Hello.
George Anzinger wrote:
>
> > then preempt_stop (cli) can be killed in entry.S:ret_from_intr()
>
> Also at least some of the trap code returns with interrupts enabled.
Then linux already have a bug - ret_from_exception: bypasses
ret_from_intr:. But as far as i can see, it does not - all users of
ret_from_exception: do preempt_stop. And so it can be shifted to
ret_from_exception.
I beleive none of the traps need current_thread_info() in regs->ebx,
so GET_THREAD_INFO(%ebx) can be killed in error_code:,
common_interrupt:,
and BUILD_INTERRUPT().
Not sure it is right time to such minor cleanups, but...
Patch on top of remove-irqlock-2.5.27-E0:
--- arch/i386/kernel/entry.S.orig Mon Jul 22 23:44:41 2002
+++ arch/i386/kernel/entry.S Tue Jul 23 00:04:02 2002
@@ -185,8 +185,10 @@
# userspace resumption stub bypassing syscall exit tracing
ALIGN
-ret_from_intr:
ret_from_exception:
+ preempt_stop
+ret_from_intr:
+ GET_THREAD_INFO(%ebx)
movl EFLAGS(%esp), %eax # mix EFLAGS and CS
movb CS(%esp), %al
testl $(VM_MASK | 3), %eax
@@ -333,14 +335,12 @@
common_interrupt:
SAVE_ALL
call do_IRQ
- GET_THREAD_INFO(%ebx)
jmp ret_from_intr
#define BUILD_INTERRUPT(name, nr) \
ENTRY(name) \
pushl $nr-256; \
SAVE_ALL \
- GET_THREAD_INFO(%ebx); \
call smp_/**/name; \
jmp ret_from_intr;
@@ -400,10 +400,8 @@
movl $(__KERNEL_DS), %edx
movl %edx, %ds
movl %edx, %es
- GET_THREAD_INFO(%ebx)
call *%edi
addl $8, %esp
- preempt_stop
jmp ret_from_exception
ENTRY(coprocessor_error)
@@ -430,7 +428,6 @@
pushl $0 # temporary storage for ORIG_EIP
call math_emulate
addl $4, %esp
- preempt_stop
jmp ret_from_exception
ENTRY(debug)
Oleg.
-
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/
This archive was generated by hypermail 2b29 : Tue Jul 23 2002 - 22:00:40 EST