On 6/20/06, Massimiliano Poletto <[EMAIL PROTECTED]> wrote:
Shaw Vrana asked me to resend the message below (originally posted to
lkml) to this list.
Note that I managed to make the problem go away by adding the boot
options "pci=usepirqmask acpi=noirq". I got the hint by reading dmesg
output and Documentation/kernel-parameters.txt. Still, maybe the
information below and my apparent solution will be of use to someone
more knowledgeable in this matter than I am.
thats odd that changing the kernel interrupt setup would make a difference.
How about compiling the e1000 driver with NAPI support? I think your
problem will go away. You don't have any choice but to run NAPI with
tg3, so please at least compare apples to apples by enabling NAPI and
re-running your test.
<snip>
BUG: soft lockup detected on CPU#0!
Pid: 0, comm: swapper
EIP: 0060:[<c035b3f6>] CPU: 0
EIP is at netif_rx+0x131/0x18b
EFLAGS: 00000282 Not tainted (2.6.16.19 #1)
EAX: 00000000 EBX: f2781a80 ECX: c0514680 EDX: c3fa4560
ESI: c3fa45e0 EDI: c051a000 EBP: 00000282 DS: 007b ES: 007b
CR0: 8005003b CR2: 081ed000 CR3: 04356da0 CR4: 000006f0
[<f88bd9ea>] e1000_clean_rx_irq+0x214/0x63b [e1000]
[<f88c116c>] e1000_intr+0x13f/0x47d [e1000]
spinning here would pretty much be expected and is exactly what NAPI
is designed to fix.
Jesse
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html