Do not use free space caching!
Hello,
Just found a big bug in the free space caching stuff that will result in early ENOSPC. I'm working on fixing this bug, but it won't be until tomorrow that I'll have it completely working, so for now make sure to mount -o clear_cache so that it just clears the cache and doesn't use it.
NOTE: It doesn't cause problems other than early ENOSPC, you won't get corruption or anything like that, tho you could possibly panic.
Sorry for the inconvenience. Thanks,
Josef
Quoted 11 lines · Josef Bacik
Hello,
Just found a big bug in the free space caching stuff that will result in early ENOSPC. I'm working on fixing this bug, but it won't be until tomorrow that I'll have it completely working, so for now make sure to mount -o
clear_cacheso that it just clears the cache and doesn't use it.NOTE: It doesn't cause problems other than early ENOSPC, you won't get corruption or anything like that, tho you could possibly panic.
Sorry for the inconvenience. Thanks,
Any chance you could provide a little more information about which kernels are affected? Is it any kernel with free space cache support (is 2.6.38.x included?) - and if so, do you plan on submitting the fix to the stable kernel series?
Signature · 1 line
Calvin Walton <[email protected]>
Quoted 11 lines · Calvin Walton
Quoted 10 lines · Josef Bacik
Hello,
Just found a big bug in the free space caching stuff that will result in early ENOSPC. I'm working on fixing this bug, but it won't be until tomorrow that I'll have it completely working, so for now make sure to mount -o
clear_cacheso that it just clears the cache and doesn't use it.NOTE: It doesn't cause problems other than early ENOSPC, you won't get corruption or anything like that, tho you could possibly panic.
Quoted 7 lines
Quoted 1 line
Sorry for the inconvenience. Thanks,
Any chance you could provide a little more information about which kernels are affected? Is it any kernel with free space cache support (is 2.6.38.x included?) - and if so, do you plan on submitting the fix to the stable kernel series?
Yeah it affects any kernel that has the free space cache feature, which I think started in .37. Course you have to have specifically enabled it, so it's not a huge problem. I've submitted a patch, but since it's currently an optional feature I don't think it needs to go to stable. Thanks,
Josef
Quoted 5 lines · Josef Bacik
Quoted 4 lines · Calvin Walton
Quoted 3 lines · Josef Bacik
Hello,
Just found a big bug in the free space caching stuff that will resu=
lt in
Quoted 1 line
Quoted 1 line
Quoted 1 line
early ENOSPC. I'm working on fixing this bug, but it won't be unti=
l
Quoted 1 line
Quoted 1 line
Quoted 1 line
tomorrow that I'll have it completely working, so for now make sure=
to
Quoted 1 line
Quoted 1 line
Quoted 1 line
mount -o
clear_cacheso that it just clears the cache and doesn't u=
se it.
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
NOTE: It doesn't cause problems other than early ENOSPC, you won't =
get
Quoted 5 lines
Quoted 5 lines
Quoted 3 lines
corruption or anything like that, tho you could possibly panic.
Sorry for the inconvenience. Thanks,
Any chance you could provide a little more information about which kernels are affected? Is it any kernel with free space cache support=
(is
Quoted 1 line
Quoted 1 line
2.6.38.x included?) - and if so, do you plan on submitting the fix t=
o
Quoted 3 lines
Quoted 2 lines
the stable kernel series?
Yeah it affects any kernel that has the free space cache feature, whi=
ch I think
Quoted 1 line
started in .37. Course you have to have specifically enabled it, so =
it's not a
Quoted 1 line
huge problem. I've submitted a patch, but since it's currently an op=
tional
Quoted 1 line
feature I don't think it needs to go to stable. Thanks,
So it will have to wait for 2.6.39? If possible please push it for=20 inclusion it in the next stable of 2.6.38, as 2.6.39 is a few months=20 away and i wont risk an early RC for my system
Thanks
Quoted 3 lines
Josef
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs=
" in
Quoted 2 lines
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" = in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Quoted 9 lines · Josef Bacik
Quoted 7 lines · Calvin Walton
Quoted 6 lines · Josef Bacik
Hello,
Just found a big bug in the free space caching stuff that will result in early ENOSPC. I'm working on fixing this bug, but it won't be until tomorrow that I'll have it completely working, so for now make sure to mount -o
clear_cacheso that it just clears the cache and doesn't use it.
Quoted 4 lines
Yeah it affects any kernel that has the free space cache feature, which I think started in .37. Course you have to have specifically enabled it, so it's not a huge problem. I've submitted a patch, but since it's currently an optional feature I don't think it needs to go to stable. Thanks,
If you don't think the full fix is appropriate for the stable kernel, what about sending a simple patch to stable that just disables the free space caching support?
That should be enough to stop any adventurous users (of the kind who don't read mailing lists) from trying the free space caching support, and possibly hitting this bug.
Signature · 1 line
Calvin Walton <[email protected]>
Excerpts from Calvin Walton's message of 2011-04-01 13:56:51 -0400:
Quoted 15 lines · Josef Bacik
Quoted 9 lines · Calvin Walton
Quoted 7 lines · Josef Bacik
Quoted 6 lines
Hello,
Just found a big bug in the free space caching stuff that will result in early ENOSPC. I'm working on fixing this bug, but it won't be until tomorrow that I'll have it completely working, so for now make sure to mount -o
clear_cacheso that it just clears the cache and doesn't use it.Quoted 4 lines
Yeah it affects any kernel that has the free space cache feature, which I think started in .37. Course you have to have specifically enabled it, so it's not a huge problem. I've submitted a patch, but since it's currently an optional feature I don't think it needs to go to stable. Thanks,
Quoted 8 lines
If you don't think the full fix is appropriate for the stable kernel, what about sending a simple patch to stable that just disables the free space caching support?
That should be enough to stop any adventurous users (of the kind who don't read mailing lists) from trying the free space caching support, and possibly hitting this bug.
We'll have a known stable patch for this shortly, I'd rather do that and send it back to the stable kernels.
-chris