[PATCH] e2fsck: return more status if fsck aborts

[PATCH] e2fsck: return more status if fsck aborts

[Eric Sandeen]

If we abort fsck (due to ENOMEM for example) we exit
with only the FSCK_ERROR flag.  It seems useful
to do the same sorts of checks as we do on normal
exit, and return whether the filesystem was modified,
whether there are still uncorrected errors, etc, even
in the abort case.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
---

diff --git a/e2fsck/util.c b/e2fsck/util.c
index fb9a87a..d761ae5 100644
--- a/e2fsck/util.c
+++ b/e2fsck/util.c
@@ -45,18 +45,37 @@ extern e2fsck_t e2fsck_global_ctx;   /* Try your very best not to use this! */
 
 void fatal_error(e2fsck_t ctx, const char *msg)
 {
+	ext2_filsys fs = ctx->fs;
+	int exit_value = FSCK_ERROR;
+
 	if (msg)
 		fprintf (stderr, "e2fsck: %s\n", msg);
-	if (ctx->fs && ctx->fs->io) {
+	if (!fs)
+		goto out;
+	if (fs->io) {
 		if (ctx->fs->io->magic == EXT2_ET_MAGIC_IO_CHANNEL)
 			io_channel_flush(ctx->fs->io);
 		else
 			fprintf(stderr, "e2fsck: io manager magic bad!\n");
 	}
+	if (ext2fs_test_changed(fs)) {
+		exit_value |= FSCK_NONDESTRUCT;
+		printf(_("\n%s: ***** FILE SYSTEM WAS MODIFIED *****\n"),
+			ctx->device_name);
+		if (ctx->mount_flags & EXT2_MF_ISROOT)
+			exit_value |= FSCK_REBOOT;
+	}
+	if (!ext2fs_test_valid(fs)) {
+		printf(_("\n%s: ********** WARNING: Filesystem still has "
+			 "errors **********\n\n"), ctx->device_name);
+		exit_value |= FSCK_UNCORRECTED;
+		exit_value &= ~FSCK_NONDESTRUCT;
+	}
+out:
 	ctx->flags |= E2F_FLAG_ABORT;
 	if (ctx->flags & E2F_FLAG_SETJMP_OK)
 		longjmp(ctx->abort_loc, 1);
-	exit(FSCK_ERROR);
+	exit(exit_value);
 }
 
 void *e2fsck_allocate_memory(e2fsck_t ctx, unsigned int size,

Re: [PATCH] e2fsck: return more status if fsck aborts

[Ted Ts'o]

On Tue, Nov 01, 2011 at 12:32:28PM -0500, Eric Sandeen wrote:
> If we abort fsck (due to ENOMEM for example) we exit
> with only the FSCK_ERROR flag.  It seems useful
> to do the same sorts of checks as we do on normal
> exit, and return whether the filesystem was modified,
> whether there are still uncorrected errors, etc, even
> in the abort case.
> 
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>

Thanks, applied.

					- Ted