Skip to contentSkip to conversation

r8169 with 64bit DMA is broken

DISCUSSION Ismael Luceno 1 participant 1 message Mar 31, 2010 2 messages not in the archive

lore
Share & tools
lore
#2 lore raw
In reply to a message not in the archive
Quoted 2 lines · El Mon, 29 Mar 2010 20:15:47 -0600 Robert Hancock <[email protected]>…

That sounds similar to a previous report.. Can you CC the thread on linux-kernel (if you've seen it)?

Sorry, I am not subscribed.

Quoted 2 lines

Also, what does dmesg show when loading the driver? There should be a string showing the chip version, something like this:

<...>

r8169 0000:05:00.0: eth0: RTL8102e at 0xffffc900100be000, 00:26:b9:a0:33:86, XID 04e00000 IRQ 28

Then, after initialization, it shows two "link up" messages, and that's all it reports. It's a nice blackhole :P.

Quoted 25 lines
Quoted 23 lines · Ismael Luceno

On 2.6.33 it worked fine, but then I decided to upgrade to 2.6.34-rc2 and it simply didn't detected the carrier correctly anymore (it was like if the cable was always plugged...).

After a short inspection, the obvious offender was commit 353176888386d9025062a12dcec08d49af10cf2c.

I tried with use_dac=0, but it didn't work (I had no time to investigate it further), but after reverting the commit it worked again.

The system is a Dell Inspiron mini 1012, lspci reports:
05:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8101E/RTL8102E PCI Express Fast Ethernet controller (rev 02)
Subsystem: Dell Device 041a Flags: bus master, fast devsel, latency 0, IRQ 28 I/O ports at 2000 [size=256]
       Memory at f0f20000 (64-bit, prefetchable) [size=4K] Memory at f0f10000 (64-bit, prefetchable) [size=64K] [virtual] Expansion ROM at f0f40000 [disabled] [size=128K]
       Capabilities: <access denied>
       Kernel driver in use: r8169
       Kernel modules: r8169