With a tigon3 5701, 2.4.21 kernel, we are getting persistent
tg3_stop_block timed out, ofs=* enable_bit=2 messages, with ofs=3400,
2400, 1400, c00. drivers/net/tg3.c:
/* To stop a block, clear the enable bit and poll till it
* clears. tp->lock is held.
*/
static int tg3_stop_block(struct tg3 *tp, unsigned long ofs, u32 enable_bit)
{
unsigned int i;
u32 val;
val = tr32(ofs);
val &= ~enable_bit;
tw32(ofs, val);
tr32(ofs); <=== why?
for (i = 0; i < MAX_WAIT_CNT; i++) {
udelay(100);
val = tr32(ofs);
if ((val & enable_bit) == 0)
break;
}
if (i == MAX_WAIT_CNT) {
printk(KERN_ERR PFX "tg3_stop_block timed out, "
"ofs=%lx enable_bit=%x\n",
ofs, enable_bit);
return -ENODEV;
}
return 0;
}
Why does the code read tr32(ofs) twice and only test after the second
read? Is there any possibility of a race here?
tw32(ofs, val); /* clear bit1 */
tr32(ofs); /* bit1 reads 0 */
for (i = 0; i < MAX_WAIT_CNT; i++) {
udelay(100);
val = tr32(ofs); /* bit1 reset to 1 */
IOW, bit1 went 1 -> 0 -> 1 but because we only test the second and
subsequent read, it appears that bit1 is locked on.
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
