- User Since
- Sep 22 2017, 3:13 PM (73 w, 3 d)
Fri, Jan 25
Tue, Jan 22
Jan 9 2019
Jan 7 2019
Jan 4 2019
Removed treatment for out of bounds constant, as it is simplified to poison before being transformed to VGETLANE.
Also removed treatment for variable index, as it is not accepted by VGETLANE.
Added extra tests.
Jan 3 2019
Dec 29 2018
Dec 28 2018
Track a single element if the index is a constant.
Dec 27 2018
Dec 21 2018
Remove empty feature. Rename a few more test files to have sb instead of specRest.
Removed empty feature
Dec 20 2018
Dec 6 2018
Trying to fix a windows bot failing test, dividing the test into multiple files and using full line matching.
Dec 3 2018
Nov 28 2018
Fixed one comment.
Nov 27 2018
NIT in the v8.4a trace tests.
Fix the tlb-rmi feature argument and test file.
Removed the register (system operand) constraint, and associated test case.
Fixed PAN, RCPC dependencies and instructions that depend on PA, not PAN.
Changed some names to be more meaningful.
Removed the prefix "has" from the features flags.
Nov 26 2018
the features implications/dependencies.
some test files.
Nov 23 2018
Indeed I did not. As you said, they are not visible to the user. I'll check with the gcc devs how they have these features split and named though.
For the feature registers (ISAR6 and MMFR2), maybe we should just make these registers always available? These registers are all defined to read as all-zeroes in older architectures, which means no features described by that register are implemented. If we continue like this then we will have to add an extra feature to pretty much every new architectural feature for the ID register.
Indeed it seems the case for the ID_ISAR6_EL1 exist, so I'll just remove the constraints over it.
However, the ID_AA64MMFR2_EL1 register was added in v8.2. So I believe we should check as to warn the user if he might get undefined behavior by accessing it when not present.
Nov 19 2018
Added FeatureRASv8_4 implies RAS.
Nov 16 2018
Nov 7 2018
Nov 6 2018
Ups, it was ment to be a clang patch, not a llvm one. Sry
Oct 30 2018
Added blank lines at end of test files.
Oct 29 2018
Re-inserted the missing tests. Small refactoring as to look alike neighbor code.
Oct 23 2018
Oct 19 2018
Oct 16 2018
No custom decoding is required
Fixed the tests as to check boundaries, and one value in the middle.
Changed tests ordering as they generated empty lines not treated by FileCheck.
To send patch....
Oct 11 2018
Removed processor and specific triple target from test.
Oct 2 2018
Updated test to use llvm-mc instead of clang.
Simplified tests. Using the standard function for emitting data symbols, if required.
Oct 1 2018
Sep 12 2018
Sep 7 2018
Sep 6 2018
Fix test march triple.