* [PATCH] fuse2fs: open read-only when ro option and image non-writable
@ 2025-10-16 20:02 Dave Dykstra
2025-10-17 19:24 ` Darrick J. Wong
0 siblings, 1 reply; 6+ messages in thread
From: Dave Dykstra @ 2025-10-16 20:02 UTC (permalink / raw)
To: linux-ext4; +Cc: Dave Dykstra, Dave Dykstra
This opens the image read-only when an ro option is given and the
image is not writable. If it is then found that a journal recovery
is needed, an error is returned then.
The ret value is set to 4 after the option checks so that if there's
an error resulting in "goto out" it won't print an error about
unrecognized options.
Also submitted as PR https://github.com/tytso/e2fsprogs/pull/250
for the issue https://github.com/tytso/e2fsprogs/issues/244.
Replaces
https://lore.kernel.org/linux-ext4/20251010214735.22683-1-dave.dykstra@cern.ch/T/#u
https://lore.kernel.org/linux-ext4/175798064776.350013.6744611652039454651.stgit@frogsfrogsfrogs/#t
Signed-off-by: Dave Dykstra <2129743+DrDaveD@users.noreply.github.com>
---
misc/fuse2fs.c | 27 +++++++++++++++++++++++----
1 file changed, 23 insertions(+), 4 deletions(-)
diff --git a/misc/fuse2fs.c b/misc/fuse2fs.c
index cb5620c7..2ae2fc1a 100644
--- a/misc/fuse2fs.c
+++ b/misc/fuse2fs.c
@@ -4696,9 +4696,19 @@ int main(int argc, char *argv[])
err = ext2fs_open2(fctx.device, options, flags, 0, 0, unix_io_manager,
&global_fs);
if (err) {
- err_printf(&fctx, "%s.\n", error_message(err));
- err_printf(&fctx, "%s\n", _("Please run e2fsck -fy."));
- goto out;
+ if (((err == EACCES) || (err == EPERM)) && fctx.ro) {
+ // read-only requested and don't have write access
+ dbg_printf(&fctx, "%s: %s\n", __func__,
+ _("Permission denied with writable, trying without.\n"));
+ flags &= ~EXT2_FLAG_RW;
+ err = ext2fs_open2(fctx.device, options, flags, 0, 0,
+ unix_io_manager, &global_fs);
+ }
+ if (err) {
+ err_printf(&fctx, "%s.\n", error_message(err));
+ err_printf(&fctx, "%s\n", _("Please run e2fsck -fy."));
+ goto out;
+ }
}
fctx.fs = global_fs;
global_fs->priv_data = &fctx;
@@ -4741,6 +4751,8 @@ int main(int argc, char *argv[])
goto out;
}
+ ret = 4;
+
if (global_fs->super->s_state & EXT2_ERROR_FS) {
err_printf(&fctx, "%s\n",
_("Errors detected; running e2fsck is required."));
@@ -4760,6 +4772,11 @@ int main(int argc, char *argv[])
_("Mounting read-only without recovering journal."));
fctx.ro = 1;
global_fs->flags &= ~EXT2_FLAG_RW;
+ } else if (fctx.ro && !(flags & EXT2_FLAG_RW)) {
+ err_printf(&fctx, "%s\n",
+ _("Journal needs recovery but filesystem could not be opened read-write."));
+ err_printf(&fctx, "%s\n", _("Please run e2fsck -fy."));
+ goto out;
} else {
log_printf(&fctx, "%s\n", _("Recovering journal."));
err = ext2fs_run_ext3_journal(&global_fs);
@@ -4833,8 +4850,10 @@ int main(int argc, char *argv[])
if (fctx.no_default_opts == 0)
fuse_opt_add_arg(&args, extra_args);
- if (fctx.ro)
+ if (fctx.ro) {
+ /* This is in case ro was implied above and not passed in */
fuse_opt_add_arg(&args, "-oro");
+ }
if (fctx.fakeroot) {
#ifdef HAVE_MOUNT_NODEV
--
2.43.5
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH] fuse2fs: open read-only when ro option and image non-writable
2025-10-16 20:02 [PATCH] fuse2fs: open read-only when ro option and image non-writable Dave Dykstra
@ 2025-10-17 19:24 ` Darrick J. Wong
2025-10-17 20:14 ` Dave Dykstra
0 siblings, 1 reply; 6+ messages in thread
From: Darrick J. Wong @ 2025-10-17 19:24 UTC (permalink / raw)
To: Dave Dykstra; +Cc: linux-ext4, Dave Dykstra
On Thu, Oct 16, 2025 at 03:02:06PM -0500, Dave Dykstra wrote:
> This opens the image read-only when an ro option is given and the
> image is not writable. If it is then found that a journal recovery
> is needed, an error is returned then.
>
> The ret value is set to 4 after the option checks so that if there's
> an error resulting in "goto out" it won't print an error about
> unrecognized options.
>
> Also submitted as PR https://github.com/tytso/e2fsprogs/pull/250
> for the issue https://github.com/tytso/e2fsprogs/issues/244.
>
> Replaces
> https://lore.kernel.org/linux-ext4/20251010214735.22683-1-dave.dykstra@cern.ch/T/#u
> https://lore.kernel.org/linux-ext4/175798064776.350013.6744611652039454651.stgit@frogsfrogsfrogs/#t
>
> Signed-off-by: Dave Dykstra <2129743+DrDaveD@users.noreply.github.com>
> ---
> misc/fuse2fs.c | 27 +++++++++++++++++++++++----
> 1 file changed, 23 insertions(+), 4 deletions(-)
>
> diff --git a/misc/fuse2fs.c b/misc/fuse2fs.c
> index cb5620c7..2ae2fc1a 100644
> --- a/misc/fuse2fs.c
> +++ b/misc/fuse2fs.c
> @@ -4696,9 +4696,19 @@ int main(int argc, char *argv[])
> err = ext2fs_open2(fctx.device, options, flags, 0, 0, unix_io_manager,
> &global_fs);
> if (err) {
> - err_printf(&fctx, "%s.\n", error_message(err));
> - err_printf(&fctx, "%s\n", _("Please run e2fsck -fy."));
> - goto out;
> + if (((err == EACCES) || (err == EPERM)) && fctx.ro) {
This is not correct. mount(8) for the kernel ext4 driver responds to
the block device being readonly by retrying with an ro mount. The user
is not required to specify 'ro':
# blockdev --setro/ dev/sda
# strace -e mount mount /dev/sda /mnt
40677<mount> 12:19:36 (+ 0.000102) mount("/dev/sda", "/mnt", "ext4", 0, NULL) = -1 EACCES (Permission denied)
40677<mount> 12:19:36 (+ 0.000285) mount("/dev/sda", "/mnt", "ext4", MS_RDONLY, NULL) = 0
> + // read-only requested and don't have write access
> + dbg_printf(&fctx, "%s: %s\n", __func__,
> + _("Permission denied with writable, trying without.\n"));
> + flags &= ~EXT2_FLAG_RW;
> + err = ext2fs_open2(fctx.device, options, flags, 0, 0,
> + unix_io_manager, &global_fs);
> + }
> + if (err) {
> + err_printf(&fctx, "%s.\n", error_message(err));
> + err_printf(&fctx, "%s\n", _("Please run e2fsck -fy."));
> + goto out;
> + }
> }
> fctx.fs = global_fs;
> global_fs->priv_data = &fctx;
> @@ -4741,6 +4751,8 @@ int main(int argc, char *argv[])
> goto out;
> }
>
> + ret = 4;
Why 4? Is this an internal mount bug?
--D
> +
> if (global_fs->super->s_state & EXT2_ERROR_FS) {
> err_printf(&fctx, "%s\n",
> _("Errors detected; running e2fsck is required."));
> @@ -4760,6 +4772,11 @@ int main(int argc, char *argv[])
> _("Mounting read-only without recovering journal."));
> fctx.ro = 1;
> global_fs->flags &= ~EXT2_FLAG_RW;
> + } else if (fctx.ro && !(flags & EXT2_FLAG_RW)) {
> + err_printf(&fctx, "%s\n",
> + _("Journal needs recovery but filesystem could not be opened read-write."));
> + err_printf(&fctx, "%s\n", _("Please run e2fsck -fy."));
> + goto out;
> } else {
> log_printf(&fctx, "%s\n", _("Recovering journal."));
> err = ext2fs_run_ext3_journal(&global_fs);
> @@ -4833,8 +4850,10 @@ int main(int argc, char *argv[])
> if (fctx.no_default_opts == 0)
> fuse_opt_add_arg(&args, extra_args);
>
> - if (fctx.ro)
> + if (fctx.ro) {
> + /* This is in case ro was implied above and not passed in */
> fuse_opt_add_arg(&args, "-oro");
> + }
>
> if (fctx.fakeroot) {
> #ifdef HAVE_MOUNT_NODEV
> --
> 2.43.5
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] fuse2fs: open read-only when ro option and image non-writable
2025-10-17 19:24 ` Darrick J. Wong
@ 2025-10-17 20:14 ` Dave Dykstra
2025-10-17 20:48 ` Dave Dykstra
0 siblings, 1 reply; 6+ messages in thread
From: Dave Dykstra @ 2025-10-17 20:14 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: linux-ext4, Dave Dykstra
On Fri, Oct 17, 2025 at 12:24:56PM -0700, Darrick J. Wong wrote:
> On Thu, Oct 16, 2025 at 03:02:06PM -0500, Dave Dykstra wrote:
...
> > diff --git a/misc/fuse2fs.c b/misc/fuse2fs.c
> > index cb5620c7..2ae2fc1a 100644
> > --- a/misc/fuse2fs.c
> > +++ b/misc/fuse2fs.c
> > @@ -4696,9 +4696,19 @@ int main(int argc, char *argv[])
> > err = ext2fs_open2(fctx.device, options, flags, 0, 0, unix_io_manager,
> > &global_fs);
> > if (err) {
> > - err_printf(&fctx, "%s.\n", error_message(err));
> > - err_printf(&fctx, "%s\n", _("Please run e2fsck -fy."));
> > - goto out;
> > + if (((err == EACCES) || (err == EPERM)) && fctx.ro) {
>
> This is not correct. mount(8) for the kernel ext4 driver responds to
> the block device being readonly by retrying with an ro mount. The user
> is not required to specify 'ro':
>
> # blockdev --setro/ dev/sda
> # strace -e mount mount /dev/sda /mnt
> 40677<mount> 12:19:36 (+ 0.000102) mount("/dev/sda", "/mnt", "ext4", 0, NULL) = -1 EACCES (Permission denied)
> 40677<mount> 12:19:36 (+ 0.000285) mount("/dev/sda", "/mnt", "ext4", MS_RDONLY, NULL) = 0
Ok then, I'll change that.
...
> > fctx.fs = global_fs;
> > global_fs->priv_data = &fctx;
> > @@ -4741,6 +4751,8 @@ int main(int argc, char *argv[])
> > goto out;
> > }
> >
> > + ret = 4;
>
> Why 4? Is this an internal mount bug?
I was just guessing; I didn't know what the bits meant. Where are they
documented? All I knew was that it needed to not include the "& 1" bit
which was included with "ret = 3" because that printed an erroneous
error just after the "out" label.
Dave
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] fuse2fs: open read-only when ro option and image non-writable
2025-10-17 20:14 ` Dave Dykstra
@ 2025-10-17 20:48 ` Dave Dykstra
2025-10-17 21:16 ` Dave Dykstra
2025-10-17 23:28 ` Darrick J. Wong
0 siblings, 2 replies; 6+ messages in thread
From: Dave Dykstra @ 2025-10-17 20:48 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: linux-ext4, Dave Dykstra
On Fri, Oct 17, 2025 at 03:14:50PM -0500, Dave Dykstra wrote:
> On Fri, Oct 17, 2025 at 12:24:56PM -0700, Darrick J. Wong wrote:
...
> > > + ret = 4;
> >
> > Why 4? Is this an internal mount bug?
>
> I was just guessing; I didn't know what the bits meant. Where are they
> documented? All I knew was that it needed to not include the "& 1" bit
> which was included with "ret = 3" because that printed an erroneous
> error just after the "out" label.
Oh, it's supposed to follow the return codes from mount(8). I think
ret = 2 is most appropriate then.
Dave
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] fuse2fs: open read-only when ro option and image non-writable
2025-10-17 20:48 ` Dave Dykstra
@ 2025-10-17 21:16 ` Dave Dykstra
2025-10-17 23:28 ` Darrick J. Wong
1 sibling, 0 replies; 6+ messages in thread
From: Dave Dykstra @ 2025-10-17 21:16 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: linux-ext4
This patch is replaced by
https://lore.kernel.org/linux-ext4/20251017211130.8507-1-dave.dykstra@cern.ch/
Dave
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] fuse2fs: open read-only when ro option and image non-writable
2025-10-17 20:48 ` Dave Dykstra
2025-10-17 21:16 ` Dave Dykstra
@ 2025-10-17 23:28 ` Darrick J. Wong
1 sibling, 0 replies; 6+ messages in thread
From: Darrick J. Wong @ 2025-10-17 23:28 UTC (permalink / raw)
To: Dave Dykstra; +Cc: linux-ext4, Dave Dykstra
On Fri, Oct 17, 2025 at 03:48:32PM -0500, Dave Dykstra wrote:
> On Fri, Oct 17, 2025 at 03:14:50PM -0500, Dave Dykstra wrote:
> > On Fri, Oct 17, 2025 at 12:24:56PM -0700, Darrick J. Wong wrote:
> ...
> > > > + ret = 4;
> > >
> > > Why 4? Is this an internal mount bug?
> >
> > I was just guessing; I didn't know what the bits meant. Where are they
> > documented? All I knew was that it needed to not include the "& 1" bit
> > which was included with "ret = 3" because that printed an erroneous
> > error just after the "out" label.
>
> Oh, it's supposed to follow the return codes from mount(8). I think
> ret = 2 is most appropriate then.
Could you please spend more time understanding the context in which fuse
servers operate, particularly when they're called as mount(8) helpers?
--D
> Dave
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2025-10-17 23:28 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-10-16 20:02 [PATCH] fuse2fs: open read-only when ro option and image non-writable Dave Dykstra
2025-10-17 19:24 ` Darrick J. Wong
2025-10-17 20:14 ` Dave Dykstra
2025-10-17 20:48 ` Dave Dykstra
2025-10-17 21:16 ` Dave Dykstra
2025-10-17 23:28 ` Darrick J. Wong
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox