| To: | "Olaf Hering" <olh@xxxxxxx>, netdev@xxxxxxxxxxx |
|---|---|
| Subject: | RE: tg3 access hangs on system shutdown with 2.6.11-rc5, tg3_stop_block timed out |
| From: | "Michael Chan" <mchan@xxxxxxxxxxxx> |
| Date: | Mon, 28 Feb 2005 13:33:38 -0800 |
| Sender: | netdev-bounce@xxxxxxxxxxx |
| Thread-index: | AcUcHx9HcO6MYrY2SKqov26LRDl5qABuvc0Q |
| Thread-topic: | tg3 access hangs on system shutdown with 2.6.11-rc5, tg3_stop_block timed out |
> > my JS20 blades hang on system reboot with current Linus tree. > I think it started after rc3. Maybe this line in dmesg gives a hint: > > tg3: tg3_stop_block timed out, ofs=4000 enable_bit=2 > Olaf, This error message means that the memory arbiter cannot be disabled for some reason. Looking at the code, tg3_abort_hw() will return -ENODEV and skip the 2 memset calls to zero out the hw_status and hw_stats memory. I don't believe skipping the memset calls can cause a hang. After tg3_abort_hw(), the chip should then be reset to a known state. The 5704S changes in 2.6.11-rc4 are mainly related to serdes link-up and tx signal. I don't see how those changes can cause a hang or the tg3_stop_block error. The stack trace did not show any tg3 function calls. Can you narrow it down further to a specific patch that is causing the problem? Michael |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [PATCH 2/2] r8169: RTL8169_registers clean-up, Jon Mason |
|---|---|
| Next by Date: | Re: tg3 access hangs on system shutdown with 2.6.11-rc5, tg3_stop_block timed out, Olaf Hering |
| Previous by Thread: | rp_filter interaction with netfilter SNAT/un-SNAT, Ian! D. Allen |
| Next by Thread: | Re: tg3 access hangs on system shutdown with 2.6.11-rc5, tg3_stop_block timed out, Olaf Hering |
| Indexes: | [Date] [Thread] [Top] [All Lists] |