linux_dsm_epyc7002/arch/x86/kernel
Jeremy Fitzhardinge ecb93d1ccd x86/paravirt: add register-saving thunks to reduce caller register pressure
Impact: Optimization

One of the problems with inserting a pile of C calls where previously
there were none is that the register pressure is greatly increased.
The C calling convention says that the caller must expect a certain
set of registers may be trashed by the callee, and that the callee can
use those registers without restriction.  This includes the function
argument registers, and several others.

This patch seeks to alleviate this pressure by introducing wrapper
thunks that will do the register saving/restoring, so that the
callsite doesn't need to worry about it, but the callee function can
be conventional compiler-generated code.  In many cases (particularly
performance-sensitive cases) the callee will be in assembler anyway,
and need not use the compiler's calling convention.

Standard calling convention is:
	 arguments	    return	scratch
x86-32	 eax edx ecx	    eax		?
x86-64	 rdi rsi rdx rcx    rax		r8 r9 r10 r11

The thunk preserves all argument and scratch registers.  The return
register is not preserved, and is available as a scratch register for
unwrapped callee code (and of course the return value).

Wrapped function pointers are themselves wrapped in a struct
paravirt_callee_save structure, in order to get some warning from the
compiler when functions with mismatched calling conventions are used.

The most common paravirt ops, both statically and dynamically, are
interrupt enable/disable/save/restore, so handle them first.  This is
particularly easy since their calls are handled specially anyway.

XXX Deal with VMI.  What's their calling convention?

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
2009-01-30 14:51:45 -08:00
..
acpi x86: misc clean up after the percpu update 2009-01-16 14:20:26 +01:00
cpu Merge branch 'tj-percpu' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc into core/percpu 2009-01-27 12:03:24 +01:00
.gitignore
alternative.c x86: improve UP kernel when CPU-hotplug and SMP is enabled 2008-10-13 10:33:46 +02:00
amd_iommu_init.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/linux-2.6-iommu 2009-01-03 12:03:52 -08:00
amd_iommu.c AMD IOMMU: remove now unnecessary #ifdefs 2009-01-03 16:44:15 +01:00
aperture_64.c aperture_64.c: clarify that too small aperture is valid reason for this code 2008-11-28 15:24:39 +01:00
apic.c x86: move apic variables to apic.c 2009-01-27 12:56:47 +09:00
apm_32.c x86: update Alan Cox's email addresses 2009-01-05 15:19:16 +01:00
asm-offsets_32.c x86: signal: move sigframe.h to arch/x86/include/asm 2008-12-18 11:28:54 +01:00
asm-offsets_64.c x86: remove pda.h 2009-01-20 12:29:20 +09:00
asm-offsets.c
audit_64.c
bios_uv.c x86: bios_uv.c: uv_systab should be static 2008-12-29 22:08:28 -08:00
bootflag.c
check.c x86: corruption-check: some post-move cleanups 2008-10-27 18:09:45 +01:00
cpuid.c Merge branch 'cpus4096-for-linus-3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip 2009-01-03 12:04:39 -08:00
crash_dump_32.c kdump: make elfcorehdr_addr independent of CONFIG_PROC_VMCORE 2008-10-20 08:52:39 -07:00
crash_dump_64.c kdump: make elfcorehdr_addr independent of CONFIG_PROC_VMCORE 2008-10-20 08:52:39 -07:00
crash.c x86: smp.h move safe_smp_processor_id declartion to cpu.h 2009-01-07 21:48:25 +01:00
doublefault_32.c x86: fix virt_addr_valid() with CONFIG_DEBUG_VIRTUAL=y, v2 2008-10-13 10:33:15 +02:00
ds.c Merge branch 'x86/ptrace' into x86/tsc 2008-12-23 16:29:31 +01:00
dumpstack_32.c ftrace: print real return in dumpstack for function graph 2008-12-03 08:56:25 +01:00
dumpstack_64.c x86-64: Move current task from PDA to per-cpu and consolidate with 32-bit. 2009-01-19 00:38:58 +09:00
dumpstack.c ftrace: print real return in dumpstack for function graph 2008-12-03 08:56:25 +01:00
dumpstack.h ftrace: print real return in dumpstack for function graph 2008-12-03 08:56:25 +01:00
e820.c Merge branch 'linus' into release 2009-01-09 03:39:43 -05:00
early_printk.c x86: early_printk - use sizeof instead of hardcoded number 2009-01-02 10:27:46 +01:00
early-quirks.c x86: only scan the root bus in early PCI quirks 2009-01-09 12:46:22 -08:00
efi_32.c x86: fix i486 suspend to disk CR4 oops 2008-08-18 08:50:19 +02:00
efi_64.c x86: remove include of apic.h from hardirq_64.h 2009-01-23 11:03:29 +09:00
efi_stub_32.S
efi_stub_64.S
efi.c x86: make UV support configurable 2009-01-21 13:00:42 +01:00
entry_32.S x86: make x86_32 use tlb_64.c 2009-01-21 17:26:06 +09:00
entry_64.S x86: fix paravirt clobber in entry_64.S 2009-01-30 14:51:44 -08:00
es7000_32.c Merge branch 'linus' into x86/quirks 2008-12-08 15:07:49 +01:00
ftrace.c tracing/function-graph-tracer: append the tracing_graph_flag 2008-12-08 15:11:45 +01:00
genapic_64.c x86: make UV support configurable 2009-01-21 13:00:42 +01:00
genapic_flat_64.c x86: fix cpu_mask_to_apicid_and to include cpu_online_mask 2008-12-18 11:59:24 +01:00
genx2apic_cluster.c x86: use logical apicid in x2apic_cluster's x2apic_cpu_mask_to_apicid_and() 2008-12-23 22:37:30 +01:00
genx2apic_phys.c Merge branch 'linus' into x86/cleanups 2009-01-02 22:41:36 +01:00
genx2apic_uv_x.c x86: uv cleanup 2009-01-21 17:26:06 +09:00
geode_32.c
head32.c x86: change static allocation of trampoline area 2008-12-08 13:49:45 +01:00
head64.c x86: remove pda_init() 2009-01-20 12:29:19 +09:00
head_32.S x86: set %fs to __KERNEL_PERCPU unconditionally for x86_32 2009-01-21 17:26:05 +09:00
head_64.S x86: move stack_canary into irq_stack 2009-01-20 12:29:20 +09:00
head.c x86, debug: remove EBDA debug printk 2008-12-12 11:08:42 +01:00
hpet.c debugobjects: add and use INIT_WORK_ON_STACK 2009-01-22 10:02:07 +01:00
i386_ksyms_32.c ftrace: rename FTRACE to FUNCTION_TRACER 2008-10-20 18:27:03 +02:00
i387.c x86: fix __cpuinit/__init tangle in init_thread_xstate() 2008-11-20 16:43:41 +01:00
i8237.c
i8253.c cpumask: convert struct clock_event_device to cpumask pointers. 2008-12-13 21:20:26 +10:30
i8259.c x86: i8259.c fix style problems 2009-01-04 18:19:37 +01:00
init_task.c take init_fs to saner place 2008-12-31 18:07:42 -05:00
io_apic.c x86: make UV support configurable 2009-01-21 13:00:42 +01:00
io_delay.c x86: add io delay quirk for Presario F700 2008-09-03 16:42:51 -07:00
ioport.c x86: ioport.c fix style problems 2009-01-04 13:23:07 +01:00
ipi.c x86: cosmetic changes apic-related files. 2008-12-16 17:40:57 -08:00
irq_32.c cpumask: update irq_desc to use cpumask_var_t 2009-01-11 19:12:46 +01:00
irq_64.c x86: remove include of apic.h from hardirq_64.h 2009-01-23 11:03:29 +09:00
irq.c x86-64: Move irq stats from PDA to per-cpu and consolidate with 32-bit. 2009-01-19 00:38:57 +09:00
irqinit_32.c x86: make x86_32 use tlb_64.c 2009-01-21 17:26:06 +09:00
irqinit_64.c x86: irqinit_64.c init_ISA_irqs should be static 2009-01-04 18:19:41 +01:00
k8.c x86, gart: fix gart detection for Fam11h CPUs 2008-10-28 17:10:27 +01:00
kdebugfs.c arch/x86/kernel/kdebugfs.c: introduce missing kfree 2008-09-10 14:03:49 +02:00
kgdb.c kgdb, x86: Avoid invoking kgdb_nmicallback twice per NMI 2008-10-06 13:50:59 -05:00
kprobes.c kprobes: check CONFIG_FREEZER instead of CONFIG_PM 2009-01-16 14:32:17 -05:00
kvm.c x86, paravirt_ops: use unsigned long instead of u32 for alloc_p*() pfn args 2008-08-22 05:34:44 +02:00
kvmclock.c x86: KVM guest: kvm_get_tsc_khz: return khz, not lpj 2008-12-31 16:55:44 +02:00
ldt.c x86: ldt.c fix style problems 2009-01-02 17:46:24 +01:00
machine_kexec_32.c kexec/i386: setup kexec page table in C 2008-10-31 10:01:57 +01:00
machine_kexec_64.c kexec jump 2008-07-26 12:00:04 -07:00
Makefile x86: initialize per-cpu GDT segment in per-cpu setup 2009-01-27 12:56:48 +09:00
mca_32.c
mfgpt_32.c trivial: fix then -> than typos in comments and documentation 2009-01-06 11:28:06 +01:00
microcode_amd.c x86: fix warning in arch/x86/kernel/microcode_amd.c 2008-12-19 01:31:54 +01:00
microcode_core.c Merge branches 'x86/apic', 'x86/cleanups', 'x86/cpufeature', 'x86/crashdump', 'x86/debug', 'x86/defconfig', 'x86/detect-hyper', 'x86/doc', 'x86/dumpstack', 'x86/early-printk', 'x86/fpu', 'x86/idle', 'x86/io', 'x86/memory-corruption-check', 'x86/microcode', 'x86/mm', 'x86/mtrr', 'x86/nmi-watchdog', 'x86/pat2', 'x86/pci-ioapic-boot-irq-quirks', 'x86/ptrace', 'x86/quirks', 'x86/reboot', 'x86/setup-memory', 'x86/signal', 'x86/sparse-fixes', 'x86/time', 'x86/uv' and 'x86/xen' into x86/core 2008-12-23 16:27:23 +01:00
microcode_intel.c x86: microcode_intel.c fix style problems 2009-01-12 11:22:40 +01:00
mmconf-fam10h_64.c x86, pci: move arch/x86/pci/pci.h to arch/x86/include/asm/pci_x86.h 2008-12-29 18:17:36 +01:00
module_32.c x86: module_32.c fix style problems 2009-01-12 11:22:55 +01:00
module_64.c x86: module_64.c fix style problems 2009-01-12 11:23:01 +01:00
mpparse.c Merge branches 'cpus4096', 'x86/cleanups' and 'x86/urgent' into x86/percpu 2009-01-15 13:18:57 +01:00
msr.c x86: msr.c fix style problems 2009-01-12 11:22:50 +01:00
nmi.c x86-64: Move irq stats from PDA to per-cpu and consolidate with 32-bit. 2009-01-19 00:38:57 +09:00
numaq_32.c x86: rename all fields of mpc_oemtable oem_X to X 2009-01-05 14:08:34 +01:00
olpc.c x86, olpc: fix endian bug in openfirmware workaround 2008-09-24 10:29:04 +02:00
paravirt_patch_32.c x86/pvops: add a paravirt_ident functions to allow special patching 2009-01-30 14:51:44 -08:00
paravirt_patch_64.c x86/pvops: add a paravirt_ident functions to allow special patching 2009-01-30 14:51:44 -08:00
paravirt-spinlocks.c x86: fix default_spin_lock_flags() prototype 2008-12-08 16:08:29 +01:00
paravirt.c x86/paravirt: add register-saving thunks to reduce caller register pressure 2009-01-30 14:51:45 -08:00
pci-calgary_64.c arch/x86/kernel/pci-calgary_64.c: change simple_strtol to simple_strtoul 2008-11-25 15:56:03 +01:00
pci-dma.c PCI: struct device - replace bus_id with dev_name(), dev_set_name() 2009-01-07 11:12:23 -08:00
pci-gart_64.c Documentation: move DMA-mapping.txt to Doc/PCI/ 2009-01-29 18:19:29 -08:00
pci-nommu.c x86: export pci-nommu's alloc_coherent 2008-09-25 11:02:25 +02:00
pci-swiotlb_64.c Merge branch 'core/iommu' into core/urgent 2009-01-05 14:17:24 +01:00
pcspeaker.c x86: use platform_device_register_simple() 2008-09-22 12:58:36 +02:00
pmtimer_64.c
probe_roms_32.c
process_32.c x86: remove idle_timestamp from 32bit irq_cpustat_t 2009-01-23 11:03:28 +09:00
process_64.c x86: remove pda.h 2009-01-20 12:29:20 +09:00
process.c Merge branch 'tracing-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip 2008-12-28 12:21:10 -08:00
ptrace.c x86, bts: memory accounting 2008-12-20 09:15:47 +01:00
pvclock.c x86: pvclock: fix shadowed variable warning 2008-10-15 14:25:14 +02:00
quirks.c x86: enable HPET on Fujitsu u9200 2008-12-16 20:36:44 +01:00
reboot_fixups_32.c
reboot.c x86: smp.h move safe_smp_processor_id declartion to cpu.h 2009-01-07 21:48:25 +01:00
relocate_kernel_32.S kexec/i386: setup kexec page table in C 2008-10-31 10:01:57 +01:00
relocate_kernel_64.S
rtc.c rtc: use bcd2bin/bin2bcd 2008-10-20 08:52:41 -07:00
scx200_32.c
setup_percpu.c Merge branch 'linus' into core/percpu 2009-01-27 12:01:51 +01:00
setup.c x86: smp.h move prefill_possible_map declartion to cpu.h 2009-01-07 13:51:21 +01:00
signal.c Revert "x86: signal: change type of paramter for sys_rt_sigreturn()" 2009-01-21 09:43:18 +01:00
smp.c Merge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip 2009-01-10 06:13:09 -08:00
smpboot.c x86: initialize per-cpu GDT segment in per-cpu setup 2009-01-27 12:56:48 +09:00
stacktrace.c tracing/stack-tracer: fix style issues 2008-11-23 11:53:48 +01:00
step.c
summit_32.c x86: mach_summit to summit 2008-07-26 16:31:35 +02:00
sys_i386_32.c x86: Introducing asm/syscalls.h 2008-07-22 14:35:57 +02:00
sys_x86_64.c Merge branches 'x86/alternatives', 'x86/cleanups', 'x86/commandline', 'x86/crashdump', 'x86/debug', 'x86/defconfig', 'x86/doc', 'x86/exports', 'x86/fpu', 'x86/gart', 'x86/idle', 'x86/mm', 'x86/mtrr', 'x86/nmi-watchdog', 'x86/oprofile', 'x86/paravirt', 'x86/reboot', 'x86/sparse-fixes', 'x86/tsc', 'x86/urgent' and 'x86/vmalloc' into x86-v28-for-linus-phase1 2008-10-06 18:17:07 +02:00
syscall_64.c x86: Fix ASM_X86__ header guards 2008-10-22 22:55:23 -07:00
syscall_table_32.S [CVE-2009-0029] Rename old_readdir to sys_old_readdir 2009-01-14 14:15:15 +01:00
tce_64.c
test_nx.c
test_rodata.c
time_32.c x86: time_32.c fix style problems 2009-01-04 13:23:05 +01:00
time_64.c x86: time_64.c fix style problems 2009-01-04 13:23:05 +01:00
tlb_uv.c Merge branch 'linus' into core/percpu 2009-01-27 12:01:51 +01:00
tls.c x86: Introducing asm/syscalls.h 2008-07-22 14:35:57 +02:00
tls.h
topology.c
trampoline_32.S
trampoline_64.S
trampoline.c x86: change static allocation of trampoline area 2008-12-08 13:49:45 +01:00
traps.c x86: remove pda.h 2009-01-20 12:29:20 +09:00
tsc_sync.c Merge branches 'x86/apic', 'x86/cleanups', 'x86/cpufeature', 'x86/crashdump', 'x86/debug', 'x86/defconfig', 'x86/detect-hyper', 'x86/doc', 'x86/dumpstack', 'x86/early-printk', 'x86/fpu', 'x86/idle', 'x86/io', 'x86/memory-corruption-check', 'x86/microcode', 'x86/mm', 'x86/mtrr', 'x86/nmi-watchdog', 'x86/pat2', 'x86/pci-ioapic-boot-irq-quirks', 'x86/ptrace', 'x86/quirks', 'x86/reboot', 'x86/setup-memory', 'x86/signal', 'x86/sparse-fixes', 'x86/time', 'x86/uv' and 'x86/xen' into x86/core 2008-12-23 16:27:23 +01:00
tsc.c Merge branches 'x86/apic', 'x86/cleanups', 'x86/cpufeature', 'x86/crashdump', 'x86/debug', 'x86/defconfig', 'x86/detect-hyper', 'x86/doc', 'x86/dumpstack', 'x86/early-printk', 'x86/fpu', 'x86/idle', 'x86/io', 'x86/memory-corruption-check', 'x86/microcode', 'x86/mm', 'x86/mtrr', 'x86/nmi-watchdog', 'x86/pat2', 'x86/pci-ioapic-boot-irq-quirks', 'x86/ptrace', 'x86/quirks', 'x86/reboot', 'x86/setup-memory', 'x86/signal', 'x86/sparse-fixes', 'x86/time', 'x86/uv' and 'x86/xen' into x86/core 2008-12-23 16:27:23 +01:00
uv_irq.c x86, UV: add uv_setup_irq() and uv_teardown_irq() functions, v3, fix 2008-10-16 16:53:13 +02:00
uv_sysfs.c x86: Add sysfs entries for UV v4 2008-10-16 16:53:13 +02:00
verify_cpu_64.S
visws_quirks.c x86: rename all fields of mpc_cpu mpc_X to X 2009-01-05 14:08:33 +01:00
vm86_32.c x86: Introducing asm/syscalls.h 2008-07-22 14:35:57 +02:00
vmi_32.c x86: fix section mismatch warning 2009-01-26 14:27:18 +01:00
vmiclock_32.c cpumask: convert struct clock_event_device to cpumask pointers. 2008-12-13 21:20:26 +10:30
vmlinux_32.lds.S x86: make vmlinux_32.lds.S use PERCPU() macro 2009-01-16 14:19:09 +01:00
vmlinux_64.lds.S x86: move stack_canary into irq_stack 2009-01-20 12:29:20 +09:00
vmlinux.lds.S
vsmp_64.c x86/paravirt: add register-saving thunks to reduce caller register pressure 2009-01-30 14:51:45 -08:00
vsyscall_64.c Merge branch 'tracing-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip 2008-12-28 12:21:10 -08:00
x8664_ksyms_64.c x86: convert pda ops to wrappers around x86 percpu accessors 2009-01-16 14:20:22 +01:00
xsave.c x86: xsave.c: restore_user_xstate should be static 2008-12-30 13:31:41 -08:00