-
gitomat
[illumos-gate] 15594 smb_node_alloc calls fsop_root excessively -- Gordon Ross <gwr⊙rc>
-
sommerfeld
good morning. I'd appreciate it if someone could look at my RTI for
code.illumos.org/c/illumos-gate/+/2891
-
fenix
→ CODE REVIEW 2891: 15713 ulwp32_t is out of sync with ulwp_t (NEW) |
illumos.org/issues/15713
-
pmooney
sommerfeld: I'm looking at that RTI. Do you have the details from the wsdiff which listed the three functions you referenced?
-
sommerfeld
yes.
-
sommerfeld
I can re-share it but I attached it as wsdiff.15713 to the RTI email.
-
sommerfeld
the changes to td_ta_thr_iter and td_thr_get_info are obviously sizeof changes. the td_thr2to32 change is as well (the code sequence is a tricksy branchless thing)
-
sommerfeld
context is : td_thr2to32+0xa5: sbbl %eax,%eax
-
sommerfeld
td_thr2to32+0xa7: andl $0x124,%eax
-
sommerfeld
td_thr2to32+0xac: addl $0x18c,%eax
-
sommerfeld
td_thr2to32+0xb1: movl %eax,0x38(%rbx)
-
sommerfeld
td_thr2to32+0xb4: cmpb $0x0,0x7e(%r12)
-
sommerfeld
which is the code corresponding to ti_p->ti_ro_size = ulwp->ul_replace?
-
sommerfeld
REPLACEMENT_SIZE32 : sizeof (ulwp32_t);
-
gitomat
[illumos-gate] 15688 blkdev and lofi need to properly init cmlb minors -- Robert Mustacchi <rm⊙fo>
-
sjorge
jperkin on the up side the no go carbon stuff seems to at least mostly build in a venv, except carbon-web which struggles with cffi and stuff -_- why can't we just have nice things wrt software portability
-
pmooney
sommerfeld: ah, please attach that to the ticket?
-
pmooney
(since it's a valuable part of the test notes)
-
sommerfeld
one moment...
-
sommerfeld
done
-
gitomat
[illumos-gate] 13180 Remove printmgr, as it doesn't work with any current java -- Peter Tribble <peter.tribble⊙gc>
-
gitomat
[illumos-gate] 15713 ulwp32_t is out of sync with ulwp_t -- Bill Sommerfeld <sommerfeld⊙ame>
-
sommerfeld
pmooney: Thanks!
-
jbk
hrm.. when doing a scrub or resilver, if i'm reading this right, cv_broadcast() is called after every I/O... that seems excessive
-
richlowe
I mean, depends on who it's waking?
-
richlowe
and why
-
jbk
it's part of it's rate limiting
-
jbk
(which is unforuntately has some other flaws)
-
jbk
i think there'll be either 1-2 threads blocked
-
gitomat
[illumos-gate] 15722 which_scm confused by git worktree -- Bill Sommerfeld <sommerfeld⊙ame>
-
gitomat
[illumos-gate] 15718 Fix bad references in ufm manuals -- Robert Mustacchi <rm⊙fo>