From: Nick Piggin <npiggin@suse.de>
To: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
akpm@linux-foundation.org, linux-mm@kvack.org,
randy.dunlap@oracle.com
Subject: Re: [patch 2/8] mm: merge populate and nopage into fault (fixes nonlinear)
Date: Thu, 24 May 2007 03:48:03 +0200 [thread overview]
Message-ID: <20070524014803.GB22998@wotan.suse.de> (raw)
In-Reply-To: <1179963439.32247.987.camel@localhost.localdomain>
On Thu, May 24, 2007 at 09:37:19AM +1000, Benjamin Herrenschmidt wrote:
> On Fri, 2007-05-18 at 08:11 -0700, Linus Torvalds wrote:
>
> > Also, the commentary says that you're planning on replacing "nopfn" too,
> > which means that returning a "struct page *" is wrong. So the patch is
> > introducing a new interface that is already known to be broken.
>
> Agreed.
Yep, I will change it to return the fault type. This makes page_mkwrite
merge cleaner too.
> > - "struct fault_data" is a stupid name. Of *course* it is data: it's a
> > struct. It can't be code. But it's not even about faults. It's about
> > missing pages.
> >
> > So call it something else. Maybe just "struct nopage". Or, "struct
> > vm_fault" at least, so that it's at least not about *random* faults.
> >
> > - drop "address" from "struct fault_data". Even if some user were to have
> > some reason to use it (doubtful), it should be called somethign long
> > and cumbersome, so that you don't use it by mistake, not realizing that
> > you should use the page index instead.
>
> I'd rather have it in, even if it's long and cumbersome :-) As I said,
> there are a few HW drivers around the tree like spufs or some weirdo IBM
> infiniband stuff that do really tricky games with nopage/nopfn and which
> can have good use of it (at the very least, it's useful for debugging to
> printk where the accesses that ended up doing the wrong thing precisely
> was done :-)
>
> > - and keep calling it "nopage".
>
> Fine by me.
I won't do this. I'll keep calling it fault, because a) it means we keep
the backwards compatible ->nopage path until all drivers are converted,
and b) the page_mkwrite conversion really will make "nopage" the wrong
name.
I don't mind about the struct naming, but struct fault or struct nopage
seems weird to me... but whatever. Maybe we can change it to struct
fault_struct to go along with struct task_struct and struct mm_struct? ;)
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2007-05-24 1:48 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-05-18 7:37 akpm, Nick Piggin
2007-05-18 10:24 ` Christoph Hellwig
2007-05-18 15:11 ` Linus Torvalds
2007-05-19 1:38 ` Nick Piggin
2007-05-22 15:12 ` Christoph Hellwig
2007-05-23 1:14 ` Linus Torvalds
2007-05-23 8:35 ` David Chinner
2007-05-23 9:35 ` Nick Piggin
2007-05-23 2:57 ` Nick Piggin
2007-05-23 23:37 ` Benjamin Herrenschmidt
2007-05-24 1:48 ` Nick Piggin [this message]
2007-05-24 2:06 ` Linus Torvalds
2007-05-24 2:24 ` Nick Piggin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20070524014803.GB22998@wotan.suse.de \
--to=npiggin@suse.de \
--cc=akpm@linux-foundation.org \
--cc=benh@kernel.crashing.org \
--cc=linux-mm@kvack.org \
--cc=randy.dunlap@oracle.com \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox