Page MenuHomeFreeBSD

mckusick (Marshall Kirk McKusick)
User

Projects

User Details

User Since
Oct 24 2014, 7:17 PM (572 w, 3 d)

Recent Activity

Yesterday

mckusick committed rG0d4642a67e59: Add --libxo support for geom status and list sub commands. (authored by mckusick).
Add --libxo support for geom status and list sub commands.
Mon, Oct 13, 9:16 PM
mckusick closed D37615: geom <class> list/status --libxo support.
Mon, Oct 13, 9:16 PM
mckusick added a comment to D53061: sendmail: Restart sendmail service on package upgrade.

I note that the sshd daemon is restarted when it is reinstalled. The sshd daemon is restarted since its update may be security related and thus the security fix should take effect as soon as the updated version is available.

Mon, Oct 13, 4:28 PM

Tue, Oct 7

mckusick accepted D37615: geom <class> list/status --libxo support.

First, sorry for letting this languish. Though I have not dealt with libxo, I assume that you have ensured that this generates the output that you need. Structurally it is properly added to the geom utility. If there are no issues raised by the other reviewers by Monday October 13th, I will commit it.

Tue, Oct 7, 9:31 PM

Sun, Sep 14

mckusick accepted D52036: sys/power: Generic sleep types.

Mentor approval.

Sun, Sep 14, 9:52 PM
mckusick accepted D52043: acpi: Use sleep types defined in sys/power.h.

Mentor approval.

Sun, Sep 14, 6:09 PM

Sep 11 2025

mckusick accepted D52044: sys/power: Sleep type reporting by PM backends.

Looks ready to go.

Sep 11 2025, 5:41 AM

Sep 6 2025

mckusick accepted D52403: fcntl(F_SETFL): Don't unconditionally invoke FIONBIO and FIOASYNC.

Interesting history of this bug.

Sep 6 2025, 12:07 AM

Aug 30 2025

mckusick closed D2844: Bail out of corrupt directory entries during boot.

This has been committed to head and MFC'ed to 14.

Aug 30 2025, 11:00 PM · srcmgr
mckusick committed rG727bc500264b: Bail out of corrupt directory entries during boot (authored by mckusick).
Bail out of corrupt directory entries during boot
Aug 30 2025, 10:58 PM

Aug 28 2025

mckusick accepted D52128: libifconfig: Add netlink based helper to bring the interface up/down.

Mentor approval.

Aug 28 2025, 5:31 PM

Aug 25 2025

mckusick accepted D51588: amdgpio: Mask and service interrupts.

Mentor approval.

Aug 25 2025, 10:34 PM
mckusick accepted D52132: netlink: Handle `ifhwioctl(SIOCSIFMTU)` failure.

Mentor approval.

Aug 25 2025, 10:15 PM
mckusick accepted D51871: netlink: Fix IFF_UP flag handling in RTM_NEWLINK's modify_link handler.

Mentor approval for commit.

Aug 25 2025, 5:41 PM

Aug 23 2025

mckusick accepted D51922: netlink/route: Support modifying IFLA_ADDRESS with RTM_NEWLINK.

Mentor approval.

Aug 23 2025, 6:25 PM
mckusick accepted D52056: netlink: Bypass refcounting when setting promiscuity.

Mentor approval - looks good to go.

Aug 23 2025, 6:19 PM

Aug 22 2025

mckusick committed rG690ae8a2025c: Bail out of corrupt directory entries during boot (authored by mckusick).
Bail out of corrupt directory entries during boot
Aug 22 2025, 5:39 AM

Aug 19 2025

mckusick accepted D50666: libifconfig: Remove ifconfig_set_fib and ifconfig_sfp_id_is_qsfp.

Mentor approval for commit.

Aug 19 2025, 11:00 PM

Aug 17 2025

mckusick accepted D51871: netlink: Fix IFF_UP flag handling in RTM_NEWLINK's modify_link handler.

Mentor approval for commit.

Aug 17 2025, 6:06 PM
mckusick accepted D51947: vfs: Introduce VN_ISDEV() macro.

I am always in favor of improvement to code clarity.

Aug 17 2025, 5:41 PM

Aug 14 2025

mckusick accepted D51871: netlink: Fix IFF_UP flag handling in RTM_NEWLINK's modify_link handler.

Mentor approval for commit.

Aug 14 2025, 2:41 AM
mckusick added a comment to D48732: acpi: Abstract over ACPI Sx sleep states with `enum sleep_type`.

I concur with @imp comment that I like the notion of abstracting them so we don't churn that too much...

Aug 14 2025, 2:33 AM

Aug 11 2025

mckusick accepted D2844: Bail out of corrupt directory entries during boot.

This is (still) a sensible thing to do.

Aug 11 2025, 6:29 AM · srcmgr

Aug 7 2025

mckusick accepted D48384: acpi_powerres: D3cold support and `acpi_d_state_to_str`.

With Mark's review and approval this looks good to go.

Aug 7 2025, 5:49 PM

Jul 28 2025

mckusick accepted D51573: Another fix for UFS with inode number sign propagation.

This change is correct.

Jul 28 2025, 8:52 PM

Jul 26 2025

mckusick accepted D51487: amdsmu: Read and dump idlemask.

Review process looks good and complete.
Adding my accept as your mentor.

Jul 26 2025, 11:02 PM
mckusick accepted D48714: amdsmu: Get and expose sysctls for metrics about last sleep.

Review process looks good and complete.
Adding my accept as your mentor.

Jul 26 2025, 10:06 PM
mckusick accepted D48683: amdsmu: Initial work on a driver for the AMD SMU.

Review process looks good and complete.
Adding my accept as your mentor.

Jul 26 2025, 6:14 PM

Jul 20 2025

mckusick accepted D50832: stand: Add support for `print_delay` environment variable in loader.

Review process is complete with all the right folks signing off.

Jul 20 2025, 2:19 PM
mckusick accepted D51428: committers-src: add myself (obiwac@).

I suggest that you add 'Sponsored by: The FreeBSD Foundation' to your commit as it helps show that they are supporting your work.

Jul 20 2025, 3:21 AM
mckusick accepted D51427: new committer (src): Aymeric Wibo (obiwac).

I too suggest that you add 'Sponsored by: The FreeBSD Foundation' to your commit as it helps show that they are supporting your work.

Jul 20 2025, 3:17 AM

Jul 6 2025

mckusick accepted D51037: makefs: Fix undefined behaviour in ffs.c.

Looks good.

Jul 6 2025, 11:30 PM

May 28 2025

mckusick accepted D50562: newfs: enable newfs to run on plain files without a bsdlabel.

This seems to be a reasonable change.

May 28 2025, 6:03 AM

May 27 2025

mckusick accepted D50520: fsck_ffs: change struct bufarea member b_index type to uint64_t.

Agree with change.

May 27 2025, 12:10 AM

May 18 2025

mckusick accepted D50401: namei(9): document the rest.

One grammer nit.

May 18 2025, 5:47 PM

May 17 2025

mckusick added a comment to D50338: vfs: vn_alloc(): Stop always calling vn_alloc_hard() and direct reclaiming.

Sorry for being out of the loop here, I just got back from a trip to the arctic where unsurprisingly there was little to no Internet.

May 17 2025, 7:25 PM

May 8 2025

mckusick added a comment to D48374: bsdinstall: Remove advice to turn off SUJ for SSDs.

There's another set of questions:

  • Is SU+J detrimental under certain conditions?
  • Should we be asking the user this question in the installer?

It's the second question that prompted this review. I don't care too much about whether SU+J is always on by default, always off by default, or on by default iff the fs is on spinning rust (or some equivalent). I do want to remove cognitive load from installer questions that aren't really required.

May 8 2025, 8:01 PM
mckusick added a comment to D50196: makefs: Ensure that FFS superblocks are reproducible.

This has been fixed in newfs for some time. Along with numerous other changes like checkhash validations, superblock integrity checks, and other fixes and improvements. In my opinion makefs/ffs should just use use newfs/mkfs.c rather than (poorly) trying to parody it.

May 8 2025, 7:58 PM

Mar 29 2025

mckusick committed rGae15f8ceaa85: Get consistent updates for UFS superblocks. (authored by mckusick).
Get consistent updates for UFS superblocks.
Mar 29 2025, 12:44 AM
mckusick closed D49276: Get consistent updates for UFS superblocks.
Mar 29 2025, 12:44 AM

Mar 20 2025

mckusick committed rG16649530b7be: Formatting and style cleanups. (authored by mckusick).
Formatting and style cleanups.
Mar 20 2025, 8:38 PM

Mar 19 2025

mckusick committed rGc2cd605e8c8a: Get consistent updates for UFS superblocks. (authored by mckusick).
Get consistent updates for UFS superblocks.
Mar 19 2025, 11:35 PM

Mar 14 2025

mckusick added inline comments to D49276: Get consistent updates for UFS superblocks.
Mar 14 2025, 11:12 PM
mckusick updated the diff for D49276: Get consistent updates for UFS superblocks.

Clarify when the superblock is a copy and fix yet another Peter Holm test failure.

Mar 14 2025, 11:11 PM
mckusick added inline comments to D49276: Get consistent updates for UFS superblocks.
Mar 14 2025, 1:58 AM
mckusick updated the diff for D49276: Get consistent updates for UFS superblocks.

Update with kib comments.
Add fixes for bugs found by Peter Holm.

Mar 14 2025, 1:57 AM

Mar 7 2025

mckusick requested review of D49276: Get consistent updates for UFS superblocks.
Mar 7 2025, 7:46 AM

Feb 16 2025

mckusick closed D48711: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106.

This has been committed to head and MFC'ed to 14 and 13. It was committed to 13 in time to be part of the 13.5 distribution.

Feb 16 2025, 8:13 AM

Feb 10 2025

mckusick committed rGdfe803fdbc54: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106 (authored by mckusick).
Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106
Feb 10 2025, 7:40 AM

Feb 7 2025

mckusick committed rG6f87402a02f0: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106 (authored by mckusick).
Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106
Feb 7 2025, 12:28 AM

Feb 6 2025

mckusick abandoned D48794: Fix kernel build with GEOM_LABEL and no FFS.

I have accepted gleb's changes, so this issue is resolved.

Feb 6 2025, 8:01 PM
mckusick commandeered D48794: Fix kernel build with GEOM_LABEL and no FFS.

I have gone with Gleb's fix. It is simple and as Warner has pointed out `It's the least bad outcome. We are doing ffs things, even if they are minimal.' As I have pointed out, your change breaks libufs and would require many additional changes after making the needed fixes in libufs to the clients of libufs.

Feb 6 2025, 7:59 PM
mckusick committed rG5572827428bb: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106 (authored by mckusick).
Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106
Feb 6 2025, 7:26 PM

Feb 1 2025

mckusick committed rG15eed13089da: Add notes for handling the January 19, 2038 date limit in UFS1 filesystems (authored by mckusick).
Add notes for handling the January 19, 2038 date limit in UFS1 filesystems
Feb 1 2025, 7:01 PM
mckusick added a comment to D48794: Fix kernel build with GEOM_LABEL and no FFS.

Actually, there is an even easier fix. Just move the declaration of vfs_ffs from ffs_alloc.c to ffs_subr.c. Then everything just works.

Feb 1 2025, 5:41 PM
mckusick added a comment to D48794: Fix kernel build with GEOM_LABEL and no FFS.

This will cause libufs to fail to build. Also, ffs_subr.c is supposed to contain all the kernel functionality needed by filesystem utilities. And ffs_oldfscompat_inode_read() is one of those functions.

Feb 1 2025, 7:12 AM

Jan 31 2025

mckusick committed rG6cd973d903c8: Update date of ffs.4 manual page. (authored by mckusick).
Update date of ffs.4 manual page.
Jan 31 2025, 1:32 AM
mckusick committed rG1111a44301da: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106 (authored by mckusick).
Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106
Jan 31 2025, 1:32 AM
mckusick added inline comments to D48711: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106.
Jan 31 2025, 1:14 AM

Jan 30 2025

mckusick updated the diff for D48711: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106.

Updates to respond to reviewer feedback,

Jan 30 2025, 10:24 PM
mckusick added a comment to D48711: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106.

I will follow up these comments with a new set of diffs reflecting suggested changes.

Jan 30 2025, 10:23 PM

Jan 29 2025

mckusick requested review of D48711: Defer the January 19, 2038 date limit in UFS1 filesystems to February 7, 2106.
Jan 29 2025, 1:58 AM

Jan 28 2025

mckusick committed rG256389eaf158: Fix build from aa90fbed151de5 by eliminating an unused variable. (authored by mckusick).
Fix build from aa90fbed151de5 by eliminating an unused variable.
Jan 28 2025, 8:49 AM
mckusick committed rGaa90fbed151d: Standardize the definition of a UFS dinode. (authored by mckusick).
Standardize the definition of a UFS dinode.
Jan 28 2025, 1:41 AM

Jan 25 2025

mckusick committed rG6a356edd4fc3: Fix backward compatibility with UFS1 filesystems created before June 2002 (authored by mckusick).
Fix backward compatibility with UFS1 filesystems created before June 2002
Jan 25 2025, 9:11 PM

Jan 24 2025

mckusick committed rG544ad91ef9f6: Use ffs_sbget() to replace open-coded superblock reload. (authored by mckusick).
Use ffs_sbget() to replace open-coded superblock reload.
Jan 24 2025, 1:23 AM

Jan 23 2025

mckusick committed rGbb14579988c5: Eliminate unneeded variables and computations. (authored by mckusick).
Eliminate unneeded variables and computations.
Jan 23 2025, 6:45 AM
mckusick committed rG26c483aec039: Fix backward compatibility with UFS1 filesystems created before June 2002 (authored by mckusick).
Fix backward compatibility with UFS1 filesystems created before June 2002
Jan 23 2025, 6:41 AM
mckusick committed rGf31c37f8acb2: Use ffs_sbget() to replace open-coded superblock reload. (authored by mckusick).
Use ffs_sbget() to replace open-coded superblock reload.
Jan 23 2025, 6:39 AM
mckusick committed rG996fe40e05c9: Eliminate unneeded variables and computations. (authored by mckusick).
Eliminate unneeded variables and computations.
Jan 23 2025, 6:37 AM

Jan 17 2025

mckusick closed D48472: Fix backward compatibility with UFS1 filesystems created before June 2002.

Committed as 661ca921e8cd56b17fc6615bc7e596e56e0e7c31

Jan 17 2025, 4:22 AM

Jan 16 2025

mckusick committed rG661ca921e8cd: Fix backward compatibility with UFS1 filesystems created before June 2002 (authored by mckusick).
Fix backward compatibility with UFS1 filesystems created before June 2002
Jan 16 2025, 6:45 PM

Jan 15 2025

mckusick requested review of D48472: Fix backward compatibility with UFS1 filesystems created before June 2002.
Jan 15 2025, 9:53 PM
mckusick committed rGc18a6c15f415: Use ffs_sbget() to replace open-coded superblock reload. (authored by mckusick).
Use ffs_sbget() to replace open-coded superblock reload.
Jan 15 2025, 9:11 PM

Jan 13 2025

mckusick accepted D48378: ffs_reallocblks(): ensure that pref cg is valid.

These changes look correct to me. I am surprised that it has taken this long to trip over this error case.

Jan 13 2025, 5:05 AM

Jan 12 2025

mckusick committed rGdf48361e7792: Eliminate unneeded variables and computations. (authored by mckusick).
Eliminate unneeded variables and computations.
Jan 12 2025, 9:52 PM

Jan 9 2025

mckusick added a comment to D48329: ls -h: humanize the total as well.

I agree that -h changes the format of all numbers, So it should just be "x MB".

Jan 9 2025, 6:26 AM
mckusick added a comment to D48374: bsdinstall: Remove advice to turn off SUJ for SSDs.

I largely agree with phk's analysis. SU+J does more writing than just SU (the extra writes being the journal data). Fsck runs very quickly on SSDs since random seeks do not slow it down as they do spinning rust. It would be worthwhile to run a timing test of fsck -p on a large SSD to see if it takes more than say 30 seconds but my guess is that it will not.

Jan 9 2025, 6:07 AM

Dec 18 2024

mckusick added a comment to D47887: style(9): deprecate blank lines without local vars.

Sorry, late to comment.

Dec 18 2024, 9:42 PM

Nov 13 2024

mckusick added a reverting change for rG8733bc277a38: vfs: don't provoke recycling non-free vnodes without a good reason: rG2ca9c96dc0cf: Revert commit 8733bc277a383cf59f38a83956f4f523869cfc90.
Nov 13 2024, 7:53 PM
mckusick committed rG2ca9c96dc0cf: Revert commit 8733bc277a383cf59f38a83956f4f523869cfc90 (authored by mckusick).
Revert commit 8733bc277a383cf59f38a83956f4f523869cfc90
Nov 13 2024, 7:52 PM
mckusick added a reverting change for rG8733bc277a38: vfs: don't provoke recycling non-free vnodes without a good reason: rGab05a1cf321a: Revert commit 8733bc277a383cf59f38a83956f4f523869cfc90.
Nov 13 2024, 7:47 PM
mckusick committed rGab05a1cf321a: Revert commit 8733bc277a383cf59f38a83956f4f523869cfc90 (authored by mckusick).
Revert commit 8733bc277a383cf59f38a83956f4f523869cfc90
Nov 13 2024, 7:47 PM

Oct 12 2024

mckusick committed rG51b1d399b2de: Ensure that soft updates are not enabled by default when using mdmfs(8) (authored by mckusick).
Ensure that soft updates are not enabled by default when using mdmfs(8)
Oct 12 2024, 10:33 PM
mckusick committed rG54416c6e1379: Ensure that soft updates are not enabled by default when using mdmfs(8) (authored by mckusick).
Ensure that soft updates are not enabled by default when using mdmfs(8)
Oct 12 2024, 10:30 PM

Oct 3 2024

mckusick committed rG5b21d4ad060a: Ensure that soft updates are not enabled by default when using mdmfs(8) (authored by mckusick).
Ensure that soft updates are not enabled by default when using mdmfs(8)
Oct 3 2024, 3:08 PM

Jul 28 2024

mckusick added a comment to D45625: getblk: avoid bufobj interlock when waiting on buf lock.

Sorry to be slow to comment. Like D45624, this change looks functionally correct. Again, the main issue for me is to understand what sort of workload it helps. Do you have an example benchmark whose performance is improved with this change?

Jul 28 2024, 11:42 PM
mckusick added a comment to D45624: lockmgr: introduce lock_sleepgen.

Sorry to be slow to comment. This change looks functionally correct. The main issue for me is to understand what sort of workload it helps. Do you have an example benchmark whose performance is improved with this change?

Jul 28 2024, 11:42 PM
mckusick committed rG21161df3e032: Do not allow snapshots on UFS filesystems using gjournal. (authored by mckusick).
Do not allow snapshots on UFS filesystems using gjournal.
Jul 28 2024, 7:07 PM
mckusick committed rGb45276b1693d: Do not allow snapshots on UFS filesystems using gjournal. (authored by mckusick).
Do not allow snapshots on UFS filesystems using gjournal.
Jul 28 2024, 7:04 PM

Jul 27 2024

mckusick added a comment to D45987: tmpfs: Account for whiteouts during rename/rmdir.

I do not remember why the change to allow the removing of directories with whiteouts became allowed. The changes in commit 996c772f581f56248 are quite extensive. If you can let me know which file(s) in that commit have the relevant changes, I can go back through the changes made in those files for the Lite/2 release to see if any explanation shows up.

Sure, I'm referring to:

--- a/sys/ufs/ufs/ufs_lookup.c
+++ b/sys/ufs/ufs/ufs_lookup.c
@@ -906,7 +961,7 @@ ufs_dirempty(ip, parentino, cred)
                if (dp->d_reclen == 0)
                        return (0);
                /* skip empty entries */
-               if (dp->d_ino == 0)
+               if (dp->d_ino == 0 || dp->d_ino == WINO)
                        continue;
                /* accept only "." and ".." */
 #              if (BYTE_ORDER == LITTLE_ENDIAN)

part of function ufs_dirempty(), which is (now, but if I'm not mistaken also was then) called only by ufs_rename() and ufs_rmdir().

Barring other reasons I can't see now, I think we'd better revert that behavior, both for logical reasons (the directory references no file but still has whiteout entries that we are going to lose silently) and practical ones (e.g., on Linux, whiteouts are implemented as device files, and thus rmdir() on directories holding some will fail with ENOTEMPTY; this is also to ease consistency in ZFS' code).

Jul 27 2024, 1:54 AM

Jul 25 2024

mckusick added a comment to D45987: tmpfs: Account for whiteouts during rename/rmdir.

I suspect that removing a directory with whiteouts was allowed for UFS only because not doing so would confuse most users not aware of whiteouts, as these are not printed by a regular ls. I doubt there was any other reason than that. It appears the test on WINO wasn't in 4.4BSD and was introduced in the Lite/2 commit (996c772f581f56248). Adding @mckusick as he might have some recollection of what lead to this decision.

Jul 25 2024, 6:32 AM
mckusick committed rGcab41e0dd31d: Do not allow snapshots on UFS filesystems using gjournal. (authored by mckusick).
Do not allow snapshots on UFS filesystems using gjournal.
Jul 25 2024, 6:11 AM

May 22 2024

mckusick committed rG017d52e38ab3: Enable soft updates by default for UFS2 filesystems. (authored by mckusick).
Enable soft updates by default for UFS2 filesystems.
May 22 2024, 7:23 PM

May 21 2024

mckusick accepted D45245: getblk: fail faster with GB_LOCK_NOWAIT.

I concur that this seems like a reasonable optimization.

May 21 2024, 5:23 AM
mckusick accepted D45248: lockmgr: make lockmgr_disowned public and use it.

Useful abstraction of internal state.

May 21 2024, 5:23 AM
mckusick accepted D45249: buf: define and use BUF_DISOWNED.

Better to avoid use of internal lock state.

May 21 2024, 5:23 AM

May 16 2024

mckusick committed rG6b2af2d88ffd: Enable soft updates by default for UFS2 filesystems. (authored by mckusick).
Enable soft updates by default for UFS2 filesystems.
May 16 2024, 11:45 PM
mckusick committed rG7465f9fc0646: Enable soft updates by default for UFS2 filesystems. (authored by mckusick).
Enable soft updates by default for UFS2 filesystems.
May 16 2024, 6:56 AM

May 15 2024

mckusick accepted D45208: vfs_domount_update(): postpone setting MNT_UNION until VFS_MOUNT() is done.

This looks like the correct fix. One other unrelated nit.

May 15 2024, 7:47 PM