Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760377AbYGHX2W (ORCPT ); Tue, 8 Jul 2008 19:28:22 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1757913AbYGHXPM (ORCPT ); Tue, 8 Jul 2008 19:15:12 -0400 Received: from gw.goop.org ([64.81.55.164]:51220 "EHLO mail.goop.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757512AbYGHXOr (ORCPT ); Tue, 8 Jul 2008 19:14:47 -0400 Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Subject: [PATCH 52 of 55] xen64: set up syscall and sysenter entrypoints for 64-bit X-Mercurial-Node: cbecd5041ab1ad7111f75189c54fb13d8216c16a Message-Id: In-Reply-To: Date: Tue, 08 Jul 2008 15:07:14 -0700 From: Jeremy Fitzhardinge To: Ingo Molnar Cc: LKML , x86@kernel.org, Stephen Tweedie , Eduardo Habkost , Mark McLoughlin , x86@kernel.org Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 7054 Lines: 285 We set up entrypoints for syscall and sysenter. sysenter is only used for 32-bit compat processes, whereas syscall can be used in by both 32 and 64-bit processes. Signed-off-by: Jeremy Fitzhardinge --- arch/x86/xen/enlighten.c | 4 + arch/x86/xen/setup.c | 42 +++++++++++++- arch/x86/xen/smp.c | 1 arch/x86/xen/xen-asm_64.S | 131 +++++++++++++++++++++++++++++++++++++++++++-- arch/x86/xen/xen-ops.h | 3 + 5 files changed, 175 insertions(+), 6 deletions(-) diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -1139,6 +1139,10 @@ .iret = xen_iret, .irq_enable_sysexit = xen_sysexit, +#ifdef CONFIG_X86_64 + .usergs_sysret32 = xen_sysret32, + .usergs_sysret64 = xen_sysret64, +#endif .load_tr_desc = paravirt_nop, .set_ldt = xen_set_ldt, diff --git a/arch/x86/xen/setup.c b/arch/x86/xen/setup.c --- a/arch/x86/xen/setup.c +++ b/arch/x86/xen/setup.c @@ -86,9 +86,11 @@ */ static void __init fiddle_vdso(void) { +#if defined(CONFIG_X86_32) || defined(CONFIG_IA32_EMULATION) extern const char vdso32_default_start; u32 *mask = VDSO32_SYMBOL(&vdso32_default_start, NOTE_MASK); *mask |= 1 << VDSO_NOTE_NONEGSEG_BIT; +#endif } static __cpuinit int register_callback(unsigned type, const void *func) @@ -106,13 +108,46 @@ { int cpu = smp_processor_id(); extern void xen_sysenter_target(void); + int ret; - if (!boot_cpu_has(X86_FEATURE_SEP) || - register_callback(CALLBACKTYPE_sysenter, - xen_sysenter_target) != 0) { +#ifdef CONFIG_X86_32 + if (!boot_cpu_has(X86_FEATURE_SEP)) { + return; + } +#else + if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL && + boot_cpu_data.x86_vendor != X86_VENDOR_CENTAUR) { + return; + } +#endif + + ret = register_callback(CALLBACKTYPE_sysenter, xen_sysenter_target); + if(ret != 0) { clear_cpu_cap(&cpu_data(cpu), X86_FEATURE_SEP); clear_cpu_cap(&boot_cpu_data, X86_FEATURE_SEP); } +} + +void __cpuinit xen_enable_syscall(void) +{ +#ifdef CONFIG_X86_64 + int cpu = smp_processor_id(); + int ret; + extern void xen_syscall_target(void); + extern void xen_syscall32_target(void); + + ret = register_callback(CALLBACKTYPE_syscall, xen_syscall_target); + if (ret != 0) { + printk("failed to set syscall: %d\n", ret); + clear_cpu_cap(&cpu_data(cpu), X86_FEATURE_SYSCALL); + clear_cpu_cap(&boot_cpu_data, X86_FEATURE_SYSCALL); + } else { + ret = register_callback(CALLBACKTYPE_syscall32, + xen_syscall32_target); + if (ret != 0) + printk("failed to set 32-bit syscall: %d\n", ret); + } +#endif /* CONFIG_X86_64 */ } void __init xen_arch_setup(void) @@ -131,6 +166,7 @@ BUG(); xen_enable_sysenter(); + xen_enable_syscall(); set_iopl.iopl = 1; rc = HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl); diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -69,6 +69,7 @@ preempt_disable(); xen_enable_sysenter(); + xen_enable_syscall(); cpu = smp_processor_id(); smp_store_cpu_info(cpu); diff --git a/arch/x86/xen/xen-asm_64.S b/arch/x86/xen/xen-asm_64.S --- a/arch/x86/xen/xen-asm_64.S +++ b/arch/x86/xen/xen-asm_64.S @@ -15,6 +15,8 @@ #include #include +#include +#include #include @@ -138,9 +140,132 @@ mov 8+8(%rsp),%r11 ret $16 +hypercall_iret = hypercall_page + __HYPERVISOR_iret * 32 +/* + Xen64 iret frame: + + ss + rsp + rflags + cs + rip <-- standard iret frame + + flags + + rcx } + r11 }<-- pushed by hypercall page +rsp -> rax } + */ ENTRY(xen_iret) pushq $0 - jmp hypercall_page + __HYPERVISOR_iret * 32 +1: jmp hypercall_iret +ENDPATCH(xen_iret) +RELOC(xen_iret, 1b+1) +/* + sysexit is not used for 64-bit processes, so it's + only ever used to return to 32-bit compat userspace. + */ ENTRY(xen_sysexit) - ud2a + pushq $__USER32_DS + pushq %rcx + pushq $X86_EFLAGS_IF + pushq $__USER32_CS + pushq %rdx + + pushq $VGCF_in_syscall +1: jmp hypercall_iret +ENDPATCH(xen_sysexit) +RELOC(xen_sysexit, 1b+1) + +ENTRY(xen_sysret64) + /* We're already on the usermode stack at this point, but still + with the kernel gs, so we can easily switch back */ + movq %rsp, %gs:pda_oldrsp + movq %gs:pda_kernelstack,%rsp + + pushq $__USER_DS + pushq %gs:pda_oldrsp + pushq %r11 + pushq $__USER_CS + pushq %rcx + + pushq $VGCF_in_syscall +1: jmp hypercall_iret +ENDPATCH(xen_sysret64) +RELOC(xen_sysret64, 1b+1) + +ENTRY(xen_sysret32) + /* We're already on the usermode stack at this point, but still + with the kernel gs, so we can easily switch back */ + movq %rsp, %gs:pda_oldrsp + movq %gs:pda_kernelstack, %rsp + + pushq $__USER32_DS + pushq %gs:pda_oldrsp + pushq %r11 + pushq $__USER32_CS + pushq %rcx + + pushq $VGCF_in_syscall +1: jmp hypercall_iret +ENDPATCH(xen_sysret32) +RELOC(xen_sysret32, 1b+1) + +/* + Xen handles syscall callbacks much like ordinary exceptions, + which means we have: + - kernel gs + - kernel rsp + - an iret-like stack frame on the stack (including rcx and r11): + ss + rsp + rflags + cs + rip + r11 + rsp-> rcx + + In all the entrypoints, we undo all that to make it look + like a CPU-generated syscall/sysenter and jump to the normal + entrypoint. + */ + +.macro undo_xen_syscall + mov 0*8(%rsp),%rcx + mov 1*8(%rsp),%r11 + mov 5*8(%rsp),%rsp +.endm + +/* Normal 64-bit system call target */ +ENTRY(xen_syscall_target) + undo_xen_syscall + jmp system_call_after_swapgs +ENDPROC(xen_syscall_target) + +#ifdef CONFIG_IA32_EMULATION + +/* 32-bit compat syscall target */ +ENTRY(xen_syscall32_target) + undo_xen_syscall + jmp ia32_cstar_target +ENDPROC(xen_syscall32_target) + +/* 32-bit compat sysenter target */ +ENTRY(xen_sysenter_target) + undo_xen_syscall + jmp ia32_sysenter_target +ENDPROC(xen_sysenter_target) + +#else /* !CONFIG_IA32_EMULATION */ + +ENTRY(xen_syscall32_target) +ENTRY(xen_sysenter_target) + lea 16(%rsp), %rsp /* strip %rcx,%r11 */ + mov $-ENOSYS, %rax + pushq $VGCF_in_syscall + jmp hypercall_iret +ENDPROC(xen_syscall32_target) +ENDPROC(xen_sysenter_target) + +#endif /* CONFIG_IA32_EMULATION */ diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h --- a/arch/x86/xen/xen-ops.h +++ b/arch/x86/xen/xen-ops.h @@ -26,6 +26,7 @@ void __init xen_arch_setup(void); void __init xen_init_IRQ(void); void xen_enable_sysenter(void); +void xen_enable_syscall(void); void xen_vcpu_restore(void); void __init xen_build_dynamic_phys_to_machine(void); @@ -70,6 +71,8 @@ /* These are not functions, and cannot be called normally */ void xen_iret(void); void xen_sysexit(void); +void xen_sysret32(void); +void xen_sysret64(void); void xen_adjust_exception_frame(void); #endif /* XEN_OPS_H */ -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/