linux-kernel - Re: [PATCH] zram: introduce per-device debug

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: <20160513062303.GA21204@bbox>
Date:	Fri, 13 May 2016 15:23:03 +0900
From:	Minchan Kim <minchan@...nel.org>
To:	Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>
CC:	Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	<linux-kernel@...r.kernel.org>, <linux-mm@...ck.org>
Subject: Re: [PATCH] zram: introduce per-device debug_stat sysfs node

On Fri, May 13, 2016 at 10:09:29AM +0900, Sergey Senozhatsky wrote:
> Hello Minchan,
> 
> On (05/13/16 08:41), Minchan Kim wrote:
> [..]
> 
> will fix and update, thanks!
> 
> 
> > > @@ -719,6 +737,8 @@ compress_again:
> > >  		zcomp_strm_release(zram->comp, zstrm);
> > >  		zstrm = NULL;
> > >  
> > > +		atomic64_inc(&zram->stats.num_recompress);
> > > +
> > 
> > It should be below "goto compress_again".
> 
> I moved it out of goto intentionally. this second zs_malloc()
> 
> 		handle = zs_malloc(meta->mem_pool, clen,
> 				GFP_NOIO | __GFP_HIGHMEM |
> 				__GFP_MOVABLE);
> 
> can take some time to complete, which will slow down zram for a bit,
> and _theoretically_ this second zs_malloc() still can fail. yes, we
> would do the error print out pr_err("Error allocating memory ... ")
> and inc the `failed_writes' in zram_bvec_rw(), but zram_bvec_write()
> has several more error return paths that can inc the `failed_writes'.
> so by just looking at the stats we won't be able to tell that we had
> failed fast path allocation combined with failed slow path allocation
> (IOW, `goto recompress' never happened).
> 
> so I'm thinking about changing its name to num_failed_fast_compress
> or num_failed_fast_write, or something similar and thus count the number
> of times we fell to "!handle" branch, not the number of goto-s.
> what do you think? or do you want it to be num_recompress specifically?

Sorry, I don't get your point.
What's the problem with below?

        goto compress_again
        so
        atomic_inc(num_recompress)

My concern isn't a performance or something but just want to be more
readable and not error-prone which can increase num_compress although
second zs_malloc could be failed. When I tested with heavy workload,
I saw second zs_malloc can be fail but not frequently so it's not
theoretical issue.

What's the your concern with below?
Sorry if I don't get your point. Please elaborate it more.

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index a629bd8d452b..8bdcc4b2b9b8 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -737,12 +737,12 @@ static int zram_bvec_write(struct zram *zram, struct bio_vec *bvec, u32 index,
 		zcomp_strm_release(zram->comp, zstrm);
 		zstrm = NULL;
 
-		atomic64_inc(&zram->stats.num_recompress);
-
 		handle = zs_malloc(meta->mem_pool, clen,
 				GFP_NOIO | __GFP_HIGHMEM);
-		if (handle)
+		if (handle) {
+			atomic64_inc(&zram->stats.num_recompress);
 			goto compress_again;
+		}
 
 		pr_err("Error allocating memory for compressed page: %u, size=%zu\n",
 			index, clen);