User Details
- User Since
- Jun 4 2014, 10:38 AM (650 w, 1 d)
Thu, Nov 12
Tue, Nov 10
New callouts popping up on cpu 0 are already a minor scalability problem as for example this pops up when creating and destroying threads.
huh, lame on my end
I don't think the description of the deadlock fits the manapge, i provided it only for illustrative purposes. If one was to talk about something like this, I have to point out there are other ways to deadlock.
Now that I'm taking another look, even if going with sysconf, looking at the uma zone is a no-go. One day it might change (e.g., it may get removed altogether) and this will break userspace. A dedicated sysctl with the size is needed, but then I don't see any justificatin for adding a sysconf.
why spend a sysconf on it? exposing this with a sysctl seems like a more natural fit, but i'm not going to insist
This does not solve the problem though as the same scheduling can still occur -- the real fix would notice the callout is taking too long while it is taking too long and if need be redirect queued callbacks to other CPUs.
Fri, Oct 30
It is not legal to hold any vnode locks across a call to vn_fullpath. In particular having the vp locked threatens a deadlock against normal path lookup, as the reconstruction grabs one lock at a time the opposite way. So you can have the child locked and stall waiting for the parent, while the forward lookup just locked the parent and now wants the child.
Thu, Oct 22
sigh, use correct diff
I did not see any point in benchmarking this, it is strictly less work in the fast path.
- guard CTASSERT with _KERNEL. verified it fixes buildworld
Oct 20 2026
Oct 12 2026
While I agree the fence is missing, I completely disagree with the fix.
Oct 9 2026
Jul 8 2026
May 28 2026
This greatly reduces the race window, but it does not fix the problem.
Apr 18 2026
Apr 17 2026
Apr 7 2026
So the ranged bitmap proposal works, but it also has funny corner cases if fds are very sparse which perhaps are not worth it.
Mar 9 2026
Feb 27 2026
I did not co-autor the change, so please drop that tag.
Feb 25 2026
Nov 10 2025
these also need to go:
This looks legitimate.
Nov 4 2025
I checked sx and rw locks and I only see the check in the slowpath.
Oct 20 2025
Oct 18 2025
Do you have stats how many classes can be realistically present? maybe this would be way faster single-threaded by iterating an array?
Oct 16 2025
I'll be posting a different variant later today.
I think some work can be saved if the errno is ENOENT or EACCESS, but someone(tm) has to make sure errors of the sort are not returned from the stage where the syscall does the reverse lookup.
Oct 15 2025
Oct 14 2025
This *is* used by pf_rule_counter_u64_periodic. The goal is to sum up the 32 bit values frequently enough(tm) that they don't overflow.
Oct 6 2025
Oct 5 2025
Oct 4 2025
Oct 3 2025
- remove goto labels
- sort includes
- fix a bug found by pho@ in the lock smr patch: after the lock is acquired, doomed handling has to be the same as with the interlock-protected routine
Oct 1 2025
- patch up commentary
- remove a debug tunable
- upload the entire thing
- also drop ref drip around unlock
- doomed asserts
- cosmetic fixups to null_hashget_locked
- null_lock using smr
Sep 30 2025
I think this is turning into a bikeshed, so just do what you feel.
The above may be true and the pointer may be populated depending on when you got the the inode. I would just add something like "->i_din2 is NULL, some information omitted" or compatible.
maybe print that the pointer is NULL?
Sep 29 2025
Sep 28 2025
I don't have an opinion on the functionality, but the implementation looks avoidably slow.