public inbox for linux-mtd@lists.infradead.org
 help / color / mirror / Atom feed
* UBIFS defaults?
@ 2011-03-09 13:01 Artem Bityutskiy
  2011-03-10  7:04 ` Artem Bityutskiy
  2011-03-10 14:28 ` Artem Bityutskiy
  0 siblings, 2 replies; 4+ messages in thread
From: Artem Bityutskiy @ 2011-03-09 13:01 UTC (permalink / raw)
  To: Hunter Adrian (Nokia-MS/Helsinki), linux-mtd; +Cc: Bhavesh Parekh

Hi,

currently, UBIFS checks data CRC on read by default, i.e., the
'chk_data_crc' mount option is the default one. This makes people who
apply benchmarks less happy, because they just use the defaults.

Should we make the 'no_chk_data_crc' mount option to be the default
instead? The rationale is that people who do care to read the
documentation can switch to 'chk_data_crc' if they need, and people who
do quick UBIFS evaluation will end up with faster read speed by default.

See this section for more information about the 'no_chk_data_crc' mount
option: http://www.linux-mtd.infradead.org/doc/ubifs.html#L_checksumming

Would be nice to hear UBIFS users' opinion.

-- 
Best Regards,
Artem Bityutskiy (Артём Битюцкий)

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: UBIFS defaults?
  2011-03-09 13:01 UBIFS defaults? Artem Bityutskiy
@ 2011-03-10  7:04 ` Artem Bityutskiy
  2011-03-10  8:48   ` Jon Povey
  2011-03-10 14:28 ` Artem Bityutskiy
  1 sibling, 1 reply; 4+ messages in thread
From: Artem Bityutskiy @ 2011-03-10  7:04 UTC (permalink / raw)
  To: Hunter Adrian (Nokia-MS/Helsinki); +Cc: Bhavesh Parekh, linux-mtd

On Wed, 2011-03-09 at 15:01 +0200, Artem Bityutskiy wrote:
> Hi,
> 
> currently, UBIFS checks data CRC on read by default, i.e., the
> 'chk_data_crc' mount option is the default one. This makes people who
> apply benchmarks less happy, because they just use the defaults.
> 
> Should we make the 'no_chk_data_crc' mount option to be the default
> instead? The rationale is that people who do care to read the
> documentation can switch to 'chk_data_crc' if they need, and people who
> do quick UBIFS evaluation will end up with faster read speed by default.
> 
> See this section for more information about the 'no_chk_data_crc' mount
> option: http://www.linux-mtd.infradead.org/doc/ubifs.html#L_checksumming
> 
> Would be nice to hear UBIFS users' opinion.

I'm  going to change the default if no one objects.

-- 
Best Regards,
Artem Bityutskiy (Артём Битюцкий)

^ permalink raw reply	[flat|nested] 4+ messages in thread

* RE: UBIFS defaults?
  2011-03-10  7:04 ` Artem Bityutskiy
@ 2011-03-10  8:48   ` Jon Povey
  0 siblings, 0 replies; 4+ messages in thread
From: Jon Povey @ 2011-03-10  8:48 UTC (permalink / raw)
  To: dedekind1@gmail.com, Hunter Adrian (Nokia-MS/Helsinki)
  Cc: linux-mtd, Bhavesh Parekh

linux-mtd-bounces@lists.infradead.org wrote:
> On Wed, 2011-03-09 at 15:01 +0200, Artem Bityutskiy wrote:
>> Hi,
>>
>> currently, UBIFS checks data CRC on read by default, i.e., the
>> 'chk_data_crc' mount option is the default one. This makes people who
>> apply benchmarks less happy, because they just use the defaults.
>>
>> Should we make the 'no_chk_data_crc' mount option to be the default
>> instead? The rationale is that people who do care to read the
>> documentation can switch to 'chk_data_crc' if they need, and people
>> who do quick UBIFS evaluation will end up with faster read speed by
>> default.
>>
>> See this section for more information about the 'no_chk_data_crc'
>> mount option:
> http://www.linux-mtd.infradead.org/doc/ubifs.html#L_checksumming
>>
>> Would be nice to hear UBIFS users' opinion.
>
> I'm  going to change the default if no one objects.

It sounds OK to me (for my uses), but:

It sounds important, I think you should wait more than 24h for comments.

--
Jon Povey
jon.povey@racelogic.co.uk

Racelogic is a limited company registered in England. Registered number 2743719 .
Registered Office Unit 10, Swan Business Centre, Osier Way, Buckingham, Bucks, MK18 1TB .

The information contained in this electronic mail transmission is intended by Racelogic Ltd for the use of the named individual or entity to which it is directed and may contain information that is confidential or privileged. If you have received this electronic mail transmission in error, please delete it from your system without copying or forwarding it, and notify the sender of the error by reply email so that the sender's address records can be corrected. The views expressed by the sender of this communication do not necessarily represent those of Racelogic Ltd. Please note that Racelogic reserves the right to monitor e-mail communications passing through its network

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: UBIFS defaults?
  2011-03-09 13:01 UBIFS defaults? Artem Bityutskiy
  2011-03-10  7:04 ` Artem Bityutskiy
@ 2011-03-10 14:28 ` Artem Bityutskiy
  1 sibling, 0 replies; 4+ messages in thread
From: Artem Bityutskiy @ 2011-03-10 14:28 UTC (permalink / raw)
  To: Hunter Adrian (Nokia-MS/Helsinki); +Cc: Bhavesh Parekh, linux-mtd

On Wed, 2011-03-09 at 15:01 +0200, Artem Bityutskiy wrote:
> Hi,
> 
> currently, UBIFS checks data CRC on read by default, i.e., the
> 'chk_data_crc' mount option is the default one. This makes people who
> apply benchmarks less happy, because they just use the defaults.
> 
> Should we make the 'no_chk_data_crc' mount option to be the default
> instead? The rationale is that people who do care to read the
> documentation can switch to 'chk_data_crc' if they need, and people who
> do quick UBIFS evaluation will end up with faster read speed by default.
> 
> See this section for more information about the 'no_chk_data_crc' mount
> option: http://www.linux-mtd.infradead.org/doc/ubifs.html#L_checksumming
> 
> Would be nice to hear UBIFS users' opinion.


>From f7a2d4a6dea3fcec4612aa81baabf8aae0040896 Mon Sep 17 00:00:00 2001
From: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Date: Thu, 10 Mar 2011 16:26:32 +0200
Subject: [PATCH] UBIFS: do not check data crc by default

Change the default UBIFS behavior WRT data CRC checking. Currently,
UBIFS checks data CRC when reading, which slows it down quite a bit,
and this is the default option. However, it looks like in average
user does not need this feature and would prefer faster read speed
over extra reliability. And this seems to be de-facto standard that
file-systems do not check data CRC every time they read from the
media.

Thus, make UBIFS default behavior so that it does not check data
CRC. This corresponds to the no_chk_data_crc mount option. Those users
who need extra protection can always enable it using the chk_data_crc
option.

Please, read more information about this feature here:
http://www.linux-mtd.infradead.org/doc/ubifs.html#L_checksumming

Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
---
 Documentation/filesystems/ubifs.txt |    4 ++--
 fs/ubifs/super.c                    |    1 +
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/Documentation/filesystems/ubifs.txt b/Documentation/filesystems/ubifs.txt
index 12fedb7..d7b13b0 100644
--- a/Documentation/filesystems/ubifs.txt
+++ b/Documentation/filesystems/ubifs.txt
@@ -82,12 +82,12 @@ Mount options
 bulk_read		read more in one go to take advantage of flash
 			media that read faster sequentially
 no_bulk_read (*)	do not bulk-read
-no_chk_data_crc		skip checking of CRCs on data nodes in order to
+no_chk_data_crc (*)	skip checking of CRCs on data nodes in order to
 			improve read performance. Use this option only
 			if the flash media is highly reliable. The effect
 			of this option is that corruption of the contents
 			of a file can go unnoticed.
-chk_data_crc (*)	do not skip checking CRCs on data nodes
+chk_data_crc		do not skip checking CRCs on data nodes
 compr=none              override default compressor and set it to "none"
 compr=lzo               override default compressor and set it to "lzo"
 compr=zlib              override default compressor and set it to "zlib"
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index b722ebc..e5dc1e1 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -1985,6 +1985,7 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
 	INIT_LIST_HEAD(&c->old_buds);
 	INIT_LIST_HEAD(&c->orph_list);
 	INIT_LIST_HEAD(&c->orph_new);
+	c->no_chk_data_crc = 1;
 
 	c->vfs_sb = sb;
 	c->highest_inum = UBIFS_FIRST_INO;
-- 
1.7.2.3

-- 
Best Regards,
Artem Bityutskiy (Артём Битюцкий)

^ permalink raw reply related	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2011-03-10 14:30 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-03-09 13:01 UBIFS defaults? Artem Bityutskiy
2011-03-10  7:04 ` Artem Bityutskiy
2011-03-10  8:48   ` Jon Povey
2011-03-10 14:28 ` Artem Bityutskiy

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox