* Xilinx network trouble
@ 2008-04-17 10:04 Guillaume Dargaud
2008-04-17 15:56 ` Rick Moleres
0 siblings, 1 reply; 6+ messages in thread
From: Guillaume Dargaud @ 2008-04-17 10:04 UTC (permalink / raw)
To: linuxppc-dev
Our network guru is out on vacation, so I'll ask here...
I now have a working Xilinx ML405 board... except the network is flaky to
say the least.
I have tried both with the Emac Lite and the LLTemac IPs.
They both work during the DHCP state of the kernel boot, my target board
gets its IP and network parameters alright from the host DHCP server.
The Emac Lite can ping both ways, but with 500ms and 30% packet loss when
pinging from the host PC and 300ms and 70% packet loss the other way. And it
goes in hiccups: 30s or so with no pings, and then 30s or so with 90%
success... RX/TX LEDs also are dark when there's nothing going on.
When using the LLTemac, ping just doesn't work.
ifconfig shows no drops and no collisions in both cases.
I'm no network specialist so I don't know what kind of tools I can use for
further diagnosis.
Here are some of the relevant kernel options I've set for both
configurations:
$ egrep "EMAC|NET|XILINX|NFS" .config | egrep -v "^#|^$"
CONFIG_XILINX_ML405=y
CONFIG_XILINX_MLxxx=y
CONFIG_XILINX_VIRTEX_4_FX=y
CONFIG_XILINX_VIRTEX=y
CONFIG_XILINX_EMBED_CONFIG=y
CONFIG_NET=y
CONFIG_INET=y
CONFIG_INET_DIAG=y
CONFIG_INET_TCP_DIAG=y
CONFIG_XILINX_SYSACE=y
CONFIG_XILINX_DRIVERS=y
CONFIG_NEED_XILINX_IPIF=y
CONFIG_NETDEVICES=y
CONFIG_NET_ETHERNET=y
CONFIG_XILINX_EMACLITE=y
CONFIG_XILINX_GPIO=y
CONFIG_XILINX_HWICAP=y
CONFIG_XILINX_IIC=y
CONFIG_XILINX_EDK=y
CONFIG_NETWORK_FILESYSTEMS=y
CONFIG_NFS_FS=y
CONFIG_NFS_V3=y
CONFIG_ROOT_NFS=y
CONFIG_NFS_COMMON=y
or:
CONFIG_XILINX_ML405=y
CONFIG_XILINX_MLxxx=y
CONFIG_XILINX_VIRTEX_4_FX=y
CONFIG_XILINX_VIRTEX=y
CONFIG_XILINX_EMBED_CONFIG=y
CONFIG_NET=y
CONFIG_INET=y
CONFIG_INET_DIAG=y
CONFIG_INET_TCP_DIAG=y
CONFIG_XILINX_SYSACE=y
CONFIG_XILINX_DRIVERS=y
CONFIG_NEED_XILINX_LLDMA=y
CONFIG_NETDEVICES=y
CONFIG_NETDEV_1000=y
CONFIG_XILINX_LLTEMAC=y
CONFIG_XILINX_LLTEMAC_MARVELL_88E1111_GMII=y
CONFIG_XILINX_GPIO=y
CONFIG_XILINX_HWICAP=y
CONFIG_XILINX_IIC=y
CONFIG_XILINX_EDK=y
CONFIG_NETWORK_FILESYSTEMS=y
CONFIG_NFS_FS=y
CONFIG_NFS_V3=y
CONFIG_ROOT_NFS=y
CONFIG_NFS_COMMON=y
Thanks for any hint.
--
Guillaume Dargaud
http://www.gdargaud.net/
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: Xilinx network trouble
2008-04-17 10:04 Xilinx network trouble Guillaume Dargaud
@ 2008-04-17 15:56 ` Rick Moleres
2008-04-18 14:40 ` Xilinx network trouble (mostly solved) Guillaume Dargaud
0 siblings, 1 reply; 6+ messages in thread
From: Rick Moleres @ 2008-04-17 15:56 UTC (permalink / raw)
To: Guillaume Dargaud, linuxppc-dev
Which version of xps_ll_temac are you using, and which version of EDK?
Also, do both designs meet timing?
-Rick
-----Original Message-----
From: linuxppc-dev-bounces+moleres=3Dxilinx.com@ozlabs.org
[mailto:linuxppc-dev-bounces+moleres=3Dxilinx.com@ozlabs.org] On Behalf =
Of
Guillaume Dargaud
Sent: Thursday, April 17, 2008 4:04 AM
To: linuxppc-dev@ozlabs.org
Subject: Xilinx network trouble
Our network guru is out on vacation, so I'll ask here...
I now have a working Xilinx ML405 board... except the network is flaky
to=20
say the least.
I have tried both with the Emac Lite and the LLTemac IPs.
They both work during the DHCP state of the kernel boot, my target board
gets its IP and network parameters alright from the host DHCP server.
The Emac Lite can ping both ways, but with 500ms and 30% packet loss
when=20
pinging from the host PC and 300ms and 70% packet loss the other way.
And it=20
goes in hiccups: 30s or so with no pings, and then 30s or so with 90%=20
success... RX/TX LEDs also are dark when there's nothing going on.
When using the LLTemac, ping just doesn't work.
ifconfig shows no drops and no collisions in both cases.
I'm no network specialist so I don't know what kind of tools I can use
for=20
further diagnosis.
Here are some of the relevant kernel options I've set for both=20
configurations:
$ egrep "EMAC|NET|XILINX|NFS" .config | egrep -v "^#|^$"
CONFIG_XILINX_ML405=3Dy
CONFIG_XILINX_MLxxx=3Dy
CONFIG_XILINX_VIRTEX_4_FX=3Dy
CONFIG_XILINX_VIRTEX=3Dy
CONFIG_XILINX_EMBED_CONFIG=3Dy
CONFIG_NET=3Dy
CONFIG_INET=3Dy
CONFIG_INET_DIAG=3Dy
CONFIG_INET_TCP_DIAG=3Dy
CONFIG_XILINX_SYSACE=3Dy
CONFIG_XILINX_DRIVERS=3Dy
CONFIG_NEED_XILINX_IPIF=3Dy
CONFIG_NETDEVICES=3Dy
CONFIG_NET_ETHERNET=3Dy
CONFIG_XILINX_EMACLITE=3Dy
CONFIG_XILINX_GPIO=3Dy
CONFIG_XILINX_HWICAP=3Dy
CONFIG_XILINX_IIC=3Dy
CONFIG_XILINX_EDK=3Dy
CONFIG_NETWORK_FILESYSTEMS=3Dy
CONFIG_NFS_FS=3Dy
CONFIG_NFS_V3=3Dy
CONFIG_ROOT_NFS=3Dy
CONFIG_NFS_COMMON=3Dy
or:
CONFIG_XILINX_ML405=3Dy
CONFIG_XILINX_MLxxx=3Dy
CONFIG_XILINX_VIRTEX_4_FX=3Dy
CONFIG_XILINX_VIRTEX=3Dy
CONFIG_XILINX_EMBED_CONFIG=3Dy
CONFIG_NET=3Dy
CONFIG_INET=3Dy
CONFIG_INET_DIAG=3Dy
CONFIG_INET_TCP_DIAG=3Dy
CONFIG_XILINX_SYSACE=3Dy
CONFIG_XILINX_DRIVERS=3Dy
CONFIG_NEED_XILINX_LLDMA=3Dy
CONFIG_NETDEVICES=3Dy
CONFIG_NETDEV_1000=3Dy
CONFIG_XILINX_LLTEMAC=3Dy
CONFIG_XILINX_LLTEMAC_MARVELL_88E1111_GMII=3Dy
CONFIG_XILINX_GPIO=3Dy
CONFIG_XILINX_HWICAP=3Dy
CONFIG_XILINX_IIC=3Dy
CONFIG_XILINX_EDK=3Dy
CONFIG_NETWORK_FILESYSTEMS=3Dy
CONFIG_NFS_FS=3Dy
CONFIG_NFS_V3=3Dy
CONFIG_ROOT_NFS=3Dy
CONFIG_NFS_COMMON=3Dy
Thanks for any hint.
--=20
Guillaume Dargaud
http://www.gdargaud.net/
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-dev
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Xilinx network trouble (mostly solved)
2008-04-17 15:56 ` Rick Moleres
@ 2008-04-18 14:40 ` Guillaume Dargaud
2008-04-18 16:14 ` Rick Moleres
0 siblings, 1 reply; 6+ messages in thread
From: Guillaume Dargaud @ 2008-04-18 14:40 UTC (permalink / raw)
To: linuxppc-dev
Hello Rick,
> Which version of xps_ll_temac are you using, and which version of EDK?
Using xps 9.2.02i, build EDK_lm_SP2.3+0
IP version for the TriMode_MAC_GMII (xps_ll_temac) is 1.00.b. EmacLite is
1.00.a
> Also, do both designs meet timing?
As the software guy here, I have no idea. I've passed the question onto the
electronics guys...
BUT, you put me on the right tracks: looking at the driver versions I
started playing with the options you get in the [Configure IP] of the
Ethernet Mac drivers. I couldn't get any improvement with the LLTEMAC, but
after I added the buffers to the EmacLite: IT WORKED !
I'm still having trouble mounting an NFS root and ssh'ing into the target,
but I guess that's another issue...
--
Guillaume Dargaud
http://www.gdargaud.net/
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: Xilinx network trouble (mostly solved)
2008-04-18 14:40 ` Xilinx network trouble (mostly solved) Guillaume Dargaud
@ 2008-04-18 16:14 ` Rick Moleres
2008-04-21 16:16 ` Guillaume Dargaud
0 siblings, 1 reply; 6+ messages in thread
From: Rick Moleres @ 2008-04-18 16:14 UTC (permalink / raw)
To: Guillaume Dargaud, linuxppc-dev
Guillaume,
You might try v1.00a of the xps_ll_temac core. There are a couple of
issues with v1.00b (posted previously on this list) that may (or may
not) be causing the problem you see.
-Rick
-----Original Message-----
From: linuxppc-dev-bounces+moleres=3Dxilinx.com@ozlabs.org
[mailto:linuxppc-dev-bounces+moleres=3Dxilinx.com@ozlabs.org] On Behalf =
Of
Guillaume Dargaud
Sent: Friday, April 18, 2008 8:40 AM
To: linuxppc-dev@ozlabs.org
Subject: Re: Xilinx network trouble (mostly solved)
Hello Rick,
> Which version of xps_ll_temac are you using, and which version of EDK?
Using xps 9.2.02i, build EDK_lm_SP2.3+0
IP version for the TriMode_MAC_GMII (xps_ll_temac) is 1.00.b. EmacLite
is=20
1.00.a
> Also, do both designs meet timing?
As the software guy here, I have no idea. I've passed the question onto
the=20
electronics guys...
BUT, you put me on the right tracks: looking at the driver versions I=20
started playing with the options you get in the [Configure IP] of the=20
Ethernet Mac drivers. I couldn't get any improvement with the LLTEMAC,
but=20
after I added the buffers to the EmacLite: IT WORKED !
I'm still having trouble mounting an NFS root and ssh'ing into the
target,=20
but I guess that's another issue...
--=20
Guillaume Dargaud
http://www.gdargaud.net/
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-dev
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Xilinx network trouble (mostly solved)
2008-04-18 16:14 ` Rick Moleres
@ 2008-04-21 16:16 ` Guillaume Dargaud
2008-04-21 16:50 ` Stephen Neuendorffer
0 siblings, 1 reply; 6+ messages in thread
From: Guillaume Dargaud @ 2008-04-21 16:16 UTC (permalink / raw)
To: Rick Moleres, linuxppc-dev
Hi Rick,
> You might try v1.00a of the xps_ll_temac core. There are a couple of
> issues with v1.00b (posted previously on this list) that may (or may
> not) be causing the problem you see.
Thanks for the answer... but I can't figure out how to revert to 1.00a. Is
the option accessible from xps ?
As for the EmacLite, it works rather imperfectly:
ping -s 2952 works
ping -s 2953 doesn't...:
[ 2340.664293] Badness at net/core/skbuff.c:317
[ 2340.668479] NIP: c015bd24 LR: c015b3f4 CTR: 00000000
[ 2340.673410] REGS: c7d0fd20 TRAP: 0700 Not tainted (2.6.24-rc8-xlnx)
[ 2340.679876] MSR: 00029030 <EE,ME,IR,DR> CR: 28000048 XER: 20000000
[ 2340.686189] TASK = c7c35810[243] 'ping' THREAD: c7d0e000
[ 2340.691279] GPR00: 00000001 c7d0fdd0 c7c35810 c7d318e0 80000009 00000004
c027bd38 00000200
[ 2340.699571] GPR08: c7d0fdb8 c7d0e000 c027bd1c c0159910 c7d0fdb8 100b46f4
ffffffff 00000000
[ 2340.707864] GPR16: 10090000 100bd6e0 100b0000 100b0000 00000c11 7fbcf9a0
00000000 10004634
[ 2340.716158] GPR24: 1007d370 7fbced90 10003fbc 100adf18 00000000 00000000
c7c9e000 c7d318e0
[ 2340.724625] NIP [c015bd24] skb_release_all+0x6c/0x94
[ 2340.729551] LR [c015b3f4] __kfree_skb+0x18/0xc0
[ 2340.734042] Call Trace:
[ 2340.736465] [c7d0fdd0] [c0161b6c] netif_rx+0x124/0x158 (unreliable)
[ 2340.742684] [c7d0fde0] [c015b3f4] __kfree_skb+0x18/0xc0
[ 2340.747866] [c7d0fdf0] [c01466a0] SendHandler+0x44/0x94
[ 2340.753050] [c7d0fe10] [c0147284] XEmacLite_InterruptHandler+0x12c/0x140
[ 2340.759702] [c7d0fe20] [c0146078] xemaclite_interrupt+0x20/0x34
[ 2340.765577] [c7d0fe30] [c003c98c] handle_IRQ_event+0x4c/0xa0
[ 2340.771193] [c7d0fe50] [c003ca94] __do_IRQ+0xb4/0x130
[ 2340.776203] [c7d0fe70] [c0006990] do_IRQ+0x6c/0xa4
[ 2340.780954] [c7d0fe80] [c000329c] ret_from_except+0x0/0x18
[ 2340.786397] [c7d0ff40] [c0002c50] ret_from_syscall+0x0/0x3c
[ 2340.791923] Instruction dump:
[ 2340.794865] 3000ffff 7c005a2c 7c00592d 40a2fff0 817f006c 54290024
2f8b0000 419e0024
[ 2340.802549] 8009000c 5400011e 7c0000d0 54000ffe <0f000000> 7fe3fb78
7d6903a6 4e800421
(192.168.1.185): 2953 data bytes
--
Guillaume Dargaud
http://www.gdargaud.net/
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: Xilinx network trouble (mostly solved)
2008-04-21 16:16 ` Guillaume Dargaud
@ 2008-04-21 16:50 ` Stephen Neuendorffer
0 siblings, 0 replies; 6+ messages in thread
From: Stephen Neuendorffer @ 2008-04-21 16:50 UTC (permalink / raw)
To: Guillaume Dargaud, Rick Moleres, linuxppc-dev
The easiest way is to edit the .mhs by hand.
Steve
> -----Original Message-----
> From: =
linuxppc-dev-bounces+stephen.neuendorffer=3Dxilinx.com@ozlabs.org
[mailto:linuxppc-dev-
> bounces+stephen.neuendorffer=3Dxilinx.com@ozlabs.org] On Behalf Of
Guillaume Dargaud
> Sent: Monday, April 21, 2008 9:16 AM
> To: Rick Moleres; linuxppc-dev@ozlabs.org
> Subject: Re: Xilinx network trouble (mostly solved)
>=20
> Hi Rick,
>=20
> > You might try v1.00a of the xps_ll_temac core. There are a couple
of
> > issues with v1.00b (posted previously on this list) that may (or may
> > not) be causing the problem you see.
>=20
> Thanks for the answer... but I can't figure out how to revert to
1.00a. Is
> the option accessible from xps ?
>=20
> As for the EmacLite, it works rather imperfectly:
> ping -s 2952 works
> ping -s 2953 doesn't...:
> [ 2340.664293] Badness at net/core/skbuff.c:317
> [ 2340.668479] NIP: c015bd24 LR: c015b3f4 CTR: 00000000
> [ 2340.673410] REGS: c7d0fd20 TRAP: 0700 Not tainted
(2.6.24-rc8-xlnx)
> [ 2340.679876] MSR: 00029030 <EE,ME,IR,DR> CR: 28000048 XER:
20000000
> [ 2340.686189] TASK =3D c7c35810[243] 'ping' THREAD: c7d0e000
> [ 2340.691279] GPR00: 00000001 c7d0fdd0 c7c35810 c7d318e0 80000009
00000004
> c027bd38 00000200
> [ 2340.699571] GPR08: c7d0fdb8 c7d0e000 c027bd1c c0159910 c7d0fdb8
100b46f4
> ffffffff 00000000
> [ 2340.707864] GPR16: 10090000 100bd6e0 100b0000 100b0000 00000c11
7fbcf9a0
> 00000000 10004634
> [ 2340.716158] GPR24: 1007d370 7fbced90 10003fbc 100adf18 00000000
00000000
> c7c9e000 c7d318e0
> [ 2340.724625] NIP [c015bd24] skb_release_all+0x6c/0x94
> [ 2340.729551] LR [c015b3f4] __kfree_skb+0x18/0xc0
> [ 2340.734042] Call Trace:
> [ 2340.736465] [c7d0fdd0] [c0161b6c] netif_rx+0x124/0x158 (unreliable)
> [ 2340.742684] [c7d0fde0] [c015b3f4] __kfree_skb+0x18/0xc0
> [ 2340.747866] [c7d0fdf0] [c01466a0] SendHandler+0x44/0x94
> [ 2340.753050] [c7d0fe10] [c0147284]
XEmacLite_InterruptHandler+0x12c/0x140
> [ 2340.759702] [c7d0fe20] [c0146078] xemaclite_interrupt+0x20/0x34
> [ 2340.765577] [c7d0fe30] [c003c98c] handle_IRQ_event+0x4c/0xa0
> [ 2340.771193] [c7d0fe50] [c003ca94] __do_IRQ+0xb4/0x130
> [ 2340.776203] [c7d0fe70] [c0006990] do_IRQ+0x6c/0xa4
> [ 2340.780954] [c7d0fe80] [c000329c] ret_from_except+0x0/0x18
> [ 2340.786397] [c7d0ff40] [c0002c50] ret_from_syscall+0x0/0x3c
> [ 2340.791923] Instruction dump:
> [ 2340.794865] 3000ffff 7c005a2c 7c00592d 40a2fff0 817f006c 54290024
> 2f8b0000 419e0024
> [ 2340.802549] 8009000c 5400011e 7c0000d0 54000ffe <0f000000> 7fe3fb78
> 7d6903a6 4e800421
> (192.168.1.185): 2953 data bytes
> --
> Guillaume Dargaud
> http://www.gdargaud.net/
>=20
>=20
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-dev
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2008-04-21 16:51 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-04-17 10:04 Xilinx network trouble Guillaume Dargaud
2008-04-17 15:56 ` Rick Moleres
2008-04-18 14:40 ` Xilinx network trouble (mostly solved) Guillaume Dargaud
2008-04-18 16:14 ` Rick Moleres
2008-04-21 16:16 ` Guillaume Dargaud
2008-04-21 16:50 ` Stephen Neuendorffer
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).