WSL2-Linux-Kernel/arch/x86/kernel/asm-offsets_64.c

80 строки
1.6 KiB
C
Исходник Обычный вид История

#ifndef __LINUX_KBUILD_H
# error "Please do not build this file directly, build asm-offsets.c instead"
#endif
#include <asm/ia32.h>
#define __SYSCALL_64(nr, sym, qual) [nr] = 1,
static char syscalls_64[] = {
#include <asm/syscalls_64.h>
};
#define __SYSCALL_I386(nr, sym, qual) [nr] = 1,
static char syscalls_ia32[] = {
#include <asm/syscalls_32.h>
};
x86/kvm: Provide optimized version of vcpu_is_preempted() for x86-64 It was found when running fio sequential write test with a XFS ramdisk on a KVM guest running on a 2-socket x86-64 system, the %CPU times as reported by perf were as follows: 69.75% 0.59% fio [k] down_write 69.15% 0.01% fio [k] call_rwsem_down_write_failed 67.12% 1.12% fio [k] rwsem_down_write_failed 63.48% 52.77% fio [k] osq_lock 9.46% 7.88% fio [k] __raw_callee_save___kvm_vcpu_is_preempt 3.93% 3.93% fio [k] __kvm_vcpu_is_preempted Making vcpu_is_preempted() a callee-save function has a relatively high cost on x86-64 primarily due to at least one more cacheline of data access from the saving and restoring of registers (8 of them) to and from stack as well as one more level of function call. To reduce this performance overhead, an optimized assembly version of the the __raw_callee_save___kvm_vcpu_is_preempt() function is provided for x86-64. With this patch applied on a KVM guest on a 2-socket 16-core 32-thread system with 16 parallel jobs (8 on each socket), the aggregrate bandwidth of the fio test on an XFS ramdisk were as follows: I/O Type w/o patch with patch -------- --------- ---------- random read 8141.2 MB/s 8497.1 MB/s seq read 8229.4 MB/s 8304.2 MB/s random write 1675.5 MB/s 1701.5 MB/s seq write 1681.3 MB/s 1699.9 MB/s There are some increases in the aggregated bandwidth because of the patch. The perf data now became: 70.78% 0.58% fio [k] down_write 70.20% 0.01% fio [k] call_rwsem_down_write_failed 69.70% 1.17% fio [k] rwsem_down_write_failed 59.91% 55.42% fio [k] osq_lock 10.14% 10.14% fio [k] __kvm_vcpu_is_preempted The assembly code was verified by using a test kernel module to compare the output of C __kvm_vcpu_is_preempted() and that of assembly __raw_callee_save___kvm_vcpu_is_preempt() to verify that they matched. Suggested-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Waiman Long <longman@redhat.com> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2017-02-20 21:36:04 +03:00
#if defined(CONFIG_KVM_GUEST) && defined(CONFIG_PARAVIRT_SPINLOCKS)
#include <asm/kvm_para.h>
#endif
int main(void)
{
#ifdef CONFIG_PARAVIRT
OFFSET(PV_CPU_usergs_sysret64, pv_cpu_ops, usergs_sysret64);
OFFSET(PV_CPU_swapgs, pv_cpu_ops, swapgs);
BLANK();
#endif
x86/kvm: Provide optimized version of vcpu_is_preempted() for x86-64 It was found when running fio sequential write test with a XFS ramdisk on a KVM guest running on a 2-socket x86-64 system, the %CPU times as reported by perf were as follows: 69.75% 0.59% fio [k] down_write 69.15% 0.01% fio [k] call_rwsem_down_write_failed 67.12% 1.12% fio [k] rwsem_down_write_failed 63.48% 52.77% fio [k] osq_lock 9.46% 7.88% fio [k] __raw_callee_save___kvm_vcpu_is_preempt 3.93% 3.93% fio [k] __kvm_vcpu_is_preempted Making vcpu_is_preempted() a callee-save function has a relatively high cost on x86-64 primarily due to at least one more cacheline of data access from the saving and restoring of registers (8 of them) to and from stack as well as one more level of function call. To reduce this performance overhead, an optimized assembly version of the the __raw_callee_save___kvm_vcpu_is_preempt() function is provided for x86-64. With this patch applied on a KVM guest on a 2-socket 16-core 32-thread system with 16 parallel jobs (8 on each socket), the aggregrate bandwidth of the fio test on an XFS ramdisk were as follows: I/O Type w/o patch with patch -------- --------- ---------- random read 8141.2 MB/s 8497.1 MB/s seq read 8229.4 MB/s 8304.2 MB/s random write 1675.5 MB/s 1701.5 MB/s seq write 1681.3 MB/s 1699.9 MB/s There are some increases in the aggregated bandwidth because of the patch. The perf data now became: 70.78% 0.58% fio [k] down_write 70.20% 0.01% fio [k] call_rwsem_down_write_failed 69.70% 1.17% fio [k] rwsem_down_write_failed 59.91% 55.42% fio [k] osq_lock 10.14% 10.14% fio [k] __kvm_vcpu_is_preempted The assembly code was verified by using a test kernel module to compare the output of C __kvm_vcpu_is_preempted() and that of assembly __raw_callee_save___kvm_vcpu_is_preempt() to verify that they matched. Suggested-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Waiman Long <longman@redhat.com> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2017-02-20 21:36:04 +03:00
#if defined(CONFIG_KVM_GUEST) && defined(CONFIG_PARAVIRT_SPINLOCKS)
OFFSET(KVM_STEAL_TIME_preempted, kvm_steal_time, preempted);
BLANK();
#endif
#define ENTRY(entry) OFFSET(pt_regs_ ## entry, pt_regs, entry)
ENTRY(bx);
ENTRY(cx);
ENTRY(dx);
ENTRY(sp);
ENTRY(bp);
ENTRY(si);
ENTRY(di);
ENTRY(r8);
ENTRY(r9);
ENTRY(r10);
ENTRY(r11);
ENTRY(r12);
ENTRY(r13);
ENTRY(r14);
ENTRY(r15);
ENTRY(flags);
BLANK();
#undef ENTRY
#define ENTRY(entry) OFFSET(saved_context_ ## entry, saved_context, entry)
ENTRY(cr0);
ENTRY(cr2);
ENTRY(cr3);
ENTRY(cr4);
ENTRY(cr8);
x86, gdt, hibernate: Store/load GDT for hibernate path. The git commite7a5cd063c7b4c58417f674821d63f5eb6747e37 ("x86-64, gdt: Store/load GDT for ACPI S3 or hibernate/resume path is not needed.") assumes that for the hibernate path the booting kernel and the resuming kernel MUST be the same. That is certainly the case for a 32-bit kernel (see check_image_kernel and CONFIG_ARCH_HIBERNATION_HEADER config option). However for 64-bit kernels it is OK to have a different kernel version (and size of the image) of the booting and resuming kernels. Hence the above mentioned git commit introduces an regression. This patch fixes it by introducing a 'struct desc_ptr gdt_desc' back in the 'struct saved_context'. However instead of having in the 'save_processor_state' and 'restore_processor_state' the store/load_gdt calls, we are only saving the GDT in the save_processor_state. For the restore path the lgdt operation is done in hibernate_asm_[32|64].S in the 'restore_registers' path. The apt reader of this description will recognize that only 64-bit kernels need this treatment, not 32-bit. This patch adds the logic in the 32-bit path to be more similar to 64-bit so that in the future the unification process can take advantage of this. [ hpa: this also reverts an inadvertent on-disk format change ] Suggested-by: "H. Peter Anvin" <hpa@zytor.com> Acked-by: "Rafael J. Wysocki" <rjw@sisk.pl> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Link: http://lkml.kernel.org/r/1367459610-9656-2-git-send-email-konrad.wilk@oracle.com Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2013-05-02 05:53:30 +04:00
ENTRY(gdt_desc);
BLANK();
#undef ENTRY
OFFSET(TSS_ist, tss_struct, x86_tss.ist);
OFFSET(TSS_sp0, tss_struct, x86_tss.sp0);
BLANK();
#ifdef CONFIG_CC_STACKPROTECTOR
DEFINE(stack_canary_offset, offsetof(union irq_stack_union, stack_canary));
BLANK();
#endif
DEFINE(__NR_syscall_max, sizeof(syscalls_64) - 1);
DEFINE(NR_syscalls, sizeof(syscalls_64));
DEFINE(__NR_syscall_compat_max, sizeof(syscalls_ia32) - 1);
DEFINE(IA32_NR_syscalls, sizeof(syscalls_ia32));
return 0;
}