summaryrefslogtreecommitdiff
path: root/arch/arm64/include/asm/fpsimdmacros.h
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2021-11-01 16:33:53 -0700
committerLinus Torvalds <torvalds@linux-foundation.org>2021-11-01 16:33:53 -0700
commit46f876322820c189ab525cfcba2519a17dbc0a6f (patch)
tree324a1c3c0872bce82bac839f64816ee810e4b5bd /arch/arm64/include/asm/fpsimdmacros.h
parent879dbe9ffebc1328717cd66eab7e4918a3f499bd (diff)
parente6359798f62da66a4a48061d2324a69ea59ff39b (diff)
downloadlinux-46f876322820c189ab525cfcba2519a17dbc0a6f.tar.gz
Merge tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 updates from Will Deacon: "There's the usual summary below, but the highlights are support for the Armv8.6 timer extensions, KASAN support for asymmetric MTE, the ability to kexec() with the MMU enabled and a second attempt at switching to the generic pfn_valid() implementation. Summary: - Support for the Arm8.6 timer extensions, including a self-synchronising view of the system registers to elide some expensive ISB instructions. - Exception table cleanup and rework so that the fixup handlers appear correctly in backtraces. - A handful of miscellaneous changes, the main one being selection of CONFIG_HAVE_POSIX_CPU_TIMERS_TASK_WORK. - More mm and pgtable cleanups. - KASAN support for "asymmetric" MTE, where tag faults are reported synchronously for loads (via an exception) and asynchronously for stores (via a register). - Support for leaving the MMU enabled during kexec relocation, which significantly speeds up the operation. - Minor improvements to our perf PMU drivers. - Improvements to the compat vDSO build system, particularly when building with LLVM=1. - Preparatory work for handling some Coresight TRBE tracing errata. - Cleanup and refactoring of the SVE code to pave the way for SME support in future. - Ensure SCS pages are unpoisoned immediately prior to freeing them when KASAN is enabled for the vmalloc area. - Try moving to the generic pfn_valid() implementation again now that the DMA mapping issue from last time has been resolved. - Numerous improvements and additions to our FPSIMD and SVE selftests" [ armv8.6 timer updates were in a shared branch and already came in through -tip in the timer pull - Linus ] * tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: (85 commits) arm64: Select POSIX_CPU_TIMERS_TASK_WORK arm64: Document boot requirements for FEAT_SME_FA64 arm64/sve: Fix warnings when SVE is disabled arm64/sve: Add stub for sve_max_virtualisable_vl() arm64: errata: Add detection for TRBE write to out-of-range arm64: errata: Add workaround for TSB flush failures arm64: errata: Add detection for TRBE overwrite in FILL mode arm64: Add Neoverse-N2, Cortex-A710 CPU part definition selftests: arm64: Factor out utility functions for assembly FP tests arm64: vmlinux.lds.S: remove `.fixup` section arm64: extable: add load_unaligned_zeropad() handler arm64: extable: add a dedicated uaccess handler arm64: extable: add `type` and `data` fields arm64: extable: use `ex` for `exception_table_entry` arm64: extable: make fixup_exception() return bool arm64: extable: consolidate definitions arm64: gpr-num: support W registers arm64: factor out GPR numbering helpers arm64: kvm: use kvm_exception_table_entry arm64: lib: __arch_copy_to_user(): fold fixups into body ...
Diffstat (limited to 'arch/arm64/include/asm/fpsimdmacros.h')
-rw-r--r--arch/arm64/include/asm/fpsimdmacros.h21
1 files changed, 12 insertions, 9 deletions
diff --git a/arch/arm64/include/asm/fpsimdmacros.h b/arch/arm64/include/asm/fpsimdmacros.h
index 00a2c0b69c2b..2509d7dde55a 100644
--- a/arch/arm64/include/asm/fpsimdmacros.h
+++ b/arch/arm64/include/asm/fpsimdmacros.h
@@ -217,28 +217,36 @@
.macro sve_flush_z
_for n, 0, 31, _sve_flush_z \n
.endm
-.macro sve_flush_p_ffr
+.macro sve_flush_p
_for n, 0, 15, _sve_pfalse \n
+.endm
+.macro sve_flush_ffr
_sve_wrffr 0
.endm
-.macro sve_save nxbase, xpfpsr, nxtmp
+.macro sve_save nxbase, xpfpsr, save_ffr, nxtmp
_for n, 0, 31, _sve_str_v \n, \nxbase, \n - 34
_for n, 0, 15, _sve_str_p \n, \nxbase, \n - 16
+ cbz \save_ffr, 921f
_sve_rdffr 0
_sve_str_p 0, \nxbase
_sve_ldr_p 0, \nxbase, -16
-
+ b 922f
+921:
+ str xzr, [x\nxbase] // Zero out FFR
+922:
mrs x\nxtmp, fpsr
str w\nxtmp, [\xpfpsr]
mrs x\nxtmp, fpcr
str w\nxtmp, [\xpfpsr, #4]
.endm
-.macro __sve_load nxbase, xpfpsr, nxtmp
+.macro sve_load nxbase, xpfpsr, restore_ffr, nxtmp
_for n, 0, 31, _sve_ldr_v \n, \nxbase, \n - 34
+ cbz \restore_ffr, 921f
_sve_ldr_p 0, \nxbase
_sve_wrffr 0
+921:
_for n, 0, 15, _sve_ldr_p \n, \nxbase, \n - 16
ldr w\nxtmp, [\xpfpsr]
@@ -246,8 +254,3 @@
ldr w\nxtmp, [\xpfpsr, #4]
msr fpcr, x\nxtmp
.endm
-
-.macro sve_load nxbase, xpfpsr, xvqminus1, nxtmp, xtmp2
- sve_load_vq \xvqminus1, x\nxtmp, \xtmp2
- __sve_load \nxbase, \xpfpsr, \nxtmp
-.endm