by Darrick J. Wong

[permalink] [raw]

Subject: Re: [PATCH 04/74] libext2fs: use ext2fs_punch() to truncate quota file

On Thu, Dec 12, 2013 at 12:07:09PM -0800, Darrick J. Wong wrote:
> On Thu, Dec 12, 2013 at 12:36:34PM -0500, Theodore Ts'o wrote:
> > On Thu, Dec 12, 2013 at 12:28:40PM -0500, Theodore Ts'o wrote:
> > > On Tue, Dec 10, 2013 at 05:18:43PM -0800, Darrick J. Wong wrote:
> > > > Use the new ext2fs_punch() call to truncate the quota file. This also
> > > > eliminates the need to fix it to work with bigalloc.
> > > >
> > > > Reviewed-by: Zheng Liu <[email protected]>
> > > > Signed-off-by: Darrick J. Wong <[email protected]>
> > >
> > > Thanks, applied.
> >
> > Hmm, I spoke too soon. This is causing a test failure when applied
> > against the maint branch. I know that you developed this versus the
> > next branch, so it's possible the bug in ext2fs_punch() was fixed
> > next. I'm going to drop this from the maint branch for now, I'll take
> > a closer look at this after going through the rest of the bug fix
> > patches, and see whether it works when applied against next (in which
> > case I'll have to see which bug fix we need to apply to backport the
> > fix to the maint branch).
>
> Hmm, I'll try to reproduce it here; which test is failing?

Is t_quota_2off the culprit?

> It's possible that I fubar'd the ordering when I resorted the patches just
> prior to sending them out. :/

I discovered that at some point, configure.in got touched, which caused
configure to be re-run during a build. Unfortunately --disable-quota is the
default, so the quota check part tests never got run, and that's why I never
noticed on subsequent runs of make check. :(

The problem is that on a ^extents filesystem (such as what t_quota_2off
creates), the end parameter of ~0ULL overflows the (end - start + 1)
calculation, effectively setting count to 0, and no blocks get punched.
libquota then zeroes the inode, and the blocks are lost.

Since block map files shouldn't really have more than 2^32 blocks anyway, I
think it's safe to clamp 'end' to 2^32 in the non-extent case.

I'll send a patch shortly.

--D

2013-12-12 22:28:20

by Andreas Dilger

[permalink] [raw]

Subject: Re: [PATCH 03/74] mke2fs: load configfile blocksize setting before 64bit checks

On Dec 10, 2013, at 6:18 PM, Darrick J. Wong <[email protected]> wrote:

> mke2fs has a series of checks to ensure that we don't create a
> filesystem too big for its blocksize -- if auto-64bit is on, then it
> turns on 64bit; otherwise it complains. Unfortunately, it performs
> these checks before looking in mke2fs.conf for a blocksize, which
> means that the checks are incorrect if the user specifies a non-4096
> blocksize in the config file and says nothing on the command line.
> The bug also has the effect of mandating a 4k block size on any block
> device larger than 4T in that situation. Therefore, read the block
> size from the config file before performing the 64bit checks.
>
> Reviewed-by: Zheng Liu <[email protected]>
> Signed-off-by: Darrick J. Wong <[email protected]>
> ---
> misc/mke2fs.c | 134 +++++++++++++++++++++++++++++++--------------------------
> 1 file changed, 72 insertions(+), 62 deletions(-)
>
>
> diff --git a/misc/mke2fs.c b/misc/mke2fs.c
> index 67c9225..19b6e85 100644
> --- a/misc/mke2fs.c
> +++ b/misc/mke2fs.c
> @@ -1780,15 +1795,67 @@ profile_error:
> }
> }
>
> + /* Get the hardware sector sizes, if available */
> + retval = ext2fs_get_device_sectsize(device_name, &lsector_size);
> + if (retval) {
> + com_err(program_name, retval,
> + _("while trying to determine hardware sector size"));
> + exit(1);
> + }
> + retval = ext2fs_get_device_phys_sectsize(device_name, &psector_size);
> + if (retval) {
> + com_err(program_name, retval,
> + _("while trying to determine physical sector size"));
> + exit(1);
> + }
> +
> + tmp = getenv("MKE2FS_DEVICE_SECTSIZE");
> + if (tmp != NULL)
> + lsector_size = atoi(tmp);
> + tmp = getenv("MKE2FS_DEVICE_PHYS_SECTSIZE");
> + if (tmp != NULL)
> + psector_size = atoi(tmp);
> +
> + /* Older kernels may not have physical/logical distinction */
> + if (!psector_size)
> + psector_size = lsector_size;
> +
> + if (blocksize <= 0) {
> + use_bsize = get_int_from_profile(fs_types, "blocksize", 4096);
> +
> + if (use_bsize == -1) {
> + use_bsize = sys_page_size;
> + if ((linux_version_code < (2*65536 + 6*256)) &&

Would be nice to have a helper to compute the linux_version_code comparison,
the above is a bit too much detail for this code.

> + (use_bsize > 4096))
> + use_bsize = 4096;
> + }
> + if (lsector_size && use_bsize < lsector_size)
> + use_bsize = lsector_size;
> + if ((blocksize < 0) && (use_bsize < (-blocksize)))
> + use_bsize = -blocksize;
> + blocksize = use_bsize;
> + fs_blocks_count /= (blocksize / 1024);
> + } else {
> + if (blocksize < lsector_size) { /* Impossible */
> + com_err(program_name, EINVAL,
> + _("while setting blocksize; too small "
> + "for device\n"));
> + exit(1);
> + } else if ((blocksize < psector_size) &&
> + (psector_size <= sys_page_size)) { /* Suboptimal */
> + fprintf(stderr, _("Warning: specified blocksize %d is "
> + "less than device physical sectorsize %d\n"),
> + blocksize, psector_size);
> + }
> + }
> +
> + fs_param.s_log_block_size =
> + int_log2(blocksize >> EXT2_MIN_BLOCK_LOG_SIZE);

Does it make sense to wrap this whole block size guessing dance into a small
helper routine, like "figure_fs_blocksize()" or similar?

Cheers, Andreas

Attachments:

signature.asc (833.00 B)
Message signed with OpenPGP using GPGMail

2013-12-12 22:33:19

On Tue, Dec 10, 2013 at 05:24:52PM -0800, Darrick J. Wong wrote:
> Use the new extended attribute APIs to display all extended attributes
> (current code does not look in the EA block) and display full names
> (current code ignores name index too).
>
> Signed-off-by: Darrick J. Wong <[email protected]>

Applied, thanks.

- Ted