A might_sleep() check caught sys_ioperm() calling kmalloc() whilst
atomic.
The issue is we call get_cpu() prior to kmalloc(). The fix is simply to
move the get_cpu() further south, which has the added bonus of shrinking
the non-preemptibility region by a few instructions. Credit to Nick
Piggin.
I also added some comments.
Patch is against current BK, please apply.
Robert Love
diff -urN linux-2.5.40/arch/i386/kernel/ioport.c linux/arch/i386/kernel/ioport.c
--- linux-2.5.40/arch/i386/kernel/ioport.c Tue Oct 1 03:06:59 2002
+++ linux/arch/i386/kernel/ioport.c Wed Oct 2 15:01:53 2002
@@ -49,8 +49,11 @@
}
}
-/*
- * this changes the io permissions bitmap in the current task.
+/**
+ * sys_ioperm - sets the I/O permission bits for the current process
+ * @from: starting port address
+ * @num: number of bytes from starting address
+ * @turn_on: value to change the permission bits to
*/
asmlinkage int sys_ioperm(unsigned long from, unsigned long num, int turn_on)
{
@@ -63,16 +66,14 @@
if (turn_on && !capable(CAP_SYS_RAWIO))
return -EPERM;
- tss = init_tss + get_cpu();
-
/*
* If it's the first ioperm() call in this thread's lifetime, set the
* IO bitmap up. ioperm() is much less timing critical than clone(),
* this is why we delay this operation until now:
*/
if (!t->ts_io_bitmap) {
- unsigned long *bitmap;
- bitmap = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL);
+ unsigned long *bitmap = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL);
+
if (!bitmap) {
ret = -ENOMEM;
goto out;
@@ -83,11 +84,14 @@
*/
memset(bitmap, 0xff, IO_BITMAP_BYTES);
t->ts_io_bitmap = bitmap;
+
/*
* this activates it in the TSS
*/
+ tss = init_tss + get_cpu();
tss->bitmap = IO_BITMAP_OFFSET;
- }
+ } else
+ tss = init_tss + get_cpu();
/*
* do it in the per-thread copy and in the TSS ...
@@ -95,8 +99,9 @@
set_bitmap(t->ts_io_bitmap, from, num, !turn_on);
set_bitmap(tss->io_bitmap, from, num, !turn_on);
-out:
put_cpu();
+
+out:
return ret;
}