pcie Realtek 8168G issues (re driver)

Discussion:

(too old to reply)

Luca Pizzamiglio

2015-02-13 16:07:26 UTC

Hi, I'm Luca,

I've some issues using a PCIe Realtek Ethernet board:
***@pci0:3:0:0: class=0x020000 card=0x012310ec chip=0x816810ec rev=0x0c hdr=0x00
vendor = 'Realtek Semiconductor Co., Ltd.'
device = 'RTL8111/8168B PCI Express Gigabit Ethernet controller'
class = network
subclass = ethernet
bar [10] = type I/O Port, range 32, base 0x1000, size 256, enabled
bar [18] = type Memory, range 64, base 0x90500000, size 4096, enabled
bar [20] = type Prefetchable Memory, range 64, base 0x90400000,
size 16384, enabled
cap 01[40] = powerspec 3 supports D0 D1 D2 D3 current D0
cap 05[50] = MSI supports 1 message, 64 bit
cap 10[70] = PCI-Express 2 endpoint IRQ 1 max data 128(128) link x1(x1)
speed 2.5(2.5) ASPM disabled(L0s/L1)
cap 11[b0] = MSI-X supports 4 messages
Table in map 0x20[0x0], PBA in map 0x20[0x800]
cap 03[d0] = VPD
ecap 0001[100] = AER 1 0 fatal 0 non-fatal 0 corrected
ecap 0002[140] = VC 1 max VC0
ecap 0003[160] = Serial 1 01000000684ce000
ecap 0018[170] = LTR 1

Rx and Tx don't work. After some minutes the interface is activated I
get kernel panic.
I've already tried to disable MSIx and MSI.
It seems a DMA problem, rx fill the 256 descriptors and the nothing
else until the panic. netstat -s shows now new packets.

I filled a bug report with more infos:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=197535

could someone kindly pointing some ideas?

Best regards,
Luca

Ben Perrault

2015-02-13 18:35:04 UTC

Permalink

Luca,

I've had the same issue with this interface on both PCIe boards and embedded in a handful of Lenovo products. The one, fairly ugly workaround I've found that makes it work well enough is disable tso ( i.e. ifconfig re0 down && ifconfig re0 -tso && ifconfig re0 up ). This also seems to stop the panics under current.

I'm not sure it will work for you - but it has on everyone of those interfaces I've dealt with.

Good luck,
-bp

Post by Luca Pizzamiglio
Hi, I'm Luca,
vendor = 'Realtek Semiconductor Co., Ltd.'
device = 'RTL8111/8168B PCI Express Gigabit Ethernet controller'
class = network
subclass = ethernet
bar [10] = type I/O Port, range 32, base 0x1000, size 256, enabled
bar [18] = type Memory, range 64, base 0x90500000, size 4096, enabled
bar [20] = type Prefetchable Memory, range 64, base 0x90400000,
size 16384, enabled
cap 01[40] = powerspec 3 supports D0 D1 D2 D3 current D0
cap 05[50] = MSI supports 1 message, 64 bit
cap 10[70] = PCI-Express 2 endpoint IRQ 1 max data 128(128) link x1(x1)
speed 2.5(2.5) ASPM disabled(L0s/L1)
cap 11[b0] = MSI-X supports 4 messages
Table in map 0x20[0x0], PBA in map 0x20[0x800]
cap 03[d0] = VPD
ecap 0001[100] = AER 1 0 fatal 0 non-fatal 0 corrected
ecap 0002[140] = VC 1 max VC0
ecap 0003[160] = Serial 1 01000000684ce000
ecap 0018[170] = LTR 1
Rx and Tx don't work. After some minutes the interface is activated I
get kernel panic.
I've already tried to disable MSIx and MSI.
It seems a DMA problem, rx fill the 256 descriptors and the nothing
else until the panic. netstat -s shows now new packets.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=197535
could someone kindly pointing some ideas?
Best regards,
Luca
_______________________________________________
http://lists.freebsd.org/mailman/listinfo/freebsd-current

Luca Pizzamiglio

2015-02-17 17:32:54 UTC

Permalink

Hi Ben,
thanks for the tip! tso was already disabled.
I tried anyway and unfortunately it crashes as before.

I filled a bug report
(https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=197535) and marius@
is giving me a big help on it.

Best regards,
Luca

Post by Ben Perrault
Luca,
I've had the same issue with this interface on both PCIe boards and embedded in a handful of Lenovo products. The one, fairly ugly workaround I've found that makes it work well enough is disable tso ( i.e. ifconfig re0 down && ifconfig re0 -tso && ifconfig re0 up ). This also seems to stop the panics under current.
I'm not sure it will work for you - but it has on everyone of those interfaces I've dealt with.
Good luck,
-bp

O. Hartmann

2015-02-17 18:31:56 UTC

Permalink

Am Tue, 17 Feb 2015 18:32:22 +0100

Post by Luca Pizzamiglio
Hi Ben,
thanks for the tip! tso was already disabled.
I tried anyway and unfortunately it crashes as before.
I filled a bug report
is giving me a big help on it.
Best regards,
Luca

Post by Ben Perrault
Luca,
I've had the same issue with this interface on both PCIe boards and embedded in a
handful of Lenovo products. The one, fairly ugly workaround I've found that makes it
work well enough is disable tso ( i.e. ifconfig re0 down && ifconfig re0 -tso &&
ifconfig re0 up ). This also seems to stop the panics under current.
I'm not sure it will work for you - but it has on everyone of those interfaces I've
dealt with.
Good luck,
-bp

_______________________________________________
http://lists.freebsd.org/mailman/listinfo/freebsd-current

In September 2014 I filed allready a bug acoording to strange behaviour with a Lenovo
ThinkPad E540 with a Realtek chip:

Bug 193743 - RTL8111/8168B PCI Express Gigabit Ethernet
controller: doesn't work properly, problems getting UP automatically

Luca Pizzamiglio

2015-02-25 15:00:34 UTC

Permalink

Hi,
thanks you all for the replies.

Unfortunately, the network chip is still not working and I updated the
PR (https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=197535) with the
last tests.
It seems that received packets are not transferred to mbuf or they are
transferred, but later, after the mbuf is already freed; moreover, the
ring entries are written without looping, overwriting and messing up
the whole kernel memory. It looks like a DMA issues, but

Apparently it seems a hardware error, but using a Linux distro, it works :(

Has someone maybe any other ideas? In the meanwhile I'll get another
board with the same chip :O

Best regards,
Luca

On Tue, Feb 17, 2015 at 7:31 PM, O. Hartmann

Post by O. Hartmann
Am Tue, 17 Feb 2015 18:32:22 +0100

Post by Ben Perrault
Luca,
I've had the same issue with this interface on both PCIe boards and embedded in a
handful of Lenovo products. The one, fairly ugly workaround I've found that makes it
work well enough is disable tso ( i.e. ifconfig re0 down && ifconfig re0 -tso &&
ifconfig re0 up ). This also seems to stop the panics under current.
I'm not sure it will work for you - but it has on everyone of those interfaces I've
dealt with.
Good luck,
-bp

_______________________________________________
http://lists.freebsd.org/mailman/listinfo/freebsd-current

In September 2014 I filed allready a bug acoording to strange behaviour with a Lenovo
Bug 193743 - RTL8111/8168B PCI Express Gigabit Ethernet
controller: doesn't work properly, problems getting UP automatically