x86: only enable interrupts when kernel state has been set up

The sysenter path tries to enable interrupts immediately. Unfortunately this doesn't work in a paravirt environment, because not enough kernel state has been set up at that point (namely, pointing %fs to the kernel percpu data segment). To fix this, defer ENABLE_INTERRUPTS until after the kernel state has been set up. Unfortunately this means that we're running with interrupts disabled for a while without calling the IRQ tracing code, but that can't be called without setting up %fs either. Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>

x86: only enable interrupts when kernel state has been set up
The sysenter path tries to enable interrupts immediately. Unfortunately this doesn't work in a paravirt environment, because not enough kernel state has been set up at that point (namely, pointing %fs to the kernel percpu data segment). To fix this, defer ENABLE_INTERRUPTS until after the kernel state has been set up. Unfortunately this means that we're running with interrupts disabled for a while without calling the IRQ tracing code, but that can't be called without setting up %fs either. Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Jeremy Fitzhardinge · Ingo Molnar
1 parent 687c805409
Showing 1 changed file with 10 additions and 9 deletions Side-by-side Diff
arch/x86/kernel/entry_32.S
@@ -291,10 +291,10 @@
 	movl TSS_sysenter_sp0(%esp),%esp
 sysenter_past_esp:
 	/*
-	 * No need to follow this irqs on/off section: the syscall
-	 * disabled irqs and here we enable it straight after entry:
+	 * Interrupts are disabled here, but we can't trace it until
+	 * enough kernel state to call TRACE_IRQS_OFF can be called - but
+	 * we immediately enable interrupts at that point anyway.
 	 */
-	ENABLE_INTERRUPTS(CLBR_NONE)
 	pushl $(__USER_DS)
 	CFI_ADJUST_CFA_OFFSET 4
 	/*CFI_REL_OFFSET ss, 0*/
@@ -302,6 +302,7 @@
 	CFI_ADJUST_CFA_OFFSET 4
 	CFI_REL_OFFSET esp, 0
 	pushfl
+	orl $X86_EFLAGS_IF, (%esp)
 	CFI_ADJUST_CFA_OFFSET 4
 	pushl $(__USER_CS)
 	CFI_ADJUST_CFA_OFFSET 4
@@ -315,6 +316,11 @@
 	CFI_ADJUST_CFA_OFFSET 4
 	CFI_REL_OFFSET eip, 0
  
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	ENABLE_INTERRUPTS(CLBR_NONE)
+
 /*
  * Load the potential sixth argument from user stack.
  * Careful about security.
  
@@ -322,14 +328,12 @@
 	cmpl $__PAGE_OFFSET-3,%ebp
 	jae syscall_fault
 1:	movl (%ebp),%ebp
+	movl %ebp,PT_EBP(%esp)
 .section __ex_table,"a"
 	.align 4
 	.long 1b,syscall_fault
 .previous
  
-	pushl %eax
-	CFI_ADJUST_CFA_OFFSET 4
-	SAVE_ALL
 	GET_THREAD_INFO(%ebp)
  
 	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
@@ -543,9 +547,6 @@
  
 	RING0_INT_FRAME			# can't unwind into user space anyway
 syscall_fault:
-	pushl %eax			# save orig_eax
-	CFI_ADJUST_CFA_OFFSET 4
-	SAVE_ALL
 	GET_THREAD_INFO(%ebp)
 	movl $-EFAULT,PT_EAX(%esp)
 	jmp resume_userspace
...	...	@@ -291,10 +291,10 @@
291	291	movl TSS_sysenter_sp0(%esp),%esp
292	292	sysenter_past_esp:
293	293	/*
294		- * No need to follow this irqs on/off section: the syscall
295		- * disabled irqs and here we enable it straight after entry:
	294	+ * Interrupts are disabled here, but we can't trace it until
	295	+ * enough kernel state to call TRACE_IRQS_OFF can be called - but
	296	+ * we immediately enable interrupts at that point anyway.
296	297	*/
297		- ENABLE_INTERRUPTS(CLBR_NONE)
298	298	pushl $(__USER_DS)
299	299	CFI_ADJUST_CFA_OFFSET 4
300	300	/CFI_REL_OFFSET ss, 0/
...	...	@@ -302,6 +302,7 @@
302	302	CFI_ADJUST_CFA_OFFSET 4
303	303	CFI_REL_OFFSET esp, 0
304	304	pushfl
	305	+ orl $X86_EFLAGS_IF, (%esp)
305	306	CFI_ADJUST_CFA_OFFSET 4
306	307	pushl $(__USER_CS)
307	308	CFI_ADJUST_CFA_OFFSET 4
...	...	@@ -315,6 +316,11 @@
315	316	CFI_ADJUST_CFA_OFFSET 4
316	317	CFI_REL_OFFSET eip, 0
317	318
	319	+ pushl %eax
	320	+ CFI_ADJUST_CFA_OFFSET 4
	321	+ SAVE_ALL
	322	+ ENABLE_INTERRUPTS(CLBR_NONE)
	323	+
318	324	/*
319	325	* Load the potential sixth argument from user stack.
320	326	* Careful about security.
321	327
...	...	@@ -322,14 +328,12 @@
322	328	cmpl $__PAGE_OFFSET-3,%ebp
323	329	jae syscall_fault
324	330	1: movl (%ebp),%ebp
	331	+ movl %ebp,PT_EBP(%esp)
325	332	.section __ex_table,"a"
326	333	.align 4
327	334	.long 1b,syscall_fault
328	335	.previous
329	336
330		- pushl %eax
331		- CFI_ADJUST_CFA_OFFSET 4
332		- SAVE_ALL
333	337	GET_THREAD_INFO(%ebp)
334	338
335	339	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
...	...	@@ -543,9 +547,6 @@
543	547
544	548	RING0_INT_FRAME # can't unwind into user space anyway
545	549	syscall_fault:
546		- pushl %eax # save orig_eax
547		- CFI_ADJUST_CFA_OFFSET 4
548		- SAVE_ALL
549	550	GET_THREAD_INFO(%ebp)
550	551	movl $-EFAULT,PT_EAX(%esp)
551	552	jmp resume_userspace