Skip to contentSkip to conversation

x86/traps: Have read_cr0() only once in the #NM handler

PATCH In review Borislav Petkov 4 participants 4 messages Jan 17, 2019 – Jan 22, 2019 x86

lore
Share & tools
lore

Borislav Petkov

author maintainer
#0 lore raw

PATCH 1/1 x86/traps: Have read_cr0() only once in the #NM handler

From: Borislav Petkov <[email protected]>

... instead of twice in the code. In any case, CR0 ends up being read once anyway:

1. The CONFIG_MATH_EMULATION case does so and exits.
2. The normal case does it once too.

However, read it on function entry instead to make the code even simpler to follow.

No functional changes.

Cc Andy Lutomirski <[email protected]>
Cc Sebastian Andrzej Siewior <[email protected]>
---
 arch/x86/kernel/traps.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
arch/x86/kernel/traps.c +2 -3
index 9b7c4ca8f0a7..2684a9d11e66 100644
@@ -880,12 +880,12 @@ do_spurious_interrupt_bug(struct pt_regs *regs, long error_code)
 dotraplinkage void
 do_device_not_available(struct pt_regs *regs, long error_code)
 {
-	unsigned long cr0;
+	unsigned long cr0 = read_cr0();
 
 	RCU_LOCKDEP_WARN(!rcu_is_watching(), "entry code didn't wake RCU");
 
 #ifdef CONFIG_MATH_EMULATION
-	if (!boot_cpu_has(X86_FEATURE_FPU) && (read_cr0() & X86_CR0_EM)) {
+	if (!boot_cpu_has(X86_FEATURE_FPU) && (cr0 & X86_CR0_EM)) {
 		struct math_emu_info info = { };
 
 		cond_local_irq_enable(regs);
@@ -897,7 +897,6 @@ do_device_not_available(struct pt_regs *regs, long error_code)
 #endif
 
 	/* This should not happen. */
-	cr0 = read_cr0();
 	if (WARN(cr0 & X86_CR0_TS, "CR0.TS was set")) {
 		/* Try to fix it up and carry on. */
 		write_cr0(cr0 & ~X86_CR0_TS);
Signature · 1 line
2.19.1
#3 lore raw
In reply to Borislav Petkov
x86/traps: Have read_cr0() only once in the #NM handler
Commit-ID:  ee35b9b9f6d52ba134b4e75442531935f295be7a
Gitweb:     https://git.kernel.org/tip/ee35b9b9f6d52ba134b4e75442531935f295be7a
Author:     Borislav Petkov <[email protected]>
AuthorDate: Thu, 17 Jan 2019 13:02:05 +0100
Committer:  Borislav Petkov <[email protected]>
CommitDate: Tue, 22 Jan 2019 14:13:35 +0100

x86/traps: Have read_cr0() only once in the #NM handler

... instead of twice in the code. In any case, CR0 ends up being read once anyway:

1. The CONFIG_MATH_EMULATION case does so and exits.
2. The normal case does it once too.

However, read it on function entry instead to make the code even simpler to follow.

No functional changes.

---
 arch/x86/kernel/traps.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
arch/x86/kernel/traps.c +2 -3
index 9b7c4ca8f0a7..2684a9d11e66 100644
@@ -880,12 +880,12 @@ do_spurious_interrupt_bug(struct pt_regs *regs, long error_code)
 dotraplinkage void
 do_device_not_available(struct pt_regs *regs, long error_code)
 {
-	unsigned long cr0;
+	unsigned long cr0 = read_cr0();
 
 	RCU_LOCKDEP_WARN(!rcu_is_watching(), "entry code didn't wake RCU");
 
 #ifdef CONFIG_MATH_EMULATION
-	if (!boot_cpu_has(X86_FEATURE_FPU) && (read_cr0() & X86_CR0_EM)) {
+	if (!boot_cpu_has(X86_FEATURE_FPU) && (cr0 & X86_CR0_EM)) {
 		struct math_emu_info info = { };
 
 		cond_local_irq_enable(regs);
@@ -897,7 +897,6 @@ do_device_not_available(struct pt_regs *regs, long error_code)
 #endif
 
 	/* This should not happen. */
-	cr0 = read_cr0();
 	if (WARN(cr0 & X86_CR0_TS, "CR0.TS was set")) {
 		/* Try to fix it up and carry on. */
 		write_cr0(cr0 & ~X86_CR0_TS);