linux-kernel - Re: [PATCH] use clear_page()/copy_page() in favor of memset()/memcpy() on whole pages

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <201009150211.09089.rjw@sisk.pl>
Date:	Wed, 15 Sep 2010 02:11:08 +0200
From:	"Rafael J. Wysocki" <rjw@...k.pl>
To:	Andrew Morton <akpm@...ux-foundation.org>
Cc:	"Jan Beulich" <JBeulich@...ell.com>, linux-kernel@...r.kernel.org,
	Miklos Szeredi <miklos@...redi.hu>
Subject: Re: [PATCH] use clear_page()/copy_page() in favor of memset()/memcpy() on whole pages

On Wednesday, September 15, 2010, Andrew Morton wrote:
> On Thu, 02 Sep 2010 13:30:37 +0100
> "Jan Beulich" <JBeulich@...ell.com> wrote:
> 
> > After all that's what they are intended for.
> > 
> > ...
> >
> > --- linux-2.6.36-rc3/fs/fuse/dev.c
> > +++ 2.6.36-rc3-use-clear_page/fs/fuse/dev.c
> > @@ -811,7 +811,7 @@ static int fuse_copy_page(struct fuse_co
> >  
> >  	if (page && zeroing && count < PAGE_SIZE) {
> >  		void *mapaddr = kmap_atomic(page, KM_USER1);
> > -		memset(mapaddr, 0, PAGE_SIZE);
> > +		clear_page(mapaddr);
> >  		kunmap_atomic(mapaddr, KM_USER1);
> >  	}
> >  	while (count) {
> 
> fuse wanted to use clear_highpage() here.  But clear_highpage() uses
> KM_USER0.  I don't immediately see why fuse uses KM_USER1 here?
> 
> >
> > ...
> >
> > --- linux-2.6.36-rc3/kernel/power/snapshot.c
> > +++ 2.6.36-rc3-use-clear_page/kernel/power/snapshot.c
> > @@ -988,7 +988,7 @@ static void copy_data_page(unsigned long
> >  			 */
> >  			safe_copy_page(buffer, s_page);
> >  			dst = kmap_atomic(d_page, KM_USER0);
> > -			memcpy(dst, buffer, PAGE_SIZE);
> > +			copy_page(dst, buffer);
> >  			kunmap_atomic(dst, KM_USER0);
> >  		} else {
> >  			safe_copy_page(page_address(d_page), s_page);
> > @@ -1636,7 +1636,7 @@ int snapshot_read_next(struct snapshot_h
> >  		memory_bm_position_reset(&orig_bm);
> >  		memory_bm_position_reset(&copy_bm);
> >  	} else if (handle->cur <= nr_meta_pages) {
> > -		memset(buffer, 0, PAGE_SIZE);
> > +		clear_page(buffer);
> >  		pack_pfns(buffer, &orig_bm);
> >  	} else {
> >  		struct page *page;
> > @@ -1650,7 +1650,7 @@ int snapshot_read_next(struct snapshot_h
> >  			void *kaddr;
> >  
> >  			kaddr = kmap_atomic(page, KM_USER0);
> > -			memcpy(buffer, kaddr, PAGE_SIZE);
> > +			copy_page(buffer, kaddr);
> >  			kunmap_atomic(kaddr, KM_USER0);
> >  			handle->buffer = buffer;
> >  		} else {
> > @@ -1933,7 +1933,7 @@ static void copy_last_highmem_page(void)
> >  		void *dst;
> >  
> >  		dst = kmap_atomic(last_highmem_page, KM_USER0);
> > -		memcpy(dst, buffer, PAGE_SIZE);
> > +		copy_page(dst, buffer);
> >  		kunmap_atomic(dst, KM_USER0);
> >  		last_highmem_page = NULL;
> >  	}
> > @@ -2219,9 +2219,9 @@ swap_two_pages_data(struct page *p1, str
> >  
> >  	kaddr1 = kmap_atomic(p1, KM_USER0);
> >  	kaddr2 = kmap_atomic(p2, KM_USER1);
> > -	memcpy(buf, kaddr1, PAGE_SIZE);
> > -	memcpy(kaddr1, kaddr2, PAGE_SIZE);
> > -	memcpy(kaddr2, buf, PAGE_SIZE);
> > +	copy_page(buf, kaddr1);
> > +	copy_page(kaddr1, kaddr2);
> > +	copy_page(kaddr2, buf);
> >  	kunmap_atomic(kaddr1, KM_USER0);
> >  	kunmap_atomic(kaddr2, KM_USER1);
> >  }
> 
> The page-copying functions in snapshot.c are magical.  The changes look
> OK to me but I'd ask Rafael to double-check, please.

Well, they _look_ OK to me too.

Rafael
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/