x86: Use MOVL when reading segment registers
Subsystems (3)
PATCH v2 1/3 x86: Use MOVL when reading segment registers
Use MOVL when reading segment registers to avoid 0x66 operand-size override insn prefix. The segment value is always 16-bit and gets zero-extended to the full 32-bit size.
--- arch/x86/include/asm/segment.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/include/asm/segment.h +1 -1
@@ -348,7 +348,7 @@ static inline void __loadsegment_fs(unsigned short value)* Save a segment register away: */ #define savesegment (seg, value) \ - asm("mov %%" #seg ",%0":"=r" (value) : : "memory") + asm("movl %%" #seg ",%k0" : "=r" (value) : : "memory")#endif /* !__ASSEMBLER__ */ #endif /* __KERNEL__ */
Signature · 1 line
2.51.1
PATCH v2 2/3 x86/hyperv: Use savesegment() instead of inline asm() to save segment registers
Use standard savesegment() utility macro to save segment registers.
--- arch/x86/hyperv/ivm.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)
arch/x86/hyperv/ivm.c +5 -4
@@ -25,6 +25,7 @@#include <asm/e820/api.h> #include <asm/desc.h> #include <asm/msr.h> +#include <asm/segment.h> #include <uapi/asm/vmx.h> #ifdef CONFIG_AMD_MEM_ENCRYPT
@@ -315,16 +316,16 @@ int hv_snp_boot_ap(u32 apic_id, unsigned long start_ip, unsigned int cpu) vmsa->gdtr.base = gdtr.address; vmsa->gdtr.limit = gdtr.size; - asmvolatile ("movl %%es, %%eax;" :"=a" (vmsa->es.selector)); +savesegment (es, vmsa->es.selector);hv_populate_vmcb_seg (vmsa->es, vmsa->gdtr.base); - asmvolatile ("movl %%cs, %%eax;" :"=a" (vmsa->cs.selector)); +savesegment (cs, vmsa->cs.selector);hv_populate_vmcb_seg (vmsa->cs, vmsa->gdtr.base); - asmvolatile ("movl %%ss, %%eax;" :"=a" (vmsa->ss.selector)); +savesegment (ss, vmsa->ss.selector);hv_populate_vmcb_seg (vmsa->ss, vmsa->gdtr.base); - asmvolatile ("movl %%ds, %%eax;" :"=a" (vmsa->ds.selector)); +savesegment (ds, vmsa->ds.selector);hv_populate_vmcb_seg (vmsa->ds, vmsa->gdtr.base); vmsa->efer =native_read_msr (MSR_EFER );
Signature · 1 line
2.51.1
PATCH v2 3/3 x86/hyperv: Remove ASM_CALL_CONSTRAINT with VMMCALL insn
In reply to Uros BizjakUnlike CALL instruction, VMMCALL does not push to the stack, so it's OK to allow the compiler to insert it before the frame pointer gets set up by the containing function. ASM_CALL_CONSTRAINT is for CALLs that must be inserted after the frame pointer is set up, so it is over-constraining here and can be removed.
---
v2: Expand commit message and include ASM_CALL_CONSTRAINT explanation
--- arch/x86/hyperv/ivm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/hyperv/ivm.c +1 -1
@@ -392,7 +392,7 @@ u64 hv_snp_hypercall(u64 control, u64 param1, u64 param2) register u64 __r8 asm("r8") = param2; asmvolatile ("vmmcall" - : "=a" (hv_status), ASM_CALL_CONSTRAINT, + :"=a" (hv_status),"+c" (control),"+d" (param1),"+r" (__r8) : :"cc" ,"memory" ,"r9" ,"r10" ,"r11" );
Signature · 1 line
2.51.1
H. Peter Anvin
reviewerRe: [PATCH v2 3/3] x86/hyperv: Remove ASM_CALL_CONSTRAINT with VMMCALL insn
Quoted 28 lines · Uros Bizjak
Unlike CALL instruction, VMMCALL does not push to the stack, so it's OK to allow the compiler to insert it before the frame pointer gets set up by the containing function.
ASM_CALL_CONSTRAINTis for CALLs that must be inserted after the frame pointer is set up, so it is over-constraining here and can be removed.Signed-off-by Uros Bizjak <[email protected]>Tested-by Michael Kelley <[email protected]>Cc "K. Y. Srinivasan" <[email protected]>Cc Haiyang Zhang <[email protected]>Cc Wei Liu <[email protected]>Cc Dexuan Cui <[email protected]>Cc Thomas Gleixner <[email protected]>Cc Ingo Molnar <[email protected]>Cc Borislav Petkov <[email protected]>Cc Dave Hansen <[email protected]>Cc "H. Peter Anvin" <[email protected]>---
v2: Expand commit message and includeASM_CALL_CONSTRAINTexplanation
---arch/x86/hyperv/ivm.c| 2 +-1 file changed, 1 insertion(+), 1 deletion(-)arch/x86/hyperv/ivm.c +0
-0
arch/x86/hyperv/ivm.c u64 hv_snp_hypercall(u64 control, u64 param1, u64 param2) · line 398
Quoted 8 lines
arch/x86/hyperv/ivm.c +1
-1@@ -392,7 +392,7 @@ u64 hv_snp_hypercall(u64 control, u64 param1, u64 param2) register u64 __r8 asm("r8") = param2; asmvolatile ("vmmcall" - : "=a" (hv_status), ASM_CALL_CONSTRAINT, + :"=a" (hv_status),"+c" (control),"+d" (param1),"+r" (__r8) : :"cc" ,"memory" ,"r9" ,"r10" ,"r11" );
I think it would be good to have a comment at the point where ASM_CALL_CONSTRAINT is defined explaining its proper use.
Specifically, instructions like syscall, vmcall, vmfunc, vmmcall, int xx and VM-specific escape instructions are not "calls" because they either don't modify the stack or create an exception frame (kernel) or signal frame (user space) which is completely special.
Re: [PATCH v2 3/3] x86/hyperv: Remove ASM_CALL_CONSTRAINT with VMMCALL insn
Quoted 41 lines · H. Peter Anvin
Quoted 28 lines · Uros Bizjak
Unlike CALL instruction, VMMCALL does not push to the stack, so it's OK to allow the compiler to insert it before the frame pointer gets set up by the containing function.
ASM_CALL_CONSTRAINTis for CALLs that must be inserted after the frame pointer is set up, so it is over-constraining here and can be removed.Signed-off-by Uros Bizjak <[email protected]>Tested-by Michael Kelley <[email protected]>Cc "K. Y. Srinivasan" <[email protected]>Cc Haiyang Zhang <[email protected]>Cc Wei Liu <[email protected]>Cc Dexuan Cui <[email protected]>Cc Thomas Gleixner <[email protected]>Cc Ingo Molnar <[email protected]>Cc Borislav Petkov <[email protected]>Cc Dave Hansen <[email protected]>Cc "H. Peter Anvin" <[email protected]>---
v2: Expand commit message and includeASM_CALL_CONSTRAINTexplanation
---arch/x86/hyperv/ivm.c| 2 +-1 file changed, 1 insertion(+), 1 deletion(-)arch/x86/hyperv/ivm.c +0
-0arch/x86/hyperv/ivm.c u64 hv_snp_hypercall(u64 control, u64 param1, u64 param2) · line 398
Quoted 8 lines
arch/x86/hyperv/ivm.c +1
-1@@ -392,7 +392,7 @@ u64 hv_snp_hypercall(u64 control, u64 param1, u64 param2) register u64 __r8 asm("r8") = param2; asmvolatile ("vmmcall" - : "=a" (hv_status), ASM_CALL_CONSTRAINT, + :"=a" (hv_status),"+c" (control),"+d" (param1),"+r" (__r8) : :"cc" ,"memory" ,"r9" ,"r10" ,"r11" );I think it would be good to have a comment at the point where
ASM_CALL_CONSTRAINTis defined explaining its proper use.
Quoted 1 line
Specifically, instructions like syscall, vmcall, vmfunc, vmmcall, int xx and VM-specific escape instructions are not "calls" because they either don't modify the stack or create an exception frame (kernel) or signal frame (user space) which is completely special.
The existing comment already mentions CALL instruction only:
/*
* This output constraint should be used for any inline asm which has a "call"
* instruction. Otherwise the asm may be inserted before the frame pointer
* gets set up by the containing function. If you forget to do this, objtool
* may print a "call without frame pointer save/setup" warning. */
Uros.
RE: [PATCH v2 2/3] x86/hyperv: Use savesegment() instead of inline asm() to save segment registers
In reply to Uros BizjakFrom: Uros Bizjak <[email protected]> Sent: Friday, November 21, 2025 6:14 AM
Quoted 2 lines
Use standard
savesegment()utility macro to save segment registers.
Patch 1 of this series was included in the tip tree. But this patch (Patch 2) and Patch 3 have not been picked up anywhere.
Wei Liu -- could you pick these two up in the hyperv tree?
Michael
Quoted 52 lines
Signed-off-by Uros Bizjak <[email protected]>Acked-by Wei Liu <[email protected]>Tested-by Michael Kelley <[email protected]>Cc "K. Y. Srinivasan" <[email protected]>Cc Haiyang Zhang <[email protected]>Cc Dexuan Cui <[email protected]>Cc Thomas Gleixner <[email protected]>Cc Ingo Molnar <[email protected]>Cc Borislav Petkov <[email protected]>Cc Dave Hansen <[email protected]>Cc "H. Peter Anvin" <[email protected]>--- arch/x86/hyperv/ivm.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)arch/x86/hyperv/ivm.c +5
-4@@ -25,6 +25,7 @@#include <asm/e820/api.h> #include <asm/desc.h> #include <asm/msr.h> +#include <asm/segment.h> #include <uapi/asm/vmx.h> #ifdef CONFIG_AMD_MEM_ENCRYPT @@ -315,16 +316,16 @@ int hv_snp_boot_ap(u32 apic_id, unsigned long start_ip, unsigned int cpu) vmsa->gdtr.base = gdtr.address; vmsa->gdtr.limit = gdtr.size; - asmvolatile ("movl %%es, %%eax;" :"=a" (vmsa->es.selector)); +savesegment (es, vmsa->es.selector);hv_populate_vmcb_seg (vmsa->es, vmsa->gdtr.base); - asmvolatile ("movl %%cs, %%eax;" :"=a" (vmsa->cs.selector)); +savesegment (cs, vmsa->cs.selector);hv_populate_vmcb_seg (vmsa->cs, vmsa->gdtr.base); - asmvolatile ("movl %%ss, %%eax;" :"=a" (vmsa->ss.selector)); +savesegment (ss, vmsa->ss.selector);hv_populate_vmcb_seg (vmsa->ss, vmsa->gdtr.base); - asmvolatile ("movl %%ds, %%eax;" :"=a" (vmsa->ds.selector)); +savesegment (ds, vmsa->ds.selector);hv_populate_vmcb_seg (vmsa->ds, vmsa->gdtr.base);
vmsa->efer=native_read_msr(MSR_EFER);
--
2.51.1
H. Peter Anvin
reviewerRe: [PATCH v2 3/3] x86/hyperv: Remove ASM_CALL_CONSTRAINT with VMMCALL insn
In reply to Uros BizjakQuoted 5 lines · Uros Bizjak
Quoted 4 lines
I think it would be good to have a comment at the point where
ASM_CALL_CONSTRAINTis defined explaining its proper use.Specifically, instructions like syscall, vmcall, vmfunc, vmmcall, int xx and VM-specific escape instructions are not "calls" because they either don't modify the stack or create an exception frame (kernel) or signal frame (user space) which is completely special.
Quoted 9 lines
The existing comment already mentions CALL instruction only:
/*
* This output constraint should be used for any inline asm which has a "call"
* instruction. Otherwise the asm may be inserted before the frame pointer
* gets set up by the containing function. If you forget to do this, objtool
* may print a "call without frame pointer save/setup" warning. */
Yes. Some people seem to have misunderstood it to mean any instruction with "CALL" in the name.
-hpa
Re: [PATCH v2 2/3] x86/hyperv: Use savesegment() instead of inline asm() to save segment registers
In reply to Michael KelleyQuoted 8 lines · Michael Kelley
From: Uros Bizjak <[email protected]> Sent: Friday, November 21, 2025 6:14 AM
Quoted 2 lines
Use standard
savesegment()utility macro to save segment registers.Patch 1 of this series was included in the tip tree. But this patch (Patch 2) and Patch 3 have not been picked up anywhere.
Wei Liu -- could you pick these two up in the hyperv tree?
Applied patch 2 and 3. Thank you for the reminder.
Wei