2005-01-06 11:44:04

by Marco Cipullo

Subject: From last __GFP_ZERO changes

>From last __GFP_ZERO changes:

--- a/drivers/block/pktcdvd.c 2005-01-06 03:27:45 -08:00
+++ b/drivers/block/pktcdvd.c 2005-01-06 03:27:45 -08:00
@@ -135,12 +135,10 @@
goto no_bio;

for (i = 0; i < PAGES_PER_PACKET; i++) {
- pkt->pages[i] = alloc_page(GFP_KERNEL);
+ pkt->pages[i] = alloc_page(GFP_KERNEL|| __GFP_ZERO);

Is this OK?

Or should be

for (i = 0; i < PAGES_PER_PACKET; i++) {
- pkt->pages[i] = alloc_page(GFP_KERNEL);
+ pkt->pages[i] = alloc_page(GFP_KERNEL| __GFP_ZERO);


Bye
Marco

2005-01-07 19:34:56

by Martin Hicks

Subject: Re: From last __GFP_ZERO changes

On Thu, Jan 06, 2005 at 12:43:58PM +0100, Marco Cipullo wrote:
> From last __GFP_ZERO changes:
>
> --- a/drivers/block/pktcdvd.c 2005-01-06 03:27:45 -08:00
> +++ b/drivers/block/pktcdvd.c 2005-01-06 03:27:45 -08:00
> @@ -135,12 +135,10 @@
> goto no_bio;
>
> for (i = 0; i < PAGES_PER_PACKET; i++) {
> - pkt->pages[i] = alloc_page(GFP_KERNEL);
> + pkt->pages[i] = alloc_page(GFP_KERNEL|| __GFP_ZERO);
>
> Is this OK?
>
> Or should be
>
> for (i = 0; i < PAGES_PER_PACKET; i++) {
> - pkt->pages[i] = alloc_page(GFP_KERNEL);
> + pkt->pages[i] = alloc_page(GFP_KERNEL| __GFP_ZERO);

It definitely should be the latter.

CCing Christoph so he can fix this up.

mh

--
Martin Hicks Wild Open Source Inc.
[email protected] 613-266-2296

2005-01-07 20:05:49

by Christoph Lameter

Subject: Re: From last __GFP_ZERO changes

On Fri, 7 Jan 2005, Martin Hicks wrote:

>
> On Thu, Jan 06, 2005 at 12:43:58PM +0100, Marco Cipullo wrote:
> > From last __GFP_ZERO changes:
> >
> > --- a/drivers/block/pktcdvd.c 2005-01-06 03:27:45 -08:00
> > +++ b/drivers/block/pktcdvd.c 2005-01-06 03:27:45 -08:00
> > @@ -135,12 +135,10 @@
> > goto no_bio;
> >
> > for (i = 0; i < PAGES_PER_PACKET; i++) {
> > - pkt->pages[i] = alloc_page(GFP_KERNEL);
> > + pkt->pages[i] = alloc_page(GFP_KERNEL|| __GFP_ZERO);
> >
> > Is this OK?
> >
> > Or should be
> >
> > for (i = 0; i < PAGES_PER_PACKET; i++) {
> > - pkt->pages[i] = alloc_page(GFP_KERNEL);
> > + pkt->pages[i] = alloc_page(GFP_KERNEL| __GFP_ZERO);
>
> It definitely should be the latter.

Correct. Please submit the patch. I see a patch that fixes a pktcdvd.c
issue in Andrews tree so it was likely already fixed without me hearing of
it.