Re: larry: you might want to add this to lmbench (but i'm not sure) Message-ID: <199511102138.OAA04574@phaeton.artisoft.com> In-Reply-To: <199511101753.JAA10892@netcom22.netcom.com> from "Bakul Shah" at Nov 10, 95 09:53:31 am
next in thread | previous in thread | raw e-mail | index | archive | help
> I don't see what system parameter your benchmark measures > (other than pointing out if arg checking is done early or > late). Since all syscall args need to be checked sooner or > later, they may as well get checked upfront. I'll agree with this, and Ron's test has made the issue more obvious, even if I don't agree with the premise. I've actually argued this before. The problem is that the check must, in the success case, cause a reservation so that the mapping is not allowed to change between the time that the assertion succeeds and the copy takes place. This introduces its own issues, since because you might be doing multiple copies (yeah, I know, you should assemble a local buffer representation of the user space buffer and copy that instead), you can't simply imply the reservation is released like the VFS does with leases. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199511102138.OAA04574>