Support veneer branches longer than short branches
Categories
(Core :: JavaScript Engine: JIT, task, P1)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox154 | --- | fixed |
People
(Reporter: anba, Assigned: anba)
References
(Blocks 1 open bug)
Details
Attachments
(4 external files)
Support case when veneer branches are longer than short branches in AssemblerBufferWithConstantPools. This will allow to remove the branch padding added in bug 1996840.
| Assignee | ||
Comment 1•5 months ago
|
||
The pool guard branch for RISCV64 is a single jal instruction, so
guardSize should be 1.
Drive-by fix:
- Change
headerSizeto 1 to match the instruction written inWritePoolHeader.
| Assignee | ||
Comment 2•5 months ago
|
||
Add a test case when veneer branches are longer than short branches.
The test case will cause an assertion without the changes from part 3.
| Assignee | ||
Comment 3•5 months ago
|
||
Account for veneers which are longer than short branches.
ARM64 has VeneerSize = 1 which should allow the compiler to fold everything in
sizeOfPrimaryVeneers to return zero. So this change only has an effect for
RISCV64, where VeneerSize = 2.
Updated•5 months ago
|
Backed out for causing build bustages at testAssemblerBuffer.cpp
Backout link
Push with failures
Failure log(s)
| Assignee | ||
Comment 7•5 months ago
|
||
"testAssemblerBuffer.cpp" contains a variable which is only used in MOZ_ASSERT, which leads to an "unused variable" warning in opt-builds.
Comment 9•5 months ago
|
||
| bugherder | ||
https://hg.mozilla.org/mozilla-central/rev/124004f48da4
https://hg.mozilla.org/mozilla-central/rev/26498574dd66
https://hg.mozilla.org/mozilla-central/rev/76198a20ca5e
https://hg.mozilla.org/mozilla-central/rev/33dd289037a0
Updated•3 months ago
|
Description
•