* Re: no way to swapoff a deleted swap file?
@ 2008-10-17 12:17 ` Hugh Dickins
0 siblings, 0 replies; 29+ messages in thread
From: Hugh Dickins @ 2008-10-17 12:17 UTC (permalink / raw)
To: Bodo Eggert
Cc: David Newall, Peter Zijlstra, Peter Cordes, linux-kernel,
Christoph Hellwig, linux-mm
On Fri, 17 Oct 2008, Bodo Eggert wrote:
>
> Somebody might want their swapfiles to have zero links,
> _and_ the possibility of doing swapoff.
You're right, they might, and it's not an unreasonable wish.
But we've not supported it in the past, and I still don't
think it's worth adding special kernel support for it now.
> If you can do it by keeping some fds open to let
> /proc/pid/fd point to the files, I think it's OK.
I've a very strong aversion to adding strange code to abuse the
/proc/<pid>/fd space of some random kernel thread - a "kswapd"
because its name contains the substring "swap"?
Hugh
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: no way to swapoff a deleted swap file?
2008-10-17 12:17 ` Hugh Dickins
@ 2008-10-17 12:36 ` David Newall
-1 siblings, 0 replies; 29+ messages in thread
From: David Newall @ 2008-10-17 12:36 UTC (permalink / raw)
To: Hugh Dickins
Cc: Bodo Eggert, Peter Zijlstra, Peter Cordes, linux-kernel,
Christoph Hellwig, linux-mm
Hugh Dickins wrote:
> On Fri, 17 Oct 2008, Bodo Eggert wrote:
>
>> Somebody might want their swapfiles to have zero links,
>> _and_ the possibility of doing swapoff.
>>
>
> You're right, they might, and it's not an unreasonable wish.
> But we've not supported it in the past, and I still don't
> think it's worth adding special kernel support for it now.
But it is supported now. It's swapoff that's not supported, and I don't
think that matters.
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: no way to swapoff a deleted swap file?
@ 2008-10-17 12:36 ` David Newall
0 siblings, 0 replies; 29+ messages in thread
From: David Newall @ 2008-10-17 12:36 UTC (permalink / raw)
To: Hugh Dickins
Cc: Bodo Eggert, Peter Zijlstra, Peter Cordes, linux-kernel,
Christoph Hellwig, linux-mm
Hugh Dickins wrote:
> On Fri, 17 Oct 2008, Bodo Eggert wrote:
>
>> Somebody might want their swapfiles to have zero links,
>> _and_ the possibility of doing swapoff.
>>
>
> You're right, they might, and it's not an unreasonable wish.
> But we've not supported it in the past, and I still don't
> think it's worth adding special kernel support for it now.
But it is supported now. It's swapoff that's not supported, and I don't
think that matters.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: no way to swapoff a deleted swap file?
2008-10-17 12:17 ` Hugh Dickins
@ 2008-10-17 22:42 ` Bodo Eggert
-1 siblings, 0 replies; 29+ messages in thread
From: Bodo Eggert @ 2008-10-17 22:42 UTC (permalink / raw)
To: Hugh Dickins
Cc: Bodo Eggert, David Newall, Peter Zijlstra, Peter Cordes,
linux-kernel, Christoph Hellwig, linux-mm
On Fri, 17 Oct 2008, Hugh Dickins wrote:
> On Fri, 17 Oct 2008, Bodo Eggert wrote:
> > Somebody might want their swapfiles to have zero links,
> > _and_ the possibility of doing swapoff.
>
> You're right, they might, and it's not an unreasonable wish.
> But we've not supported it in the past, and I still don't
> think it's worth adding special kernel support for it now.
IMO it depends on the cost. Maybe it's cheap to keep an extra fd around,
maybe you'd have to add an extra infrastructure for this. And maybe it's not
important enough for anybody to create a patch and let us know ...
--
Funny quotes:
9. Despite the cost of living, have you noticed how popular it remains?
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: no way to swapoff a deleted swap file?
@ 2008-10-17 22:42 ` Bodo Eggert
0 siblings, 0 replies; 29+ messages in thread
From: Bodo Eggert @ 2008-10-17 22:42 UTC (permalink / raw)
To: Hugh Dickins
Cc: Bodo Eggert, David Newall, Peter Zijlstra, Peter Cordes,
linux-kernel, Christoph Hellwig, linux-mm
On Fri, 17 Oct 2008, Hugh Dickins wrote:
> On Fri, 17 Oct 2008, Bodo Eggert wrote:
> > Somebody might want their swapfiles to have zero links,
> > _and_ the possibility of doing swapoff.
>
> You're right, they might, and it's not an unreasonable wish.
> But we've not supported it in the past, and I still don't
> think it's worth adding special kernel support for it now.
IMO it depends on the cost. Maybe it's cheap to keep an extra fd around,
maybe you'd have to add an extra infrastructure for this. And maybe it's not
important enough for anybody to create a patch and let us know ...
--
Funny quotes:
9. Despite the cost of living, have you noticed how popular it remains?
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: no way to swapoff a deleted swap file?
2008-10-17 12:17 ` Hugh Dickins
@ 2008-10-18 0:31 ` Peter Cordes
-1 siblings, 0 replies; 29+ messages in thread
From: Peter Cordes @ 2008-10-18 0:31 UTC (permalink / raw)
To: Hugh Dickins
Cc: Bodo Eggert, David Newall, Peter Zijlstra, linux-kernel,
Christoph Hellwig, linux-mm
On Fri, Oct 17, 2008 at 01:17:17PM +0100, Hugh Dickins wrote:
> On Fri, 17 Oct 2008, Bodo Eggert wrote:
> >
> > Somebody might want their swapfiles to have zero links,
> > _and_ the possibility of doing swapoff.
>
> You're right, they might, and it's not an unreasonable wish.
> But we've not supported it in the past, and I still don't
> think it's worth adding special kernel support for it now.
I'd be inclined to agree with not bloating the kernel to support
this, even though it would have been convenient for me in one case. I
do have an idea for supporting this without bloat, see below. In case
anyone wants more details about how I painted myself into that corner,
here's the backstory to my feature request.
I was cleaning out /var/tmp until I had it down to everything I
wanted to keep. Then I was going to rsync it to somewhere else,
umount, mkfs (with a newer mkfs.xfs -l version=2,lazy-count=1 -n
version=2, etc...), and copy my files back. I think I forgot to
/etc/init.d/swapspace stop before rm -r on my swapfile directory.
I run swapspace(8), http://pqxx.org/development/swapspace
http://packages.debian.org/sid/swapspace. I have a ~700MB swap
partition that will actually get used most of the time, and let
swapspace(8) make swap files in /var/cache/swapspace. (I symlink or
bind mount /var/cache so it's actually on /var/tmp, since it seems to
fit better there.)
Usually dynamic swap file creation just delays the inevitable OOM and
makes thrashing worse. (because one can't create swap files without
writing them full of zeros, not even with xfs_io resvsp, except with
old XFS without unwritten extent tracking. But that's another feature
request...) In the rare case where you're running a few bloated things
and want to let them swap out while doing something unusual that takes
almost all your RAM, it's fairly convenient. Basically I like the
idea, even if in practice it's clunky and would be better to just use
some big swap files I could delete if I need the disk space. It works
well when your swap isn't filling up as fast as your disk can write.
> > If you can do it by keeping some fds open to let
> > /proc/pid/fd point to the files, I think it's OK.
>
> I've a very strong aversion to adding strange code to abuse the
> /proc/<pid>/fd space of some random kernel thread - a "kswapd"
> because its name contains the substring "swap"?
Yes, because it has swap in the name :P. I was just guessing it had
something to do with swap areas.
Here's another idea: swapoff(2) takes a text string. It could be
overloaded to parse the string as a numeric index into the list of
swap areas (as listed in /proc/swaps). It could do this as a fallback
iff the string was numeric and didn't exist as a pathname relative to
the CWD. (You don't want "/nonexistant/file" to be treated the same as
"0" of course, or you'd always be swapoff()ing the first swap file
instead of returning ENOENT.)
Before rebooting I tried
# swapoff "/var/tmp/EXP/cache/swap/1 (deleted)"
Then I looked at the source to see that that sys_swapoff only tries
to use the string as a path in the filesystem. But that doesn't have
to be the case, if there's enough information in data structures
associated with a swap area to get whatever swapoff() needs to let go
of the file.
This would just add some kernel code that wouldn't usually be in the
CPU icache, and wouldn't allocate any extra objects like file
descriptors or VFS entries to lie around to support this.
It's a little ugly, though, to change a system call that normally
accepts only paths to accept things that aren't paths. It would be
even worse to use swapoff("/sys/swaps/2") if /sys/swaps didn't exist
in the VFS, and was parsed specially by sys_swapoff(). So don't do
that. :) swapoff("0") doesn't seem too bad, if only by comparison
with something horrible.
--
#define X(x,y) x##y
Peter Cordes ; e-mail: X(peter@cor , des.ca)
"The gods confound the man who first found out how to distinguish the hours!
Confound him, too, who in this place set up a sundial, to cut and hack
my day so wretchedly into small pieces!" -- Plautus, 200 BC
^ permalink raw reply [flat|nested] 29+ messages in thread* Re: no way to swapoff a deleted swap file?
@ 2008-10-18 0:31 ` Peter Cordes
0 siblings, 0 replies; 29+ messages in thread
From: Peter Cordes @ 2008-10-18 0:31 UTC (permalink / raw)
To: Hugh Dickins
Cc: Bodo Eggert, David Newall, Peter Zijlstra, linux-kernel,
Christoph Hellwig, linux-mm
On Fri, Oct 17, 2008 at 01:17:17PM +0100, Hugh Dickins wrote:
> On Fri, 17 Oct 2008, Bodo Eggert wrote:
> >
> > Somebody might want their swapfiles to have zero links,
> > _and_ the possibility of doing swapoff.
>
> You're right, they might, and it's not an unreasonable wish.
> But we've not supported it in the past, and I still don't
> think it's worth adding special kernel support for it now.
I'd be inclined to agree with not bloating the kernel to support
this, even though it would have been convenient for me in one case. I
do have an idea for supporting this without bloat, see below. In case
anyone wants more details about how I painted myself into that corner,
here's the backstory to my feature request.
I was cleaning out /var/tmp until I had it down to everything I
wanted to keep. Then I was going to rsync it to somewhere else,
umount, mkfs (with a newer mkfs.xfs -l version=2,lazy-count=1 -n
version=2, etc...), and copy my files back. I think I forgot to
/etc/init.d/swapspace stop before rm -r on my swapfile directory.
I run swapspace(8), http://pqxx.org/development/swapspace
http://packages.debian.org/sid/swapspace. I have a ~700MB swap
partition that will actually get used most of the time, and let
swapspace(8) make swap files in /var/cache/swapspace. (I symlink or
bind mount /var/cache so it's actually on /var/tmp, since it seems to
fit better there.)
Usually dynamic swap file creation just delays the inevitable OOM and
makes thrashing worse. (because one can't create swap files without
writing them full of zeros, not even with xfs_io resvsp, except with
old XFS without unwritten extent tracking. But that's another feature
request...) In the rare case where you're running a few bloated things
and want to let them swap out while doing something unusual that takes
almost all your RAM, it's fairly convenient. Basically I like the
idea, even if in practice it's clunky and would be better to just use
some big swap files I could delete if I need the disk space. It works
well when your swap isn't filling up as fast as your disk can write.
> > If you can do it by keeping some fds open to let
> > /proc/pid/fd point to the files, I think it's OK.
>
> I've a very strong aversion to adding strange code to abuse the
> /proc/<pid>/fd space of some random kernel thread - a "kswapd"
> because its name contains the substring "swap"?
Yes, because it has swap in the name :P. I was just guessing it had
something to do with swap areas.
Here's another idea: swapoff(2) takes a text string. It could be
overloaded to parse the string as a numeric index into the list of
swap areas (as listed in /proc/swaps). It could do this as a fallback
iff the string was numeric and didn't exist as a pathname relative to
the CWD. (You don't want "/nonexistant/file" to be treated the same as
"0" of course, or you'd always be swapoff()ing the first swap file
instead of returning ENOENT.)
Before rebooting I tried
# swapoff "/var/tmp/EXP/cache/swap/1 (deleted)"
Then I looked at the source to see that that sys_swapoff only tries
to use the string as a path in the filesystem. But that doesn't have
to be the case, if there's enough information in data structures
associated with a swap area to get whatever swapoff() needs to let go
of the file.
This would just add some kernel code that wouldn't usually be in the
CPU icache, and wouldn't allocate any extra objects like file
descriptors or VFS entries to lie around to support this.
It's a little ugly, though, to change a system call that normally
accepts only paths to accept things that aren't paths. It would be
even worse to use swapoff("/sys/swaps/2") if /sys/swaps didn't exist
in the VFS, and was parsed specially by sys_swapoff(). So don't do
that. :) swapoff("0") doesn't seem too bad, if only by comparison
with something horrible.
--
#define X(x,y) x##y
Peter Cordes ; e-mail: X(peter@cor , des.ca)
"The gods confound the man who first found out how to distinguish the hours!
Confound him, too, who in this place set up a sundial, to cut and hack
my day so wretchedly into small pieces!" -- Plautus, 200 BC
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 29+ messages in thread* Re: no way to swapoff a deleted swap file?
2008-10-18 0:31 ` Peter Cordes
@ 2008-10-18 5:18 ` Willy Tarreau
-1 siblings, 0 replies; 29+ messages in thread
From: Willy Tarreau @ 2008-10-18 5:18 UTC (permalink / raw)
To: Peter Cordes
Cc: Hugh Dickins, Bodo Eggert, David Newall, Peter Zijlstra,
linux-kernel, Christoph Hellwig, linux-mm
On Fri, Oct 17, 2008 at 09:31:17PM -0300, Peter Cordes wrote:
> On Fri, Oct 17, 2008 at 01:17:17PM +0100, Hugh Dickins wrote:
> > On Fri, 17 Oct 2008, Bodo Eggert wrote:
> > >
> > > Somebody might want their swapfiles to have zero links,
> > > _and_ the possibility of doing swapoff.
> >
> > You're right, they might, and it's not an unreasonable wish.
> > But we've not supported it in the past, and I still don't
> > think it's worth adding special kernel support for it now.
>
> I'd be inclined to agree with not bloating the kernel to support
> this, even though it would have been convenient for me in one case. I
> do have an idea for supporting this without bloat, see below. In case
> anyone wants more details about how I painted myself into that corner,
> here's the backstory to my feature request.
(...)
I have another idea which might be simpler to implement in userspace.
What happened to you is a typical accident, you did not run on purpose
on a deleted swap file. So we should at least ensure that such types
of accidents could not happen easily.
If swapon did set the immutable bit on a file just after enabling swap
to it, it would at least prevent accidental removal of that file. Swapoff
would have to clean that bit, and swapon would have to clean it upon
startup too (in case of unplanned reboots).
That way, you could still remove such files on purpose provided you do
a preliminary "chattr -i" on them, but "rm -rf" would keep them intact.
It would also prevent accidental modifications, such as "ls .>swapfile"
instead of "ls ./swapfile".
Regards,
Willy
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: no way to swapoff a deleted swap file?
@ 2008-10-18 5:18 ` Willy Tarreau
0 siblings, 0 replies; 29+ messages in thread
From: Willy Tarreau @ 2008-10-18 5:18 UTC (permalink / raw)
To: Peter Cordes
Cc: Hugh Dickins, Bodo Eggert, David Newall, Peter Zijlstra,
linux-kernel, Christoph Hellwig, linux-mm
On Fri, Oct 17, 2008 at 09:31:17PM -0300, Peter Cordes wrote:
> On Fri, Oct 17, 2008 at 01:17:17PM +0100, Hugh Dickins wrote:
> > On Fri, 17 Oct 2008, Bodo Eggert wrote:
> > >
> > > Somebody might want their swapfiles to have zero links,
> > > _and_ the possibility of doing swapoff.
> >
> > You're right, they might, and it's not an unreasonable wish.
> > But we've not supported it in the past, and I still don't
> > think it's worth adding special kernel support for it now.
>
> I'd be inclined to agree with not bloating the kernel to support
> this, even though it would have been convenient for me in one case. I
> do have an idea for supporting this without bloat, see below. In case
> anyone wants more details about how I painted myself into that corner,
> here's the backstory to my feature request.
(...)
I have another idea which might be simpler to implement in userspace.
What happened to you is a typical accident, you did not run on purpose
on a deleted swap file. So we should at least ensure that such types
of accidents could not happen easily.
If swapon did set the immutable bit on a file just after enabling swap
to it, it would at least prevent accidental removal of that file. Swapoff
would have to clean that bit, and swapon would have to clean it upon
startup too (in case of unplanned reboots).
That way, you could still remove such files on purpose provided you do
a preliminary "chattr -i" on them, but "rm -rf" would keep them intact.
It would also prevent accidental modifications, such as "ls .>swapfile"
instead of "ls ./swapfile".
Regards,
Willy
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: no way to swapoff a deleted swap file?
2008-10-18 5:18 ` Willy Tarreau
@ 2008-10-18 20:45 ` Hugh Dickins
-1 siblings, 0 replies; 29+ messages in thread
From: Hugh Dickins @ 2008-10-18 20:45 UTC (permalink / raw)
To: Willy Tarreau
Cc: Peter Cordes, Bodo Eggert, David Newall, Peter Zijlstra,
linux-kernel, Christoph Hellwig, linux-mm
On Sat, 18 Oct 2008, Willy Tarreau wrote:
> (...)
> I have another idea which might be simpler to implement in userspace.
> What happened to you is a typical accident, you did not run on purpose
> on a deleted swap file. So we should at least ensure that such types
> of accidents could not happen easily.
>
> If swapon did set the immutable bit on a file just after enabling swap
> to it, it would at least prevent accidental removal of that file. Swapoff
> would have to clean that bit, and swapon would have to clean it upon
> startup too (in case of unplanned reboots).
>
> That way, you could still remove such files on purpose provided you do
> a preliminary "chattr -i" on them, but "rm -rf" would keep them intact.
That's a good idea, thank you Willy:
much more to my taste than previous suggestions.
But I'm still not tempted to build it into the swapon and swapoff,
neither at the command nor at the kernel level. Let's leave it as
advice to sufferers on how to address the issue if it troubles them.
I did play with immutable on swapfiles back around 2.6.8. Prior
to that we left i_sem downed on a swapfile to protect it against
truncation (freeing its pages!) while in use - which caused
anyone idly touching it to hang, not very nice.
I experimented with setting immutable in sys_swapon, clearing it
in sys_swapoff, but I see from old mails that I didn't find that
satisfactory: I haven't actually recorded why not, but I think it
was partly a difficulty in getting the locking right, and partly
what happened if the user also made it immutable while swapped on -
oh, yes, and immutable gets written back to the filesystem which is
inconvenient when we crash, as you observe. So we ended up adding
an additional swapfile flag just at the VFS level.
Hmm, I suppose it would be very easy to make that additional swapfile
flag prohibit unlinking just as immutable does: patch below should do
that. What do you guys think - should we include this? It does then
(barring races which I don't propose to worry about) make an unlinked
swapfile impossible, which earlier had seemed a reasonable option.
> It would also prevent accidental modifications, such as "ls .>swapfile"
> instead of "ls ./swapfile".
That we do already protect against with the swapfile flag: we don't
actually protect against writing (that's just a permission thing,
same as when swapping to block device), but we do protect against
truncation, which would otherwise end up with swap corrupting
blocks of other files.
Hugh
--- 2.6.27/fs/namei.c 2008-10-09 23:13:53.000000000 +0100
+++ linux/fs/namei.c 2008-10-18 21:33:01.000000000 +0100
@@ -1407,7 +1407,7 @@ static int may_delete(struct inode *dir,
if (IS_APPEND(dir))
return -EPERM;
if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
- IS_IMMUTABLE(victim->d_inode))
+ IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
return -EPERM;
if (isdir) {
if (!S_ISDIR(victim->d_inode->i_mode))
^ permalink raw reply [flat|nested] 29+ messages in thread* Re: no way to swapoff a deleted swap file?
@ 2008-10-18 20:45 ` Hugh Dickins
0 siblings, 0 replies; 29+ messages in thread
From: Hugh Dickins @ 2008-10-18 20:45 UTC (permalink / raw)
To: Willy Tarreau
Cc: Peter Cordes, Bodo Eggert, David Newall, Peter Zijlstra,
linux-kernel, Christoph Hellwig, linux-mm
On Sat, 18 Oct 2008, Willy Tarreau wrote:
> (...)
> I have another idea which might be simpler to implement in userspace.
> What happened to you is a typical accident, you did not run on purpose
> on a deleted swap file. So we should at least ensure that such types
> of accidents could not happen easily.
>
> If swapon did set the immutable bit on a file just after enabling swap
> to it, it would at least prevent accidental removal of that file. Swapoff
> would have to clean that bit, and swapon would have to clean it upon
> startup too (in case of unplanned reboots).
>
> That way, you could still remove such files on purpose provided you do
> a preliminary "chattr -i" on them, but "rm -rf" would keep them intact.
That's a good idea, thank you Willy:
much more to my taste than previous suggestions.
But I'm still not tempted to build it into the swapon and swapoff,
neither at the command nor at the kernel level. Let's leave it as
advice to sufferers on how to address the issue if it troubles them.
I did play with immutable on swapfiles back around 2.6.8. Prior
to that we left i_sem downed on a swapfile to protect it against
truncation (freeing its pages!) while in use - which caused
anyone idly touching it to hang, not very nice.
I experimented with setting immutable in sys_swapon, clearing it
in sys_swapoff, but I see from old mails that I didn't find that
satisfactory: I haven't actually recorded why not, but I think it
was partly a difficulty in getting the locking right, and partly
what happened if the user also made it immutable while swapped on -
oh, yes, and immutable gets written back to the filesystem which is
inconvenient when we crash, as you observe. So we ended up adding
an additional swapfile flag just at the VFS level.
Hmm, I suppose it would be very easy to make that additional swapfile
flag prohibit unlinking just as immutable does: patch below should do
that. What do you guys think - should we include this? It does then
(barring races which I don't propose to worry about) make an unlinked
swapfile impossible, which earlier had seemed a reasonable option.
> It would also prevent accidental modifications, such as "ls .>swapfile"
> instead of "ls ./swapfile".
That we do already protect against with the swapfile flag: we don't
actually protect against writing (that's just a permission thing,
same as when swapping to block device), but we do protect against
truncation, which would otherwise end up with swap corrupting
blocks of other files.
Hugh
--- 2.6.27/fs/namei.c 2008-10-09 23:13:53.000000000 +0100
+++ linux/fs/namei.c 2008-10-18 21:33:01.000000000 +0100
@@ -1407,7 +1407,7 @@ static int may_delete(struct inode *dir,
if (IS_APPEND(dir))
return -EPERM;
if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
- IS_IMMUTABLE(victim->d_inode))
+ IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
return -EPERM;
if (isdir) {
if (!S_ISDIR(victim->d_inode->i_mode))
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 29+ messages in thread* Re: no way to swapoff a deleted swap file?
2008-10-18 20:45 ` Hugh Dickins
@ 2008-10-18 20:49 ` Christoph Hellwig
-1 siblings, 0 replies; 29+ messages in thread
From: Christoph Hellwig @ 2008-10-18 20:49 UTC (permalink / raw)
To: Hugh Dickins
Cc: Willy Tarreau, Peter Cordes, Bodo Eggert, David Newall,
Peter Zijlstra, linux-kernel, Christoph Hellwig, linux-mm
On Sat, Oct 18, 2008 at 09:45:14PM +0100, Hugh Dickins wrote:
> --- 2.6.27/fs/namei.c 2008-10-09 23:13:53.000000000 +0100
> +++ linux/fs/namei.c 2008-10-18 21:33:01.000000000 +0100
> @@ -1407,7 +1407,7 @@ static int may_delete(struct inode *dir,
> if (IS_APPEND(dir))
> return -EPERM;
> if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
> - IS_IMMUTABLE(victim->d_inode))
> + IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
> return -EPERM;
> if (isdir) {
> if (!S_ISDIR(victim->d_inode->i_mode))
Looks reasonable.
^ permalink raw reply [flat|nested] 29+ messages in thread* Re: no way to swapoff a deleted swap file?
@ 2008-10-18 20:49 ` Christoph Hellwig
0 siblings, 0 replies; 29+ messages in thread
From: Christoph Hellwig @ 2008-10-18 20:49 UTC (permalink / raw)
To: Hugh Dickins
Cc: Willy Tarreau, Peter Cordes, Bodo Eggert, David Newall,
Peter Zijlstra, linux-kernel, Christoph Hellwig, linux-mm
On Sat, Oct 18, 2008 at 09:45:14PM +0100, Hugh Dickins wrote:
> --- 2.6.27/fs/namei.c 2008-10-09 23:13:53.000000000 +0100
> +++ linux/fs/namei.c 2008-10-18 21:33:01.000000000 +0100
> @@ -1407,7 +1407,7 @@ static int may_delete(struct inode *dir,
> if (IS_APPEND(dir))
> return -EPERM;
> if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
> - IS_IMMUTABLE(victim->d_inode))
> + IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
> return -EPERM;
> if (isdir) {
> if (!S_ISDIR(victim->d_inode->i_mode))
Looks reasonable.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 29+ messages in thread* Re: no way to swapoff a deleted swap file?
2008-10-18 20:49 ` Christoph Hellwig
@ 2008-10-18 20:56 ` Willy Tarreau
-1 siblings, 0 replies; 29+ messages in thread
From: Willy Tarreau @ 2008-10-18 20:56 UTC (permalink / raw)
To: Christoph Hellwig
Cc: Hugh Dickins, Peter Cordes, Bodo Eggert, David Newall,
Peter Zijlstra, linux-kernel, linux-mm
On Sat, Oct 18, 2008 at 04:49:48PM -0400, Christoph Hellwig wrote:
> On Sat, Oct 18, 2008 at 09:45:14PM +0100, Hugh Dickins wrote:
> > --- 2.6.27/fs/namei.c 2008-10-09 23:13:53.000000000 +0100
> > +++ linux/fs/namei.c 2008-10-18 21:33:01.000000000 +0100
> > @@ -1407,7 +1407,7 @@ static int may_delete(struct inode *dir,
> > if (IS_APPEND(dir))
> > return -EPERM;
> > if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
> > - IS_IMMUTABLE(victim->d_inode))
> > + IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
> > return -EPERM;
> > if (isdir) {
> > if (!S_ISDIR(victim->d_inode->i_mode))
>
> Looks reasonable.
I like the idea and the simplicity a lot !
Willy
^ permalink raw reply [flat|nested] 29+ messages in thread* Re: no way to swapoff a deleted swap file?
@ 2008-10-18 20:56 ` Willy Tarreau
0 siblings, 0 replies; 29+ messages in thread
From: Willy Tarreau @ 2008-10-18 20:56 UTC (permalink / raw)
To: Christoph Hellwig
Cc: Hugh Dickins, Peter Cordes, Bodo Eggert, David Newall,
Peter Zijlstra, linux-kernel, linux-mm
On Sat, Oct 18, 2008 at 04:49:48PM -0400, Christoph Hellwig wrote:
> On Sat, Oct 18, 2008 at 09:45:14PM +0100, Hugh Dickins wrote:
> > --- 2.6.27/fs/namei.c 2008-10-09 23:13:53.000000000 +0100
> > +++ linux/fs/namei.c 2008-10-18 21:33:01.000000000 +0100
> > @@ -1407,7 +1407,7 @@ static int may_delete(struct inode *dir,
> > if (IS_APPEND(dir))
> > return -EPERM;
> > if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
> > - IS_IMMUTABLE(victim->d_inode))
> > + IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
> > return -EPERM;
> > if (isdir) {
> > if (!S_ISDIR(victim->d_inode->i_mode))
>
> Looks reasonable.
I like the idea and the simplicity a lot !
Willy
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 29+ messages in thread