* [Bug translator/10796] New: thread-specific globals
@ 2009-10-16 22:23 jistone at redhat dot com
2009-10-23 23:42 ` [Bug translator/10796] " jistone at redhat dot com
2009-10-26 16:03 ` jkenisto at us dot ibm dot com
0 siblings, 2 replies; 3+ messages in thread
From: jistone at redhat dot com @ 2009-10-16 22:23 UTC (permalink / raw)
To: systemtap
Similar to PR10795, it would also be nice to have native per-thread globals.
It's a common pattern to have maps indexed only by tid(), so it would be nice if
that were automatic and lockless.
Per-thread is not as easy as per-cpu though -- we'll probably need to have a
fixed pool of thread storage areas, allocating when a probe hits a thread,
deallocating when a thread dies (perhaps even on exec). Those lookups might
need some locking, but hopefully that could be more fine-grained than wrapping
the entire probe.
--
Summary: thread-specific globals
Product: systemtap
Version: unspecified
Status: NEW
Severity: enhancement
Priority: P2
Component: translator
AssignedTo: systemtap at sources dot redhat dot com
ReportedBy: jistone at redhat dot com
http://sourceware.org/bugzilla/show_bug.cgi?id=10796
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug translator/10796] thread-specific globals
2009-10-16 22:23 [Bug translator/10796] New: thread-specific globals jistone at redhat dot com
@ 2009-10-23 23:42 ` jistone at redhat dot com
2009-10-26 16:03 ` jkenisto at us dot ibm dot com
1 sibling, 0 replies; 3+ messages in thread
From: jistone at redhat dot com @ 2009-10-23 23:42 UTC (permalink / raw)
To: systemtap
------- Additional Comments From jistone at redhat dot com 2009-10-23 23:42 -------
If we get this such that it performs better than plain "array[tid()]"
mechanisms, then we should port the .return $target saving to use this feature.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10796
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug translator/10796] thread-specific globals
2009-10-16 22:23 [Bug translator/10796] New: thread-specific globals jistone at redhat dot com
2009-10-23 23:42 ` [Bug translator/10796] " jistone at redhat dot com
@ 2009-10-26 16:03 ` jkenisto at us dot ibm dot com
1 sibling, 0 replies; 3+ messages in thread
From: jkenisto at us dot ibm dot com @ 2009-10-26 16:03 UTC (permalink / raw)
To: systemtap
------- Additional Comments From jkenisto at us dot ibm dot com 2009-10-26 16:02 -------
(In reply to comment #1)
> If we get this such that it performs better than plain "array[tid()]"
> mechanisms, then we should port the .return $target saving to use this feature.
If you're going to revise that feature, consider using
kretprobe_instance->data[] to store such data between call and return. That's
what it was designed for. See PR #5916 and section 1.3.2 of
Documentation/kprobes.txt.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10796
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2009-10-26 16:03 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-10-16 22:23 [Bug translator/10796] New: thread-specific globals jistone at redhat dot com
2009-10-23 23:42 ` [Bug translator/10796] " jistone at redhat dot com
2009-10-26 16:03 ` jkenisto at us dot ibm dot com
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).