https://bugs.kde.org/show_bug.cgi?id=412377
Paul Floyd changed:
What|Removed |Added
Resolution|--- |FIXED
Status|REPORTED
https://bugs.kde.org/show_bug.cgi?id=412377
Paul Floyd changed:
What|Removed |Added
Assignee|jsew...@acm.org |pjfl...@wanadoo.fr
--
You are receiving this
https://bugs.kde.org/show_bug.cgi?id=412377
--- Comment #7 from Paul Floyd ---
In the end it was me being stupid. I missed out the conditional block in
VG_(invalidate_icache)
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=412377
Paul Floyd changed:
What|Removed |Added
CC||pjfl...@wanadoo.fr
--- Comment #6 from Paul Floyd
https://bugs.kde.org/show_bug.cgi?id=412377
--- Comment #5 from Adam Borowski ---
You've added civac which differs from cvac only by trashing the cache entry
after committing it to memory ("i" stands for "invalidate"). cvac is the
instruction I need, and I'm using it successfully. It works on
https://bugs.kde.org/show_bug.cgi?id=412377
Adam Borowski changed:
What|Removed |Added
Attachment #122885|0 |1
is obsolete|
https://bugs.kde.org/show_bug.cgi?id=412377
--- Comment #3 from Julian Seward ---
My problem is that it is safer if we don't have untested code in the tree.
That's why I didn't take your original patch. If you can post the minimum-
change patch that provides |cvac|, relative to the current
https://bugs.kde.org/show_bug.cgi?id=412377
--- Comment #2 from Adam Borowski ---
Alas, nope -- I specifically need the variant without invalidation. On ARMv8.0
and 8.1 that's cvac; what I really want is cvap but that's unsupported on any
hardware I have access to.
Mainline valgrind (ie,
https://bugs.kde.org/show_bug.cgi?id=412377
--- Comment #1 from Julian Seward ---
In c82d35f6d67ed34cf20d79f90a7400bd7f83ebad just now I added
support for 'dc civac', which is the only one of the variants I
have a way to test right now. Is that a good enough fix for your
use case(s) ?
--
You