- User Since
- Jan 15 2015, 3:04 AM (325 w, 2 d)
Fri, Apr 9
Sure thing, thank you for the review. @MaskRay please let me know if you want to further discuss the changes to the patchable-function-entry test, oterwise I will commit this sometime next week.
Wed, Apr 7
Tue, Mar 30
For patchable functions are there any other cases where we wouldn't generate a BTI at the start of the function? If there is we may be alter the test so that it is still checking that the nop is still inserted. If there is no way to avoid a BTI then we may have to alter the expected result and the comment.
I've dug up https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92424 and https://reviews.llvm.org/D72215 for context.
Reintroduced the removed fpatchable-function-entry test, but
modified so that it expects a landing pad at the beginning of the
Fri, Mar 26
@MaskRay sorry to pull you into this review - feel free to ignore if you're not interested.
Mar 12 2021
Nov 10 2020
Nov 9 2020
Diogo, Victor, thanks for the review. I am going to attribute the failure in the HWAddressSanitizer to something else, since it looks x86-related and this patch doesn't touch anything near x86 support.
Nov 2 2020
Jun 5 2020
This is the same logic that we applied to PAC mnemonics in https://reviews.llvm.org/D78372. I'll approve but I would recommend to wait for a few days in case anyone else has something to say.
Apr 24 2020
Apr 21 2020
Another rebase to pass C.I. - looks like it was broken outside this commit
Rebase to trigger C.I.
Apr 17 2020
Mar 5 2020
Jan 22 2020
Jan 16 2020
Jan 13 2020
Updated tests and rebase conflict with new "isAuthenticated" predicate
Jan 2 2020
Dec 18 2019
Sep 30 2019
Sep 27 2019
Reworded all the FP & SIMD constraints
Discussing with @chill on a chat, he was happier with the following wording:
Sep 26 2019
Sep 23 2019
Sep 18 2019
Hi, CMSE upstreaming is indeed one of our priorities. So yes, we are very interested in your feedback. And no, CMSE upstreaming is not abandoned, just going a bit slow ATM :( but any help reviewing is much appreciated! :)
Aug 9 2019
Aug 8 2019
Aug 5 2019
Aug 2 2019
Jul 25 2019
Jul 19 2019
Oooops sorry, not sure how I missed that. I've added SSBS to all the CPUs now.
Jul 18 2019
Add Cortex-A65 and a few features that were missed in the first patch. I think all the features should be in now.
Jul 11 2019
These are trace extensions that will be used by a niche group of developers. They add no instructions (only system registers) and they will not be generated by the compiler. Adding them by default does not have any side-effect, i.e. everyone will still see the same behavior in their code unless they start using these registers on purpose.
Jul 9 2019
Jan 25 2019
Same idea as https://reviews.llvm.org/D54961 but for two other command line options. Approved!
Dec 3 2018
Rebased onto master after a recent refactoring of the AArch64 target parser.
Nov 27 2018
Thank you for the review, Sam :)
Nov 19 2018
The Armv8.5-A specification is available here: https://developer.arm.com/products/architecture/cpu-architecture/a-profile/exploration-tools
Nov 16 2018
Mar 1 2018
Buildbot failing here: http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-bootstrap-msan/builds/2998
Feb 28 2018
Feb 23 2018
If nobody does it before, I'll give it a try early next week.
Feb 15 2018
I'll commit this fix now to prevent other buildbots to fail.
Related fix for a silly errata in one of the tests that is breaking some Windows buildbots:
Committed now. @rengolin many thanks for the review!
What about 32-bit integers?
Feb 13 2018
There is still the possibility that someone tries to use 't' for a vector of two doubles. Only single-precision is allowed in vector operations for 32-bit architectures, so doing something like this would be illegal:
Added tests for int vectors. Allowing integers to go to FP/vector registers is
useful because FP/int conversion instructions (i.e. VCVT) need that.
Feb 12 2018
AFAICS, the current approach just checks the size of the type, not the size of the sub-type. f64 or even integer types could still leak in, no?
To prove they're not, we need tests making sure they break if you try.
Feb 7 2018
Feb 6 2018
Feb 1 2018
Sorry, I missed the sanitizer failure in yesterday's buildbot message and thought it wasn't related to this patch. I will leave some time for @thebolt to fix it, otherwise maybe I can take a look.
Jan 31 2018
Committed now. Thanks for the patch @thebolt!
I'll do it.
Jan 29 2018
LGTM now. In other circumstances, I would wait for someone more experienced than me, but this is a small peephole optimization. I've also tested the patch and it works correctly, so I think it has very little risk.
Jan 26 2018
The logic & testing LGTM. Just a couple of coding standard nits.
Jan 23 2018
Waiting a bit just in case there's extra feedback for the last change. I will commit this patch later today.
Jan 22 2018
VT should be a vector in addTypeForNEON, so the check can be an assert
instead of a proper test in the conditional.
Thanks for the suggestion, very helpful :)
Instead of an ugly switch, reuse already-existing functionality in LLVM to do
the conversion of FP vector types into int vector types.
Jan 18 2018
Jan 17 2018
Fixed big-endian bitconvert patterns and extensive testing for half-float vectors.
Jan 16 2018
I have fixed the bug as olista01 suggested, which is more straightforward than
my previous fix.
Jan 15 2018
Hi Sjoerd, thanks for the review. I have attached some thoughts on your comments and I will upload a new patch soon.
Jan 9 2018
Dec 4 2017
Dec 1 2017
I have also updated the existing tests and added a couple more.
Use !isa<GlobalAddressSDNode> to check if the call is through a function pointer
Nov 17 2016
Thanks all for the reviews!
Cleaned up tests as per Renato's comments.
Nov 16 2016
More testing: add tests for AAPCS and for floating-point arguments to variadic/non-variadic functions.
Could you have a look at this patch? I see no reason to restrict the tail call optimizations to variadic functions: the AAPCS restrictions seem to be already checked for non-variadic functions. The LLVM regression tests passed, and I have also tested with a few benchmarks and they all seem to work fine.
Nov 11 2016
Traverse use list instead of user list to naturally avoid duplicates. Commit message modified accordingly.
Sorry, forget what I said. I didn't get what you were trying to do at first (I've never used the "uses"). I will do the change, as this makes more clear that the operand we are interested in is the condition (i.e. operand 0).
Nov 10 2016
I have investigated the -preserve-ll-uselistorder option. It sets option "ShouldPreserveUseListOrder" to true in the module printer. According to the documentation:
Nov 9 2016
Could you have a look at this fix as per your comments in D26391?