Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Linux 6.7 compat #15681

Closed
wants to merge 4 commits into from
Closed

Linux 6.7 compat #15681

wants to merge 4 commits into from

Conversation

robn
Copy link
Member

@robn robn commented Dec 16, 2023

Motivation and Context

Upcoming Linux 6.7 release will change a couple of interfaces. Adapt or die!

Fixes: #15518
Fixes: #15582

Description

Adjusts for three changes:

  • i_atime and i_mtime in struct inode are no longer directly accessible, but must use accessor functions
  • s_shrink in struct super_block is now a pointer rather than an embedded struct
  • The "shrinker" (ie inode reclaim) API has changed, such that the kernel now allocates shrinkers dynamically rather than allowing a static shrinker to be registered.

How Has This Been Tested?

Compiled against 6.7-rc5, 6.7-rc6 and 5.10.170 and light sanity checks done (create pool, write some data, scrub it, echo 3 > /proc/sys/vm/drop_caches).

No test runs yet. Also have not attempted to compile or check on 3.0-3.11 (HAVE_SINGLE_SHRINKER_CALLBACK). (I'll run more tests soon enough; I just wanted to get this up before I disappear for the night).

Types of changes

  • Bug fix (non-breaking change which fixes an issue)
  • New feature (non-breaking change which adds functionality)
  • Performance enhancement (non-breaking change which improves efficiency)
  • Code cleanup (non-breaking change which makes code smaller or more readable)
  • Breaking change (fix or feature that would cause existing functionality to change)
  • Library ABI change (libzfs, libzfs_core, libnvpair, libuutil and libzfsbootenv)
  • Documentation (a change to man pages or other documentation)

Checklist:

@philmmanjaro
Copy link

This doesn't compile when applied on top of zfs-dkms 2.2.2 and errors out regarding current_time issues:

In file included from /build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/zfs/sys/zpl.h:36,
                 from /build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/zfs/sys/policy.h:36,
                 from /build/linux67-zfs/src/zfs/2.2.2/build/module/zfs/dmu_objset.c:62:
/build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/kernel/linux/vfs_compat.h:419:1: error: return type is an incomplete type
  419 | current_time(struct inode *ip)
      | ^~~~~~~~~~~~
/build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/kernel/linux/vfs_compat.h:419:1: error: conflicting types for 'current_time'; have 'void(struct inode *)'
In file included from ./include/linux/compat.h:17,
                 from ./arch/x86/include/asm/ia32.h:8,
                 from ./arch/x86/include/asm/elf.h:10,
                 from ./include/linux/elf.h:6,
                 from ./include/linux/module.h:19,
                 from /build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/spl/sys/cred.h:27,
                 from /build/linux67-zfs/src/zfs/2.2.2/build/module/zfs/dmu_objset.c:40:
./include/linux/fs.h:1516:19: note: previous declaration of 'current_time' with type 'struct timespec64(struct inode *)'
 1516 | struct timespec64 current_time(struct inode *inode);
      |                   ^~~~~~~~~~~~
/build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/kernel/linux/vfs_compat.h: In function 'current_time':
/build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/kernel/linux/vfs_compat.h:421:17: error: implicit declaration of function 'timespec_trunc'; did you mean 'timespec64_to_ns'? [-Werror=implicit-function-declaration]
  421 |         return (timespec_trunc(current_kernel_time(), ip->i_sb->s_time_gran));
      |                 ^~~~~~~~~~~~~~
      |                 timespec64_to_ns
/build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/kernel/linux/vfs_compat.h:421:32: error: implicit declaration of function 'current_kernel_time'; did you mean 'current_time'? [-Werror=implicit-function-declaration]
  421 |         return (timespec_trunc(current_kernel_time(), ip->i_sb->s_time_gran));
      |                                ^~~~~~~~~~~~~~~~~~~
      |                                current_time
/build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/kernel/linux/vfs_compat.h:421:17: error: 'return' with a value, in function returning void [-Werror=return-type]
  421 |         return (timespec_trunc(current_kernel_time(), ip->i_sb->s_time_gran));
      |                ~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/build/linux67-zfs/src/zfs/2.2.2/build/include/os/linux/kernel/linux/vfs_compat.h:419:1: note: declared here
  419 | current_time(struct inode *ip)
      | ^~~~~~~~~~~~
cc1: some warnings being treated as errors

Build log: make.log

@robn
Copy link
Member Author

robn commented Dec 17, 2023

@philmmanjaro here's linux-6.7.compat-2.2. This builds against 5.10.170, and 6.7-rc5, and a deb-dkms package built and installed fine against 6.1.52 on a Debian 12 system. So I'm gonna need more info: system you build the dkms package on, how you did it, and then system and kernel you're trying to build and install it on.

@spikerguy
Copy link

Hi @robn
I tried the 4 patches on zfs-dkms-git as well as zfs-dkms-2.2.2 both failed to compile.
Logs attached.

zfs-2.2.2-aarch64-6.7-rc3.txt

zfs-dkms-git--aarch64-6.7-rc3.txt

If you need any other detailed please feel free to ask.

Thanks.

@robn
Copy link
Member Author

robn commented Dec 17, 2023

Ahh, sorta maybe seeing what's up. Can you post your build/build.log.kabi please? This is the output from the configure checks.

@spikerguy
Copy link

Hello
Sure.

Here:
zfs-2.2.2.-aarch64-6.7-rc3-kabi.log

@robn
Copy link
Member Author

robn commented Dec 17, 2023

Thanks! The change in cdc0e7f is implicated, and though I don't entirely see why, I did think it felt brittle when I wrote it. I'll redo it a better way.

robn added 4 commits December 18, 2023 09:50
6.7 changed the names of the time members in struct inode, so we can't
assign back to it because we don't know its name. In practice this
doesn't matter though - if we're missing current_time(), then we must be
on <4.9, and we know our fallback will need to return timespec.

Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
6.6 made i_ctime inaccessible; 6.7 has done the same for i_atime and
i_mtime. This extends the method used for ctime in b37f293 to atime
and mtime as well.

Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
In 6.7 the superblock shrinker member s_shrink has changed from being an
embedded struct to a pointer. Detect this, and don't take a reference if
it already is one.

Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
6.7 changes the shrinker API such that shrinkers must be allocated
dynamically by the kernel. To accomodate this, this commit reworks
spl_register_shrinker() to do something similar against earlier kernels.

Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
@robn
Copy link
Member Author

robn commented Dec 17, 2023

I pushed a different way to do the current_time() check. See if that does better for you!

@philmmanjaro
Copy link

Works now also on my end: zfs-linux67.txt

@spikerguy
Copy link

Hey @robn
Thanks, It compiled fine on aarch64 using 6.7-rc3.

@JohnyPeaN
Copy link

Can confirm. Compiles and runs fine on 6.7-rc5.

@behlendorf behlendorf added the Status: Code Review Needed Ready for review and testing label Dec 20, 2023
@behlendorf behlendorf added Status: Accepted Ready to integrate (reviewed, tested) and removed Status: Code Review Needed Ready for review and testing labels Dec 20, 2023
behlendorf pushed a commit that referenced this pull request Dec 20, 2023
6.6 made i_ctime inaccessible; 6.7 has done the same for i_atime and
i_mtime. This extends the method used for ctime in b37f293 to atime
and mtime as well.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes #15681
behlendorf pushed a commit that referenced this pull request Dec 20, 2023
In 6.7 the superblock shrinker member s_shrink has changed from being an
embedded struct to a pointer. Detect this, and don't take a reference if
it already is one.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes #15681
behlendorf pushed a commit that referenced this pull request Dec 20, 2023
6.7 changes the shrinker API such that shrinkers must be allocated
dynamically by the kernel. To accomodate this, this commit reworks
spl_register_shrinker() to do something similar against earlier kernels.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes #15681
@robn robn mentioned this pull request Dec 21, 2023
13 tasks
stgraber pushed a commit to zabbly/zfs that referenced this pull request Jan 24, 2024
6.7 changed the names of the time members in struct inode, so we can't
assign back to it because we don't know its name. In practice this
doesn't matter though - if we're missing current_time(), then we must be
on <4.9, and we know our fallback will need to return timespec.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
stgraber pushed a commit to zabbly/zfs that referenced this pull request Jan 24, 2024
6.6 made i_ctime inaccessible; 6.7 has done the same for i_atime and
i_mtime. This extends the method used for ctime in b37f293 to atime
and mtime as well.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
stgraber pushed a commit to zabbly/zfs that referenced this pull request Jan 24, 2024
In 6.7 the superblock shrinker member s_shrink has changed from being an
embedded struct to a pointer. Detect this, and don't take a reference if
it already is one.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
stgraber pushed a commit to zabbly/zfs that referenced this pull request Jan 24, 2024
6.7 changes the shrinker API such that shrinkers must be allocated
dynamically by the kernel. To accomodate this, this commit reworks
spl_register_shrinker() to do something similar against earlier kernels.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
@robn robn mentioned this pull request Feb 7, 2024
13 tasks
lundman pushed a commit to openzfsonwindows/openzfs that referenced this pull request Mar 13, 2024
6.7 changed the names of the time members in struct inode, so we can't
assign back to it because we don't know its name. In practice this
doesn't matter though - if we're missing current_time(), then we must be
on <4.9, and we know our fallback will need to return timespec.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
lundman pushed a commit to openzfsonwindows/openzfs that referenced this pull request Mar 13, 2024
6.6 made i_ctime inaccessible; 6.7 has done the same for i_atime and
i_mtime. This extends the method used for ctime in b37f293 to atime
and mtime as well.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
lundman pushed a commit to openzfsonwindows/openzfs that referenced this pull request Mar 13, 2024
In 6.7 the superblock shrinker member s_shrink has changed from being an
embedded struct to a pointer. Detect this, and don't take a reference if
it already is one.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
lundman pushed a commit to openzfsonwindows/openzfs that referenced this pull request Mar 13, 2024
6.7 changes the shrinker API such that shrinkers must be allocated
dynamically by the kernel. To accomodate this, this commit reworks
spl_register_shrinker() to do something similar against earlier kernels.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
lundman pushed a commit to openzfsonwindows/openzfs that referenced this pull request Mar 13, 2024
6.7 changed the names of the time members in struct inode, so we can't
assign back to it because we don't know its name. In practice this
doesn't matter though - if we're missing current_time(), then we must be
on <4.9, and we know our fallback will need to return timespec.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
lundman pushed a commit to openzfsonwindows/openzfs that referenced this pull request Mar 13, 2024
6.6 made i_ctime inaccessible; 6.7 has done the same for i_atime and
i_mtime. This extends the method used for ctime in b37f293 to atime
and mtime as well.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
lundman pushed a commit to openzfsonwindows/openzfs that referenced this pull request Mar 13, 2024
In 6.7 the superblock shrinker member s_shrink has changed from being an
embedded struct to a pointer. Detect this, and don't take a reference if
it already is one.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
lundman pushed a commit to openzfsonwindows/openzfs that referenced this pull request Mar 13, 2024
6.7 changes the shrinker API such that shrinkers must be allocated
dynamically by the kernel. To accomodate this, this commit reworks
spl_register_shrinker() to do something similar against earlier kernels.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rob Norris <robn@despairlabs.com>
Sponsored-by: https://github.com/sponsors/robn
Closes openzfs#15681
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Status: Accepted Ready to integrate (reviewed, tested)
Projects
None yet
Development

Successfully merging this pull request may close these issues.

shrinker replacements for 6.7+ Linux 6.7 super_block has s_shrink failing
6 participants