Re: grub-probe detects ext4 wronly as ext2

["Javier Martín" <lordhabbit@gmail.com>]

[-- Attachment #1.1: Type: text/plain, Size: 1703 bytes --]

El sáb, 07-02-2009 a las 20:30 +0100, Felix Zielcke escribió:
> Am Mittwoch, den 04.02.2009, 14:08 +0100 schrieb Javier Martín:
> 
> > Well, I am happy to post a diff of the patch against current SVN head
> > (r1973). I have personally confirmed (in a VM) that it:
> > 1) Still builds (and even runs! ^^)
> > 2) Works with existing ext2/3 file systems (I haven't checked ext4 FSs
> > but the "extents" bit is marked as supported, so it should work)
> > 3) Correctly rejects journal devices, which will then appear as "unknown
> > filesystem" when accessed.
> 
> FLEX_BG needs to be added to the list of ignored flags.
> As Robert already said in his last reply to this thread [0]
> 
> const char *local_error = 0;
> Please use NULL.
> 
>  +    EXT2_DRIVER_MOUNT_FAIL(0);
> 
> I share his opinion that this isn't needed.
> 
> If you fix this and write a changelog then I commit this.
> 
> [0] http://lists.gnu.org/archive/html/grub-devel/2008-08/msg00645.html

Oops... I was going to send a new version of the patch with those fixed,
but when doing a "svn up" so that it would be against HEAD, I've noticed
that Robert has just integrated a much cleaner version without the macro
and local_error thingies. Well, the only thing left to do is adding
flex_bg - here goes the patch. It also clarifies a comment and corrects
those added in my original patch and Robert's cleaned-up version that
don't end with ".  */" as they should.

-- Lazy, Oblivious, Rational Disaster -- Habbit

BTW: Robert, you're having a total mailing spree today! What's it been,
30 posts? Evolution nearly choked, and my spam filter was about to ban
you as "mass mailing - possible spam" ;)

[-- Attachment #1.2: ext2_incompat_flexbg.patch --]
[-- Type: text/x-patch, Size: 2884 bytes --]

Index: fs/ext2.c
===================================================================
--- fs/ext2.c	(revision 1977)
+++ fs/ext2.c	(working copy)
@@ -73,7 +73,7 @@
 
 /* Superblock filesystem feature flags (RW compatible)
  * A filesystem with any of these enabled can be read and written by a driver
- * that does not understand them without causing metadata/data corruption */
+ * that does not understand them without causing metadata/data corruption.  */
 #define EXT2_FEATURE_COMPAT_DIR_PREALLOC	0x0001
 #define EXT2_FEATURE_COMPAT_IMAGIC_INODES	0x0002
 #define EXT3_FEATURE_COMPAT_HAS_JOURNAL		0x0004
@@ -83,7 +83,7 @@
 /* Superblock filesystem feature flags (RO compatible)
  * A filesystem with any of these enabled can be safely read by a driver that
  * does not understand them, but should not be written to, usually because
- * additional metadata is required */
+ * additional metadata is required.  */
 #define EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER	0x0001
 #define EXT2_FEATURE_RO_COMPAT_LARGE_FILE	0x0002
 #define EXT2_FEATURE_RO_COMPAT_BTREE_DIR	0x0004
@@ -93,7 +93,7 @@
 /* Superblock filesystem feature flags (back-incompatible)
  * A filesystem with any of these enabled should not be attempted to be read
  * by a driver that does not understand them, since they usually indicate
- * metadata format changes that might confuse the reader. */
+ * metadata format changes that might confuse the reader.  */
 #define EXT2_FEATURE_INCOMPAT_COMPRESSION	0x0001
 #define EXT2_FEATURE_INCOMPAT_FILETYPE		0x0002
 #define EXT3_FEATURE_INCOMPAT_RECOVER		0x0004 /* Needs recovery */
@@ -104,17 +104,17 @@
 #define EXT4_FEATURE_INCOMPAT_FLEX_BG		0x0200
 
 /* The set of back-incompatible features this driver DOES support. Add (OR)
- * flags here as the related features are implemented into the driver */
+ * flags here as the related features are implemented into the driver.  */
 #define EXT2_DRIVER_SUPPORTED_INCOMPAT ( EXT2_FEATURE_INCOMPAT_FILETYPE \
-                                       | EXT4_FEATURE_INCOMPAT_EXTENTS )
+                                       | EXT4_FEATURE_INCOMPAT_EXTENTS  \
+                                       | EXT4_FEATURE_INCOMPAT_FLEX_BG )
 /* List of rationales for the ignored "incompatible" features:
  * needs_recovery: Not really back-incompatible - was added as such to forbid
  *                 ext2 drivers from mounting an ext3 volume with a dirty
  *                 journal because they will ignore the journal, but the next
  *                 ext3 driver to mount the volume will find the journal and
  *                 replay it, potentially corrupting the metadata written by
- *                 the ext2 drivers
- */
+ *                 the ext2 drivers. Safe to ignore for this RO driver.  */
 #define EXT2_DRIVER_IGNORED_INCOMPAT ( EXT3_FEATURE_INCOMPAT_RECOVER )
 
 

[-- Attachment #2: Esta parte del mensaje está firmada digitalmente --]
[-- Type: application/pgp-signature, Size: 835 bytes --]