Patchwork mke2fs: check that auto-detected blocksize <= sys_page_size

login
register
mail settings
Submitter Yury V. Zaytsev
Date June 8, 2011, 9:23 a.m.
Message ID <1307524998.2467.41.camel@newpride>
Download mbox | patch
Permalink /patch/99391/
State Superseded
Headers show

Comments

Yury V. Zaytsev - June 8, 2011, 9:23 a.m.
Hi!

I have recently stumbled upon a RHEL6.1 installation problem on an SSD
drive that was working fine with older kernels, where created file
systems could not be mounted and the installation was aborting.

It turned out that the drive had a firmware bug and was exposing a
minimum_io_size of 8912, whereas my x86_64 system had a page size of
4096 bytes.

mkfs in turn was silently creating file systems with 8912 bytes block
size, because it was auto-detected as opposed to specified via the -b
option from the command line.

This made the problem hard to debug, hence I attach a tentative patch to
fix this issue. I'd like to acknowledge Mike Snitzer for directing me to
this list.

I'm not subscribed, so please CC: me upon replies and be mindful that
it's my first contribution of this kind, so even though I did my best to
do it the right way, I might have screwed something up.

Thanks!
Amir Goldstein - June 8, 2011, 12:50 p.m.
On Wed, Jun 8, 2011 at 12:23 PM, Yury V. Zaytsev <yury@shurup.com> wrote:
> Hi!
>
> I have recently stumbled upon a RHEL6.1 installation problem on an SSD
> drive that was working fine with older kernels, where created file
> systems could not be mounted and the installation was aborting.
>
> It turned out that the drive had a firmware bug and was exposing a
> minimum_io_size of 8912, whereas my x86_64 system had a page size of
> 4096 bytes.
>
> mkfs in turn was silently creating file systems with 8912 bytes block
> size, because it was auto-detected as opposed to specified via the -b
> option from the command line.
>
> This made the problem hard to debug, hence I attach a tentative patch to
> fix this issue. I'd like to acknowledge Mike Snitzer for directing me to
> this list.
>
> I'm not subscribed, so please CC: me upon replies and be mindful that
> it's my first contribution of this kind, so even though I did my best to
> do it the right way, I might have screwed something up.

The only thing you screwed up is sending the patch as an attachment.
People here expect it to be as plain text in the email, so that they can
reply and comment inline.
Other than that, the fix looks good (and well explained) to me.

Cheers,
Amir.

>
> Thanks!
>
> --
> Sincerely yours,
> Yury V. Zaytsev
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

From 5dcc3266d46ee3846bd8707c7b99bd312fe2276b Mon Sep 17 00:00:00 2001
From: "Yury V. Zaytsev" <yury@shurup.com>
Date: Wed, 8 Jun 2011 11:12:54 +0200
Subject: [PATCH] mke2fs: check that auto-detected blocksize <= sys_page_size

Block size can be specified manually via the -b option or deduced
automatically. Unfortunately, the check that it is still smaller than
the system page size is only performed right after the command line
options were parsed.

Therefore, if buggy or inappropriately installed/configured hardware
hints that larger block sizes have to be used, mkfs will silently create
a file system which can not be mounted on the system in question.

By moving the check beyond the last assignment to blocksize it is now
ensured, that mkfs will issue a warning even if inappropriate blocksize
was auto-detected.

The new behavior can be easily tested, by exporting the following
variables before running mkfs:

    export MKE2FS_DEVICE_SECTSIZE=8912
    export MKE2FS_DEVICE_PHYS_SECTSIZE=8912

Signed-off-by: Yury V. Zaytsev <yury@shurup.com>
---
 misc/mke2fs.c |   24 +++++++++++++-----------
 1 files changed, 13 insertions(+), 11 deletions(-)

diff --git a/misc/mke2fs.c b/misc/mke2fs.c
index e28828e..b600617 100644
--- a/misc/mke2fs.c
+++ b/misc/mke2fs.c
@@ -1526,17 +1526,6 @@  static void PRS(int argc, char *argv[])
 		ext2fs_close(jfs);
 	}
 
-	if (blocksize > sys_page_size) {
-		if (!force) {
-			com_err(program_name, 0,
-				_("%d-byte blocks too big for system (max %d)"),
-				blocksize, sys_page_size);
-			proceed_question();
-		}
-		fprintf(stderr, _("Warning: %d-byte blocks too big for system "
-				  "(max %d), forced to continue\n"),
-			blocksize, sys_page_size);
-	}
 	if (optind < argc) {
 		fs_param.s_blocks_count = parse_num_blocks(argv[optind++],
 				fs_param.s_log_block_size);
@@ -1811,6 +1800,19 @@  got_size:
 
 	blocksize = EXT2_BLOCK_SIZE(&fs_param);
 
+	/* This check should happen beyond the last assignment to blocksize */
+	if (blocksize > sys_page_size) {
+		if (!force) {
+			com_err(program_name, 0,
+				_("%d-byte blocks too big for system (max %d)"),
+				blocksize, sys_page_size);
+			proceed_question();
+		}
+		fprintf(stderr, _("Warning: %d-byte blocks too big for system "
+				  "(max %d), forced to continue\n"),
+			blocksize, sys_page_size);
+	}
+
 	lazy_itable_init = 0;
 	if (access("/sys/fs/ext4/features/lazy_itable_init", R_OK) == 0)
 		lazy_itable_init = 1;
-- 
1.7.5.4