diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2016-03-17 20:03:47 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2016-03-17 20:03:47 -0700 |
commit | 588ab3f9afdfa1a6b1e5761c858b2c4ab6098285 (patch) | |
tree | c9aa4c4f8a63d25c3cf05330c68948dceec79cc2 /arch/arm64/lib/copy_page.S | |
parent | 3d15cfdb1b77536c205d8e49c0312219ddf162ec (diff) | |
parent | 2776e0e8ef683a42fe3e9a5facf576b73579700e (diff) | |
download | linux-588ab3f9afdfa1a6b1e5761c858b2c4ab6098285.tar.gz linux-588ab3f9afdfa1a6b1e5761c858b2c4ab6098285.tar.bz2 linux-588ab3f9afdfa1a6b1e5761c858b2c4ab6098285.zip |
Merge tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 updates from Catalin Marinas:
"Here are the main arm64 updates for 4.6. There are some relatively
intrusive changes to support KASLR, the reworking of the kernel
virtual memory layout and initial page table creation.
Summary:
- Initial page table creation reworked to avoid breaking large block
mappings (huge pages) into smaller ones. The ARM architecture
requires break-before-make in such cases to avoid TLB conflicts but
that's not always possible on live page tables
- Kernel virtual memory layout: the kernel image is no longer linked
to the bottom of the linear mapping (PAGE_OFFSET) but at the bottom
of the vmalloc space, allowing the kernel to be loaded (nearly)
anywhere in physical RAM
- Kernel ASLR: position independent kernel Image and modules being
randomly mapped in the vmalloc space with the randomness is
provided by UEFI (efi_get_random_bytes() patches merged via the
arm64 tree, acked by Matt Fleming)
- Implement relative exception tables for arm64, required by KASLR
(initial code for ARCH_HAS_RELATIVE_EXTABLE added to lib/extable.c
but actual x86 conversion to deferred to 4.7 because of the merge
dependencies)
- Support for the User Access Override feature of ARMv8.2: this
allows uaccess functions (get_user etc.) to be implemented using
LDTR/STTR instructions. Such instructions, when run by the kernel,
perform unprivileged accesses adding an extra level of protection.
The set_fs() macro is used to "upgrade" such instruction to
privileged accesses via the UAO bit
- Half-precision floating point support (part of ARMv8.2)
- Optimisations for CPUs with or without a hardware prefetcher (using
run-time code patching)
- copy_page performance improvement to deal with 128 bytes at a time
- Sanity checks on the CPU capabilities (via CPUID) to prevent
incompatible secondary CPUs from being brought up (e.g. weird
big.LITTLE configurations)
- valid_user_regs() reworked for better sanity check of the
sigcontext information (restored pstate information)
- ACPI parking protocol implementation
- CONFIG_DEBUG_RODATA enabled by default
- VDSO code marked as read-only
- DEBUG_PAGEALLOC support
- ARCH_HAS_UBSAN_SANITIZE_ALL enabled
- Erratum workaround Cavium ThunderX SoC
- set_pte_at() fix for PROT_NONE mappings
- Code clean-ups"
* tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: (99 commits)
arm64: kasan: Fix zero shadow mapping overriding kernel image shadow
arm64: kasan: Use actual memory node when populating the kernel image shadow
arm64: Update PTE_RDONLY in set_pte_at() for PROT_NONE permission
arm64: Fix misspellings in comments.
arm64: efi: add missing frame pointer assignment
arm64: make mrs_s prefixing implicit in read_cpuid
arm64: enable CONFIG_DEBUG_RODATA by default
arm64: Rework valid_user_regs
arm64: mm: check at build time that PAGE_OFFSET divides the VA space evenly
arm64: KVM: Move kvm_call_hyp back to its original localtion
arm64: mm: treat memstart_addr as a signed quantity
arm64: mm: list kernel sections in order
arm64: lse: deal with clobbered IP registers after branch via PLT
arm64: mm: dump: Use VA_START directly instead of private LOWEST_ADDR
arm64: kconfig: add submenu for 8.2 architectural features
arm64: kernel: acpi: fix ioremap in ACPI parking protocol cpu_postboot
arm64: Add support for Half precision floating point
arm64: Remove fixmap include fragility
arm64: Add workaround for Cavium erratum 27456
arm64: mm: Mark .rodata as RO
...
Diffstat (limited to 'arch/arm64/lib/copy_page.S')
-rw-r--r-- | arch/arm64/lib/copy_page.S | 63 |
1 files changed, 55 insertions, 8 deletions
diff --git a/arch/arm64/lib/copy_page.S b/arch/arm64/lib/copy_page.S index 512b9a7b980e..4c1e700840b6 100644 --- a/arch/arm64/lib/copy_page.S +++ b/arch/arm64/lib/copy_page.S @@ -18,6 +18,8 @@ #include <linux/const.h> #include <asm/assembler.h> #include <asm/page.h> +#include <asm/cpufeature.h> +#include <asm/alternative.h> /* * Copy a page from src to dest (both are page aligned) @@ -27,20 +29,65 @@ * x1 - src */ ENTRY(copy_page) - /* Assume cache line size is 64 bytes. */ - prfm pldl1strm, [x1, #64] -1: ldp x2, x3, [x1] +alternative_if_not ARM64_HAS_NO_HW_PREFETCH + nop + nop +alternative_else + # Prefetch two cache lines ahead. + prfm pldl1strm, [x1, #128] + prfm pldl1strm, [x1, #256] +alternative_endif + + ldp x2, x3, [x1] ldp x4, x5, [x1, #16] ldp x6, x7, [x1, #32] ldp x8, x9, [x1, #48] - add x1, x1, #64 - prfm pldl1strm, [x1, #64] + ldp x10, x11, [x1, #64] + ldp x12, x13, [x1, #80] + ldp x14, x15, [x1, #96] + ldp x16, x17, [x1, #112] + + mov x18, #(PAGE_SIZE - 128) + add x1, x1, #128 +1: + subs x18, x18, #128 + +alternative_if_not ARM64_HAS_NO_HW_PREFETCH + nop +alternative_else + prfm pldl1strm, [x1, #384] +alternative_endif + stnp x2, x3, [x0] + ldp x2, x3, [x1] stnp x4, x5, [x0, #16] + ldp x4, x5, [x1, #16] stnp x6, x7, [x0, #32] + ldp x6, x7, [x1, #32] stnp x8, x9, [x0, #48] - add x0, x0, #64 - tst x1, #(PAGE_SIZE - 1) - b.ne 1b + ldp x8, x9, [x1, #48] + stnp x10, x11, [x0, #64] + ldp x10, x11, [x1, #64] + stnp x12, x13, [x0, #80] + ldp x12, x13, [x1, #80] + stnp x14, x15, [x0, #96] + ldp x14, x15, [x1, #96] + stnp x16, x17, [x0, #112] + ldp x16, x17, [x1, #112] + + add x0, x0, #128 + add x1, x1, #128 + + b.gt 1b + + stnp x2, x3, [x0] + stnp x4, x5, [x0, #16] + stnp x6, x7, [x0, #32] + stnp x8, x9, [x0, #48] + stnp x10, x11, [x0, #64] + stnp x12, x13, [x0, #80] + stnp x14, x15, [x0, #96] + stnp x16, x17, [x0, #112] + ret ENDPROC(copy_page) |