:Matthew Dillon wrote:
:
:> Things like file meta-data always sync, though if the only data written
:> to a file is through such an mmap() the filesystem will not have much
:> to do there.
:
:Incidentally, I notice that files "read" through a mmap (PROT_READ,
:MAP_SHARED) don't cause the file's atime to be set, on 3.3-STABLE
:and 2.2.8-STABLE at least.
:
:This doesn't strike me as the right thing to do, even if the stat(2) man
:page says:
I think you're right, but it isn't something that I can fix this
second. I recommend filing a PR and then emailing me the PR number.
I'll assign it to myself so it shows up in my weekly reminder but
I don't think I'm going to get to it for a couple of months.
:I'm inclined to the view that it would be nice if mtime was set at
:the first write reference that would change the underlying object,
:but not synced out to disk at that time. This means running
:processes do at least see the mtime change immediately, just as
:they'd see the changes if they mmap'd the file.
:
: Richard
:--
:Richard Wendland [EMAIL PROTECTED]
I would hesitate to delve that deeply into the VFS system from a VM
fault for performance reasons. The Open Group specification you
quoted seems reasonable to me, but the mtime specification is not
100% achieveable % in a coherent VM/buffer-cache because once
the page is dirtied the process can make further modifications to it
(that essentially modify the underlying file) without any further
faulting.
-Matt
Matthew Dillon
<[EMAIL PROTECTED]>
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
