Closed
Bug 791366
Opened 12 years ago
Closed 12 years ago
Implement memory reporters missing on BSDs
Categories
(Core :: XPCOM, defect)
Tracking
()
RESOLVED
FIXED
mozilla18
People
(Reporter: jbeich, Assigned: jbeich)
References
Details
Attachments
(3 files)
741 bytes,
patch
|
n.nethercote
:
review+
|
Details | Diff | Splinter Review |
2.53 KB,
patch
|
n.nethercote
:
review+
|
Details | Diff | Splinter Review |
2.93 KB,
text/plain
|
Details |
No description provided.
getrusage() comes from 4.2BSD and at least 4.4BSD has ru_majflt/ru_minflt. I'm not sure if there any incompatible implementations.
Attachment #661433 -
Flags: review?(justin.lebar+bug)
NS_SUCCEEDED is probably superfluous
Attachment #661434 -
Flags: review?(justin.lebar+bug)
Attachment #661434 -
Flags: review?(justin.lebar+bug) → review?(n.nethercote)
Attachment #661433 -
Flags: review?(justin.lebar+bug) → review?(n.nethercote)
Looking at MapsReporter some bits can be ported to struct kinfo_vmentry (only FreeBSD) - using kve_protection for rwx flags - using kve_path for pathname when KVME_TYPE_VNODE - Swap is when KVME_TYPE_SWAP - using kve_resident * pagesize for Rss - everything besides KVME_TYPE_{VNODE,DEVICE} for Anon mappings but others need more poking or mimicking procstat(1). Any hope on other BSDs?
Comment 5•12 years ago
|
||
> Attachment #661434 [details] [diff] - Flags: review?(justin.lebar+bug@gmail.com) → review?
> (n.nethercote@gmail.com)
This is probably wise; njn's review queue is probably less crazy than mine atm.
Comment 6•12 years ago
|
||
Comment on attachment 661433 [details] [diff] [review] Bug 791366 - Define HAVE_PAGE_FAULT_REPORTERS on every Unix. Review of attachment 661433 [details] [diff] [review]: ----------------------------------------------------------------- Seems ok; we can back out and try something else if it causes problems.
Attachment #661433 -
Flags: review?(n.nethercote) → review+
Comment 7•12 years ago
|
||
Comment on attachment 661434 [details] [diff] [review] Bug 791366 - Implement Vsize and Resident memory reports on BSDs. Review of attachment 661434 [details] [diff] [review]: ----------------------------------------------------------------- I won't pretend to understand how GetKinfoProcSelf(), but that doesn't worry me because this is BSD-only code. The rest of the patch looks fine, thanks.
Attachment #661434 -
Flags: review?(n.nethercote) → review+
Comment 8•12 years ago
|
||
(In reply to Jan Beich from comment #4) > Looking at MapsReporter some bits can be ported to struct kinfo_vmentry > (only FreeBSD) TBH I wouldn't bother. MapsReporter on Linux has provided us with very little in the way of useful information.
Keywords: checkin-needed
Comment 9•12 years ago
|
||
https://hg.mozilla.org/integration/mozilla-inbound/rev/5cdc27b83300 https://hg.mozilla.org/integration/mozilla-inbound/rev/8fce157732f3
Comment 10•12 years ago
|
||
https://hg.mozilla.org/mozilla-central/rev/5cdc27b83300 https://hg.mozilla.org/mozilla-central/rev/8fce157732f3
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla18
Comment 11•11 years ago
|
||
Can this bug be closed? All the blocking bugs are fixed.
You need to log in
before you can comment on or make changes to this bug.
Description
•