- User Since
- Mar 7 2015, 6:48 PM (403 w, 6 d)
Split out vn_path_to_global_path_hardlink into its own commit.
I changed from DISTVERSION to PORTVERSION so that the package versioning matches the compiled-in version numbers. The porter's handbook doesn't seem to discourage this, just noting that 'Only one of PORTVERSION and DISTVERSION can be set at a time.'
Thu, Dec 1
Update containernetworking-plugins so that portmap works with the example pf.conf from v0.1
Wed, Nov 30
I will also reach out to pho@ for advice on stress testing.
Added a comment to note that vnode type validation happens in vfs_domount_first. Also fixed a couple of line-too-long formatting issues.
Tue, Nov 29
Added a couple of comments above
Mon, Nov 28
Add vn_path_to_global_path_hardlink. This works for trivial tests and makes
the code in vfs_domount more readable. I'll update my podman test VM and
stress test it a little.
Updated to include requested changes. I'm still working on a regular-file equivalent to vn_path_to_global_path.
I have tested this with my main use-cases (podman and buildah) and simple tests are working. I plan to run some of the podman integration tests to get a better signal.
THere were a couple of places in vfs_cache.c which assumed that mount
points are directories and paniced if they were files. I relaxes these
checks to allow file mounts.
Wed, Nov 23
This is a work in progress but I wanted a sanity check for the changes to vfs_mount.c while I figure out how to write ATF tests for it.
Tue, Nov 22
Fri, Nov 18
I added a new field to struct image_args. Since this was added to the end of the structure, it should make it possible to MFC to stable/13 since the structure is owned and zeroed by do_execve and any imgact modules should not be affected.
Thu, Nov 10
Fri, Nov 4
Nov 1 2022
Aug 19 2022
Aug 3 2022
FIx a typo in Makefie.packages
Add an example use-case for kerberos-lib to the commit message.
Updated Makefile.package to add entries for kerberos-lib
Jul 29 2022
Jun 17 2022
Jun 13 2022
May 30 2022
Sep 11 2021
Dec 13 2020
Jun 22 2020
Jun 15 2020
Not sure if this is still active, approving anyway.
Jun 14 2020
Feb 3 2019
Sep 30 2017
Looks good to me.
Apr 25 2017
May 8 2016
I think the patch is probably ok. As you point out, it does open a race where some other client can remove the file while we still have it open but its not clear how likely that is in the real world.
Feb 1 2016
Looks good to me.
Jan 31 2016
Jan 4 2016
May 10 2015
I know I suggested adding an option for the high flag bits but now I wonder if it would be cleaner to just have all 64 bits of flags in an option which would be or'ed with the flags argument from nmount. This would cut out most of the bit masking.