This adds a splits the non-incremental threshold parameter into one for small heaps and one for large. What counts as large and small are controlled by the existing parameters that were previously used for dynamic heap growth. I also renamed the parameter from "non-incremental threshold" to "incremental limit".
The small heap parameter is increased to 1.4. This avoids non-incremental GC on facebook and reddit as reported in the dependent bugs and allows us to remove the splay latency hack that was previously necessary.
Differential Revision: https://phabricator.services.mozilla.com/D72903
A while back we added a heuristic to delay a GC if it would cause a reset (bug 1367455), but it was turned off after it caused a performance regression and hasn't been reenabled since.
Telemetry shows that that only 0.2% of GCs get reset anyway so we should probably just remove this.
Depends on D71328
Differential Revision: https://phabricator.services.mozilla.com/D71329
Set the flag on the affected classes, which are:
- CallbackTimeoutHandler
- nsJSArgArray
- CallbackObject
- Console
- MessageEvent
- IDBIndexCursor
- ExtendableMessageEvent
- JSPurpleBuffer
Differential Revision: https://phabricator.services.mozilla.com/D68196
--HG--
extra : moz-landing-system : lando
If we're in the middle of shutting down a content process, don't GC if we get a
memory pressure event. Shutting down the process is a better way to free memory!
This doesn't seem to happen much, presumably because once we've started
shutting down the content process we ignore new messages, like ones telling
us to do memory-pressure.
Differential Revision: https://phabricator.services.mozilla.com/D65755
--HG--
extra : moz-landing-system : lando
Hopefully this will reduce the number of shutdown hangs in child processes.
Differential Revision: https://phabricator.services.mozilla.com/D65131
--HG--
extra : moz-landing-system : lando
This changeset is a simple find and replace of `MOZ_FALLTHROUGH` and `[[fallthrough]]`.
Unfortunately, the MOZ_FALLTHROUGH_ASSERT macro (to assert on case fallthrough in debug builds) is still necessary after switching from [[clang::fallthrough]] to [[fallthrough]] because:
* MOZ_ASSERT(false) followed by [[fallthrough]] triggers a -Wunreachable-code warning in DEBUG builds
* but MOZ_ASSERT(false) without [[fallthrough]] triggers a -Wimplicit-fallthrough warning in NDEBUG builds.
Differential Revision: https://phabricator.services.mozilla.com/D56440
--HG--
extra : moz-landing-system : lando
The inclusions were removed with the following very crude script and the
resulting breakage was fixed up by hand. The manual fixups did either
revert the changes done by the script, replace a generic header with a more
specific one or replace a header with a forward declaration.
find . -name "*.idl" | grep -v web-platform | grep -v third_party | while read path; do
interfaces=$(grep "^\(class\|interface\).*:.*" "$path" | cut -d' ' -f2)
if [ -n "$interfaces" ]; then
if [[ "$interfaces" == *$'\n'* ]]; then
regexp="\("
for i in $interfaces; do regexp="$regexp$i\|"; done
regexp="${regexp%%\\\|}\)"
else
regexp="$interfaces"
fi
interface=$(basename "$path")
rg -l "#include.*${interface%%.idl}.h" . | while read path2; do
hits=$(grep -v "#include.*${interface%%.idl}.h" "$path2" | grep -c "$regexp" )
if [ $hits -eq 0 ]; then
echo "Removing ${interface} from ${path2}"
grep -v "#include.*${interface%%.idl}.h" "$path2" > "$path2".tmp
mv -f "$path2".tmp "$path2"
fi
done
fi
done
Differential Revision: https://phabricator.services.mozilla.com/D55442
--HG--
extra : moz-landing-system : lando
The inclusions were removed with the following very crude script and the
resulting breakage was fixed up by hand. The manual fixups did either
revert the changes done by the script, replace a generic header with a more
specific one or replace a header with a forward declaration.
find . -name "*.idl" | grep -v web-platform | grep -v third_party | while read path; do
interfaces=$(grep "^\(class\|interface\).*:.*" "$path" | cut -d' ' -f2)
if [ -n "$interfaces" ]; then
if [[ "$interfaces" == *$'\n'* ]]; then
regexp="\("
for i in $interfaces; do regexp="$regexp$i\|"; done
regexp="${regexp%%\\\|}\)"
else
regexp="$interfaces"
fi
interface=$(basename "$path")
rg -l "#include.*${interface%%.idl}.h" . | while read path2; do
hits=$(grep -v "#include.*${interface%%.idl}.h" "$path2" | grep -c "$regexp" )
if [ $hits -eq 0 ]; then
echo "Removing ${interface} from ${path2}"
grep -v "#include.*${interface%%.idl}.h" "$path2" > "$path2".tmp
mv -f "$path2".tmp "$path2"
fi
done
fi
done
Differential Revision: https://phabricator.services.mozilla.com/D55442
--HG--
extra : moz-landing-system : lando
This makes the function into a state machine implemented with a switch statement and gets rid of the recursive call. This results in a little duplication of logic but hopefully it is easier to understand overall. What do you think?
Differential Revision: https://phabricator.services.mozilla.com/D54750
--HG--
extra : moz-landing-system : lando
This adds two AUTO_PROFILER_LABEL_DYNAMIC_... macros and updates select
usages of the old macros to use the new ones. These new macros cause
the dynamic string of the label to be included in BHR stacks.
We don't want to do this all of the time, as in many cases we may not
be interested enough in the dynamic string or it may be sensitive
information, but it is rather important information for certain cases.
This uses the same buffer that we use for the strings for JS frames,
and if we fail to fit into that buffer we just append the raw label.
If the string is too long for our static buffer (128 bytes), we just
leave it truncated, as it should be stable and we may be able to infer
from the truncated form what the full form would be.
Differential Revision: https://phabricator.services.mozilla.com/D51665
--HG--
extra : moz-landing-system : lando
The problem is that a GC can be triggered in a number of ways, so scheduling zones in nsJSContext::GarbageCollectNow is not effective. This patch schedules these zones in the GC callback, which is always called at the start of GC. This means we have to change the internal logic slightly to allow the GC callback to schedule addional zones; previously this state was saved and restored before/after running the callback.
Differential Revision: https://phabricator.services.mozilla.com/D51145
--HG--
extra : moz-landing-system : lando
I'm not sure what if there's a preference either way, but all the other globals in nsJSEnvironement.cpp are static so I made the CycleCollectorStats global static too.
Differential Revision: https://phabricator.services.mozilla.com/D47919
--HG--
extra : moz-landing-system : lando
This needs to be defined before FireForgetSkippable for the subsequnt patches.
Differential Revision: https://phabricator.services.mozilla.com/D47918
--HG--
extra : moz-landing-system : lando
This doesn't fix the underlying problem but makes RunNextCollectorTimer more aggressive in what it does. Testing shows this substantially reduces the max GC heap size while running these tests as well as vsize and RSS. Hopefully this will fix the problem with the tests failing while we work out a good solution for the underlying issue.
Differential Revision: https://phabricator.services.mozilla.com/D44378
--HG--
extra : moz-landing-system : lando
Converted the following to StaticPrefs so that we can easily test variations:
NS_GC_DELAY
NS_SHRINK_GC_BUFFERS_DELAY
NS_FIRST_GC_DELAY
NS_FULL_GC_DELAY
NS_INTERSLICE_GC_DELAY
NS_USER_INTERACTION_INTERVAL
Differential Revision: https://phabricator.services.mozilla.com/D43112
--HG--
extra : moz-landing-system : lando
Simon Giesecke
Kris Maglione
Tom Schuster
Steve Fink
Jon Coppeard
Noemi Erli
Sylvestre Ledru
Andreas Farre
Tetsuharu OHZEKI
Sebastian Streich
Andrew McCreight
Ciure Andrei
Nazım Can Altınova
Chris Peterson
Jeff Walden
Gabriele Svelto
Dorel Luca
Doug Thayer
Karl Tomlinson
Andrew Creskey