2006-08-28 20:49:45

by Andrew Morton

[permalink] [raw]
Subject: Re: [PATCH RFP-V4 00/13] remap_file_pages protection support - 4th attempt

On Sat, 26 Aug 2006 19:33:35 +0200
Blaisorblade <[email protected]> wrote:

> Again, about 4 month since last time (for lack of time) I'm sending for final
> review and for inclusion into -mm protection support for remap_file_pages (in
> short "RFP prot support"), i.e. setting per-pte protections (beyond file
> offset) through this syscall.

This all looks a bit too fresh and TODO-infested for me to put it in -mm at
this time.

I could toss them in to get some testing underway, but that makes life
complex for other ongoing MM work. (And there's a _lot_ of that - I
presently have >180 separate patches which alter ./mm/*).

Also, it looks like another round of detailed review is needed before this
work will really start to settle into its final form.

So.. I'll await version 5, sorry. Please persist.


2006-08-29 08:22:48

by Blaisorblade

[permalink] [raw]
Subject: Re: [PATCH RFP-V4 00/13] remap_file_pages protection support - 4th attempt

Andrew Morton <[email protected]> ha scritto:

> On Sat, 26 Aug 2006 19:33:35 +0200
> Blaisorblade <[email protected]> wrote:
>
> > Again, about 4 month since last time (for lack of time) I'm
> sending for final
> > review and for inclusion into -mm protection support for
> remap_file_pages (in
> > short "RFP prot support"), i.e. setting per-pte protections
> (beyond file
> > offset) through this syscall.

> This all looks a bit too fresh and TODO-infested for me to put it
> in -mm at
> this time.

It is possible, subsequent rounds of review should be near to each
other, but calling the code "new" is maybe exaggerate. I do not
remember all these TODOs but I may forget (and I don't have my box
right now, so I can't check).

> I could toss them in to get some testing underway, but that makes
> life
> complex for other ongoing MM work. (And there's a _lot_ of that -
> I
> presently have >180 separate patches which alter ./mm/*).

That's fine. If this can help I could try to base next version
against -mm.

> Also, it looks like another round of detailed review is needed
> before this
> work will really start to settle into its final form.

That's ok, I prefer reviews to testing right now. Almost all but 1
patch (which is marked) is unit tested on i386, x86_64 and uml (but
if I don't have a multithreaded concurrent fault tester), so it's
time to catch remaining bugs by review.

> So.. I'll await version 5, sorry. Please persist.

I'll try. I just hope we'll not have it next summer (I know it's my
problem, I'm not complaining on you).

Thanks!
Bye
--
Paolo Giarrusso

Chiacchiera con i tuoi amici in tempo reale!
http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com