I sent a subscription to the linux-modules list which seemed to be
successful, but when I sent a message, it bounced. I sent the message
to linux-modules@vger.kernel.org. Did I do something wrong?
Hi,
Le 29/05/2022 à 13:35, Leslie Rhorer a écrit :
I sent a subscription to the linux-modules list which seemed to be
successful, but when I sent a message, it bounced. I sent the message
to linux-modules@vger.kernel.org. Did I do something wrong?
Obviously you did it right as I got your message through the modules
list (I'm not subscribed to linux-kernel list).
Christophe
From: Randy Dunlap <hidden> Date: 2022-05-29 15:41:30
On 5/29/22 08:21, Christophe Leroy wrote:
Hi,
Le 29/05/2022 à 13:35, Leslie Rhorer a écrit :
quoted
I sent a subscription to the linux-modules list which seemed to be
successful, but when I sent a message, it bounced. I sent the message
to linux-modules@vger.kernel.org. Did I do something wrong?
Obviously you did it right as I got your message through the modules
list (I'm not subscribed to linux-kernel list).
Christophe
OK, well the message, as I said, was bounced. It's rather urgent. Is
there a limit on the size of the message? Even though I removed much of
the content, the message was still rather large, since it contained a
filtered output of dmesg. I will try splitting t up. Can attachments
be made in the list? Zipping the contents of the dmesg output will cut
the size greatly. I will also try zipping the dmesg text.
On 5/29/2022 10:41 AM, Randy Dunlap wrote:
On 5/29/22 08:21, Christophe Leroy wrote:
quoted
Hi,
Le 29/05/2022 à 13:35, Leslie Rhorer a écrit :
quoted
I sent a subscription to the linux-modules list which seemed to be
successful, but when I sent a message, it bounced. I sent the message
to linux-modules@vger.kernel.org. Did I do something wrong?
Obviously you did it right as I got your message through the modules
list (I'm not subscribed to linux-kernel list).
Christophe
If this is the wrong list, please let me know the correct one so I
can subscribe and post there.
Two of my servers are having bizarre problems after upgrading to
Debian Bullseye. After attempting numerous fixes, I decided to wipe the
systems and do a fresh install. I did a fresh install of he first from
a netinst thumb drive, but the .iso does not contain the bnx2c drivers
needed for the ASUS PEB-10G/57811-1S 10GbE SFP+ Network Adapters used by
both systems. I tried adding what I thought to be the correct non-free
drivers, but I still got errors about the bnx2x drivers and the
interface shows no carrier. Since no carrier is detected on the
interface, no LAN or internet access is possible. The system is
effectively dead, although it boots up and of course I have local
console access.
After several different approaches, I changed from the netinst
image to a full non-free image. It still would not properly load the
bnx2x drivers. I tried downloading all six bnx2x e1, e1h, and e2
release 7.13.15.0 and 7.13.21.0 firmware files and adding them to the
initram image. It no longer complains about missing files, but it
doesn't activate the NIC, either. These servers are both essentially
dead in the water until I can get this fixed.
The only thing I spot from dmesg that seems possibly related is:
[ 1.420748] r8169 0000:0b:00.0: can't disable ASPM; OS doesn't have
ASPM control
[ 1.433467] r8169 0000:0b:00.0 eth0: RTL8168f/8111f,
50:46:5d:65:15:9c, XID 480, IRQ 39
[ 1.433469] r8169 0000:0b:00.0 eth0: jumbo features [frames: 9194
bytes, tx checksumming: ko]
...
[ 1.458361] r8169 0000:0b:00.0 enp11s0: renamed from eth0
[ 1.465675] bnx2x: disagrees about version of symbol module_layout
...
[ 12.976323] r8169 0000:0b:00.0: firmware: failed to load
rtl_nic/rtl8168f-1.fw (-2)
[ 12.976478] firmware_class: See https://wiki.debian.org/Firmware for
information about missing firmware
[ 12.976632] r8169 0000:0b:00.0: Direct firmware load for
rtl_nic/rtl8168f-1.fw failed with error -2
[ 12.976636] r8169 0000:0b:00.0: Unable to load firmware
rtl_nic/rtl8168f-1.fw (-2)
[ 12.977326] RTL8211E Gigabit Ethernet r8169-0-b00:00: attached PHY
driver [RTL8211E Gigabit Ethernet] (mii_bus:phy_addr=r8169-0-b00:00,
irq=IGNORE)
[ 13.046445] r8169 0000:0b:00.0 enp11s0: Link is Down
I would not expect the RTL8168 or RTL8169 firmware to be necessary,
but is it? Either way, why isn't it on the image? I would ask the same
thing about the bnx2x drivers, but they seem to have a bit of a history
of being fragmented. All that aside, how can I get this fixed in a hurry?
This is an AMD system with an 8 core FX-8350 CPU. The install
image is Debian GNU/Linux 11.3.0 "Bullseye" - Unofficial amd64 DVD
Binary-1 with firmware 20220326-11:23. I pulled the extra firmware from
git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/bnx2x/
Below is the output of dmesg after removing quite a few of what I
am certain are unrelated lines:
[ 0.000000] Linux version 5.10.0-13-amd64
(debian-kernel@lists.debian.org) (gcc-10 (Debian 10.2.1-6) 10.2.1
20210110, GNU ld (GNU Binutils for Debian) 2.35.2) #1 SMP Debian
5.10.106-1 (2022-03-17)
[ 1.465675] bnx2x: disagrees about version of symbol module_layout
[ 12.075903] bnx2x: disagrees about version of symbol module_layout
Those two messages means that you are trying to use modules that were
built for a different kernel version.
As far as I can see you are using kernel 5.10
You have to use bnx2 modules for that exact kernel.
Christophe
If this is the wrong list, please let me know the correct one so I
can subscribe and post there.
Two of my servers are having bizarre problems after upgrading to
Debian Bullseye. After attempting numerous fixes, I decided to wipe the
systems and do a fresh install. I did a fresh install of he first from
a netinst thumb drive, but the .iso does not contain the bnx2c drivers
needed for the ASUS PEB-10G/57811-1S 10GbE SFP+ Network Adapters used by
both systems. I tried adding what I thought to be the correct non-free
drivers, but I still got errors about the bnx2x drivers and the
interface shows no carrier. Since no carrier is detected on the
interface, no LAN or internet access is possible. The system is
effectively dead, although it boots up and of course I have local
console access.
After several different approaches, I changed from the netinst
image to a full non-free image. It still would not properly load the
bnx2x drivers. I tried downloading all six bnx2x e1, e1h, and e2
release 7.13.15.0 and 7.13.21.0 firmware files and adding them to the
initram image. It no longer complains about missing files, but it
doesn't activate the NIC, either. These servers are both essentially
dead in the water until I can get this fixed.
The only thing I spot from dmesg that seems possibly related is:
...
[ 12.976323] r8169 0000:0b:00.0: firmware: failed to load
rtl_nic/rtl8168f-1.fw (-2)
[ 12.976478] firmware_class: See https://wiki.debian.org/Firmware for
information about missing firmware
[ 12.976632] r8169 0000:0b:00.0: Direct firmware load for
rtl_nic/rtl8168f-1.fw failed with error -2
[ 12.976636] r8169 0000:0b:00.0: Unable to load firmware
rtl_nic/rtl8168f-1.fw (-2)
Do you have that file rtl8168f-1.fw in your setup ? Is it at the correct
place ?
By the way, this seems to be more a Debian problem than a kernel
problem. Have you asked your question to Debian support ?
Christophe
Below is the output of dmesg after removing quite a few of what I
am certain are unrelated lines:
[ 0.000000] Linux version 5.10.0-13-amd64
(debian-kernel@lists.debian.org) (gcc-10 (Debian 10.2.1-6) 10.2.1
20210110, GNU ld (GNU Binutils for Debian) 2.35.2) #1 SMP Debian
5.10.106-1 (2022-03-17)
quoted
[ 1.465675] bnx2x: disagrees about version of symbol module_layout
quoted
[ 12.075903] bnx2x: disagrees about version of symbol module_layout
Those two messages means that you are trying to use modules that were
built for a different kernel version.
That is pretty clear
As far as I can see you are using kernel 5.10
You have to use bnx2 modules for that exact kernel.
Right. I thought of compiling them myself, but so far I have not been
successful in loading the kernel headers. I have tried adding a line to
the sources.list file, but so far I haven't found a directory in the
installation drive that apt doesn't complain about.
If this is the wrong list, please let me know the correct one so I
can subscribe and post there.
Two of my servers are having bizarre problems after upgrading to
Debian Bullseye. After attempting numerous fixes, I decided to wipe the
systems and do a fresh install. I did a fresh install of he first from
a netinst thumb drive, but the .iso does not contain the bnx2c drivers
needed for the ASUS PEB-10G/57811-1S 10GbE SFP+ Network Adapters used by
both systems. I tried adding what I thought to be the correct non-free
drivers, but I still got errors about the bnx2x drivers and the
interface shows no carrier. Since no carrier is detected on the
interface, no LAN or internet access is possible. The system is
effectively dead, although it boots up and of course I have local
console access.
After several different approaches, I changed from the netinst
image to a full non-free image. It still would not properly load the
bnx2x drivers. I tried downloading all six bnx2x e1, e1h, and e2
release 7.13.15.0 and 7.13.21.0 firmware files and adding them to the
initram image. It no longer complains about missing files, but it
doesn't activate the NIC, either. These servers are both essentially
dead in the water until I can get this fixed.
The only thing I spot from dmesg that seems possibly related is:
...
[ 12.976323] r8169 0000:0b:00.0: firmware: failed to load
rtl_nic/rtl8168f-1.fw (-2)
[ 12.976478] firmware_class: See https://wiki.debian.org/Firmware for
information about missing firmware
[ 12.976632] r8169 0000:0b:00.0: Direct firmware load for
rtl_nic/rtl8168f-1.fw failed with error -2
[ 12.976636] r8169 0000:0b:00.0: Unable to load firmware
rtl_nic/rtl8168f-1.fw (-2)
Do you have that file rtl8168f-1.fw in your setup ? Is it at the correct
place ?
I will check. Without remote LAN support, it isn't very easy.
By the way, this seems to be more a Debian problem than a kernel
problem. Have you asked your question to Debian support ?
They and I do not get along. More importantly, while at first glance
this might seem to be the case, I don't think it is. The github
repository at
git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/bnx2x/
Is not a Debian thing. It is a kernel module thing for all distros.
It is from there I downloaded the firmware (after it failed to load
during install), and it is that which is failing. As I said, it could
possibly be the realtek drivers also need to be loaded, and they are
not, but I really would not expect it to be the case.
I was able to get the 1G interface on the motherboard up, so now I have
LAN connectoivity and apt is working.
It is kernel 5.10.0-13-amd64
On 5/30/2022 1:08 AM, Christophe Leroy wrote:
Le 30/05/2022 à 01:46, Leslie Rhorer a écrit :
quoted
Below is the output of dmesg after removing quite a few of what I
am certain are unrelated lines:
[ 0.000000] Linux version 5.10.0-13-amd64
(debian-kernel@lists.debian.org) (gcc-10 (Debian 10.2.1-6) 10.2.1
20210110, GNU ld (GNU Binutils for Debian) 2.35.2) #1 SMP Debian
5.10.106-1 (2022-03-17)
quoted
[ 1.465675] bnx2x: disagrees about version of symbol module_layout
quoted
[ 12.075903] bnx2x: disagrees about version of symbol module_layout
Those two messages means that you are trying to use modules that were
built for a different kernel version.
As far as I can see you are using kernel 5.10
You have to use bnx2 modules for that exact kernel.
Christophe
I may have a valid module ready to load copied from an archive of the
kernel I am using on this system. I have placed it in
/lib/modules/5.10.0-13-amd64/kernel/drivers/net/ethernet/bnx2x/
How do I proceed with attempting to load the driver? Is there some way
I can insure the system will still boot with the existing LAN intact?
It looks like the system may try to eliminate the 1G interface when a
module for the 10G exists, whether the 10G is working or not.
On 5/30/2022 4:50 AM, Leslie Rhorer wrote:
I was able to get the 1G interface on the motherboard up, so now I
have LAN connectoivity and apt is working.
It is kernel 5.10.0-13-amd64
On 5/30/2022 1:08 AM, Christophe Leroy wrote:
quoted
Le 30/05/2022 à 01:46, Leslie Rhorer a écrit :
quoted
Below is the output of dmesg after removing quite a few of what I
am certain are unrelated lines:
[ 0.000000] Linux version 5.10.0-13-amd64
(debian-kernel@lists.debian.org) (gcc-10 (Debian 10.2.1-6) 10.2.1
20210110, GNU ld (GNU Binutils for Debian) 2.35.2) #1 SMP Debian
5.10.106-1 (2022-03-17)
quoted
[ 1.465675] bnx2x: disagrees about version of symbol module_layout
quoted
[ 12.075903] bnx2x: disagrees about version of symbol module_layout
Those two messages means that you are trying to use modules that were
built for a different kernel version.
As far as I can see you are using kernel 5.10
You have to use bnx2 modules for that exact kernel.
Christophe
From: Luis Chamberlain <mcgrof@kernel.org> Date: 2022-06-03 15:09:23
On Thu, Jun 02, 2022 at 11:00:00PM -0500, Leslie Rhorer wrote:
I may have a valid module ready to load copied from an archive of the
kernel I am using on this system. I have placed it in
Huh? You should not be copying around modules loosely from other
kernels.
/lib/modules/5.10.0-13-amd64/kernel/drivers/net/ethernet/bnx2x/
How do I proceed with attempting to load the driver? Is there some way I
can insure the system will still boot with the existing LAN intact? It looks
like the system may try to eliminate the 1G interface when a module for the
10G exists, whether the 10G is working or not.
Sounds like you want to engage with the bnx2x developers and report
whatever it is you are seeing. But note, you should probably try a
more recent kernel.
./scripts/get_maintainer.pl -f drivers/net/ethernet/broadcom/bnx2x/
Ariel Elior [off-list ref] (supporter:BROADCOM BNX2X 10 GIGABIT ETHERNET DRIVER)
Sudarsana Kalluru [off-list ref] (supporter:BROADCOM BNX2X 10 GIGABIT ETHERNET DRIVER)
Manish Chopra [off-list ref] (supporter:BROADCOM BNX2X 10 GIGABIT ETHERNET DRIVER)
"David S. Miller" [off-list ref] (maintainer:NETWORKING DRIVERS)
Eric Dumazet [off-list ref] (maintainer:NETWORKING DRIVERS)
Jakub Kicinski [off-list ref] (maintainer:NETWORKING DRIVERS)
Paolo Abeni [off-list ref] (maintainer:NETWORKING DRIVERS)
netdev@vger.kernel.org (open list:BROADCOM BNX2X 10 GIGABIT ETHERNET DRIVER)
linux-kernel@vger.kernel.org (open list)
Luis
On Thu, Jun 02, 2022 at 11:00:00PM -0500, Leslie Rhorer wrote:
quoted
I may have a valid module ready to load copied from an archive of the
kernel I am using on this system. I have placed it in
Huh? You should not be copying around modules loosely from other
kernels.
I know that. This not from "another" kernel. It is from an archive of
the same kernel. IOW, I have another machine running a 5.10.0-15-amd64
kernel that was upgraded from 5.10.0-13-amd64, and still has the 0-13
bnx2x driver still in the archives.
quoted
/lib/modules/5.10.0-13-amd64/kernel/drivers/net/ethernet/bnx2x/
How do I proceed with attempting to load the driver? Is there some way I
can insure the system will still boot with the existing LAN intact? It looks
like the system may try to eliminate the 1G interface when a module for the
10G exists, whether the 10G is working or not.
Sounds like you want to engage with the bnx2x developers and report
whatever it is you are seeing.
I have been trying (and apparently so have others) with no success.
But note, you should probably try a more recent kernel.
The same problem exists there. Indeed, I need to wipe the other server
and reload its OS, but I will run into precisely the same issue, since
it is the 5.10.0-13-amd64 that is on the installation DVD. I am
physically disabled, so having to reconfigure hardware for a reinstall
is a real problem.
I suppose I could upgrade the new server to 5.10.0-15-amd64, but then I
still need to install the 5.10.0-15-amd64 driver. No matter what, I
need to insert the module. When I look up the man for insmod, it says,
"don't use insmod, let the kernel do it", or words to that effect. So
now, what? One way or anotehr, I need to get this system working, and
then I need to do the same with the other system.