This was used to label IndexedDB work and work in storage/mozStorage*.
I don't think this deserves its own category; categories are most useful for
the main thread, and most of the time-consuming database-related work happens
on helper threads. The main thread pieces are mostly for asynchronicity-
coordination and don't usually take up time.
This patch labels IndexedDB work as DOM instead (which is maybe debatable) and
the rest as OTHER.
MozReview-Commit-ID: 3UYhFFbi3Ry
--HG--
extra : rebase_source : 5c88dfd67274103de01fe44191f49776017738f9
The next changeset is going to move over more annotations that Gecko developers
would count as "layout" into the LAYOUT category, and which is currently marked
as GRAPHICS.
We can add a subcategory for style resolution once we have subcategories, but
for now I think it makes more sense to put style resolution into the same bucket
as reflow and display list building.
MozReview-Commit-ID: 7r9eICVBA1Z
--HG--
extra : rebase_source : ce2df7a07522e99b0ccb59e40a8eae590ebfe834
Categories are useful to indicate: This much % of time was spent in this category.
The EVENTS category isn't a very good match for this. This category is currently
only set on labels of functions that handle the processing of an event. But
those functions are usually closer to the base of the stack, and the actual CPU
work during the processing of an event is usually in another category closer to
the top of the stack, e.g. in JS if we're running an event handler, or in LAYOUT
if we're hit testing the position of the event.
This changeset removes the EVENTS category and replaces all uses of it with the
OTHER category.
MozReview-Commit-ID: JPm5hQiBkvp
--HG--
extra : rebase_source : 66f8ee003d2f70111f4cff16d6e2d906ef4bf10b
They're very similar as far as most users of the profiler are concerned, I'd
say, and I don't believe it's worth giving them two different colors in the
activity graphs.
MozReview-Commit-ID: HTqjp56naL3
--HG--
extra : rebase_source : cf8d64bc3e76ed9bb07100081aebfc404845b8bc
This was used to label IndexedDB work and work in storage/mozStorage*.
I don't think this deserves its own category; categories are most useful for
the main thread, and most of the time-consuming database-related work happens
on helper threads. The main thread pieces are mostly for asynchronicity-
coordination and don't usually take up time.
This patch labels IndexedDB work as DOM instead (which is maybe debatable) and
the rest as OTHER.
MozReview-Commit-ID: 3UYhFFbi3Ry
--HG--
extra : rebase_source : f83946138d8311ea5aa91f537a1d8e420e784068
The next changeset is going to move over more annotations that Gecko developers
would count as "layout" into the LAYOUT category, and which is currently marked
as GRAPHICS.
We can add a subcategory for style resolution once we have subcategories, but
for now I think it makes more sense to put style resolution into the same bucket
as reflow and display list building.
MozReview-Commit-ID: 7r9eICVBA1Z
--HG--
extra : rebase_source : f447dcbb9d81be81a418c7464ef814ce4778073b
Categories are useful to indicate: This much % of time was spent in this category.
The EVENTS category isn't a very good match for this. This category is currently
only set on labels of functions that handle the processing of an event. But
those functions are usually closer to the base of the stack, and the actual CPU
work during the processing of an event is usually in another category closer to
the top of the stack, e.g. in JS if we're running an event handler, or in LAYOUT
if we're hit testing the position of the event.
This changeset removes the EVENTS category and replaces all uses of it with the
OTHER category.
MozReview-Commit-ID: JPm5hQiBkvp
--HG--
extra : rebase_source : fb1ff20aeebb7ac494227e62ba2101039578808c
They're very similar as far as most users of the profiler are concerned, I'd
say, and I don't believe it's worth giving them two different colors in the
activity graphs.
MozReview-Commit-ID: HTqjp56naL3
--HG--
extra : rebase_source : ed1dc54efd143ad64bc21cfc39853ae90477cc9d
This was used to label IndexedDB work and work in storage/mozStorage*.
I don't think this deserves its own category; categories are most useful for
the main thread, and most of the time-consuming database-related work happens
on helper threads. The main thread pieces are mostly for asynchronicity-
coordination and don't usually take up time.
This patch labels IndexedDB work as DOM instead (which is maybe debatable) and
the rest as OTHER.
MozReview-Commit-ID: 3UYhFFbi3Ry
--HG--
extra : rebase_source : 16abc3c4bd8ed9ac55b5c188bd10ee26b0566330
The next changeset is going to move over more annotations that Gecko developers
would count as "layout" into the LAYOUT category, and which is currently marked
as GRAPHICS.
We can add a subcategory for style resolution once we have subcategories, but
for now I think it makes more sense to put style resolution into the same bucket
as reflow and display list building.
MozReview-Commit-ID: 7r9eICVBA1Z
--HG--
extra : rebase_source : 53ce9912d3219ce8ce5dc411e03bb5ef5e2c7b64
Categories are useful to indicate: This much % of time was spent in this category.
The EVENTS category isn't a very good match for this. This category is currently
only set on labels of functions that handle the processing of an event. But
those functions are usually closer to the base of the stack, and the actual CPU
work during the processing of an event is usually in another category closer to
the top of the stack, e.g. in JS if we're running an event handler, or in LAYOUT
if we're hit testing the position of the event.
This changeset removes the EVENTS category and replaces all uses of it with the
OTHER category.
MozReview-Commit-ID: JPm5hQiBkvp
--HG--
extra : rebase_source : 34fb2ca94151403a6d7ffd5a8b840f00a8bb4afb
They're very similar as far as most users of the profiler are concerned, I'd
say, and I don't believe it's worth giving them two different colors in the
activity graphs.
MozReview-Commit-ID: HTqjp56naL3
--HG--
extra : rebase_source : f172424042fab18a514201ba4b6c67c03c209cdb
This also changes many references to the 'pseudo stack' to refer to the 'label
stack' instead. The label stack is one of the two stacks that are managed by
the profiling stack, the other stack being the JS interpreter stack.
MozReview-Commit-ID: Ed0YMMeCBY8
--HG--
extra : rebase_source : 5675d670f424c7d7dda04bafc2b3431fa2485e3c
The term "entry" is already used for elements in the profile buffer.
MozReview-Commit-ID: 1aB22V6veQh
--HG--
extra : rebase_source : c664eb4d6bed6cb74ba8a1b67ea99bd8ca57bcf7
extra : source : 3264c0cc0027b240b55bd3aebf27263b1e1d1cc0
The name Cpp was confusing, because C++ functions are in the native stack, not
in the pseudo stack. The pseudo stack only contains frames for manually
instrumented code that uses AutoProfilerLabel, and JS frames.
MozReview-Commit-ID: 9ptfhREo0qy
--HG--
extra : rebase_source : 76a1a32acb4c946aeb2ad45e904e419c1c9e2ad1
This adds JS::Zone to TypeDecls.h. Arguably that is pretty borderline, but
even though the Zone type is only used in a dozen or so places in Gecko,
it's a central concept.
--HG--
extra : rebase_source : 2267ad26c8888f3db9213acdb64ce7a56b1c16e5
extra : source : 7f7eb82a83f47c8e1a4b67527c42f88e35f9ff4a
This requires moving some things around. RootingContext is an existing
superclass of JSContext whose members are exposed in a header file, so we can
use it to expose the location of the geckoProfiler_ member to non-JS code.
MozReview-Commit-ID: 3oClAEVMsDr
--HG--
extra : rebase_source : f5943a710cddaa65cfdcb13370f95387caaf6892
extra : source : 2f5b804cef7b0888ac6121f5645bbb910750ed72
Before this patch, the writes to stackPointer were done using sequentially-
consistent stores. On x86 these compile to 'xchg' instructions, which showed up
as the most expensive part of the PseudoStack overhead.
MozReview-Commit-ID: IP9w9ievEXZ
--HG--
extra : rebase_source : 6b08b1ce98b59812e9da6da055a9a8daa18a1200
Only one thread ever modifies a PseudoStack, so we don't need to enforce
synchronization of writes from different threads. We can just read the old
value, add one to it, and then do an atomic store with the new value, because
we know that the current value of stackPointer can't have changed in the
meantime.
On its own, this patch actually seems to make things slower. But combined with
the next patch (which changes the memory ordering to ReleaseAcquire) it doesn't.
(I haven't checked whether the next patch on its own would give just as much
improvements with and without this patch.)
MozReview-Commit-ID: 3WIdyJC9kcj
--HG--
extra : rebase_source : c4e88746a0239cab3d0f9cf6f7b7fb10732a62ab
These annotations aren't doing anything useful. The important thing with
the PseudoStack is that, during pushes, the stack pointer incrementing happens
after the new entry is written, and this is ensured by the stack pointer being
Atomic.
The patch also improves the comments on PseudoStack.
--HG--
extra : rebase_source : 100f8a5e4b750c15fac66175550c4c284a141f16
There are three flags in ProfileEntry::Flags, which suggests there are 2**3 = 8
combinations. But there are only actually 4 valid combinations.
This patch converts the three flags to a single "kind" enum, which makes things
clearer. Note also that the patch moves the condition at the start of
AddPseudoEntry() to its callsite, for consistency with the earlier JS_OSR entry
kind check.
--HG--
extra : rebase_source : 0950769ee1530291860ef3be47d240df5939e871
This patch:
- renames flags_ as flagsAndCategory_ because it contains both the flags and
the category;
- uses << for some bitfield definitions, because it's easier to read;
- removes some dead methods from ProfileEntry;
- removes the unnecessary JS_FRIEND_API from setPC().
--HG--
extra : rebase_source : 24e24b2f6c948b770293ea497be4933f58be3722
- The profiler gives the JS engine a reference to the pseudo-stack via
SetContextProfiilngStack(). That function now takes a |PseudoStack*| instead
of a |ProfileEntry*| and pointer to the stack pointer.
- PseudoStack now has a |kMaxEntries| field, which is easier to work with than
|mozilla::ArrayLength(entries)|.
- AddressOfStackPointer() is no longer needed.
- The patch also neatens up the push operations significantly. PseudoStack now
has pushCppFrame() and pushJsFrame(), which nicely encapsulate the two main
cases. These delegate to the updated initCppFrame() and initJsFrame()
functions in ProfileEntry.
- Renames max_stck in testProfileStrings.cpp as peakStackPointer, which is a
clearer name.
- Removes a couple of checks from testProfileStrings.cpp. These checks made
sense when the pseudo-stack was accessed via raw manipulation, but are not
applicable now because we can't artificially limit the maximum stack size so
easily.
This includes renaming its fields to match SpiderMonkey naming conventions
instead of Gecko naming conventions.
This patch is just about moving the code. The next patch will change
SpiderMonkey to actually use PseudoStack directly.
--HG--
extra : rebase_source : 27e77ddf950201eb6bdba60003218056442cf7ab
ProfileEntry has |string|, which can be static or dynamic, and |dynamicString|.
If |string| is dynamic, the FRAME_LABEL_COPY flag must be set, and it will be
copied into profiler output.
But there is only one place that uses dynamic |string| values, in SpiderMonkey.
And that place doesn't use |dynamicString|. So this patch changes that place to
use an empty |string| and put the old dynamic |string| value in
|dynamicString|. This in turn removes the need for FRAME_LABEL_COPY.
One minor wrinkle is that when |dynamicString| is used the old code put a space
between |string| and |dynamicString|. The new code omits the space if |string|
is empty.
The patch also renames ProfileEntry::string as ProfileEntry::label_, which
better matches how it's used, and ProfileEntry::dynamicString as
ProfileEntry::dynamicString_ so the getter can be renamed dynamicString().
Steve Fink
Markus Stange
Csoregi Natalia
Noemi Erli
Emilio Cobos Álvarez
Jeff Walden
Andreea Pavel
Jason Orendorff
Cosmin Sabou
Nicholas Nethercote
Carsten "Tomcat" Book