Mesa hm2/hm2_7i96s.0: error finishing read

More
3 years 11 months ago #256233 by scotth
Running 5.19.0-2-rt-amd64 #1 SMP PREEMPT_RT Debian 5.19.11-1 (2022-09-24) x86_64 x86_64 x86_64 GNU/Linux.

Using the patch Rodw recommended.
forum.linuxcnc.org/27-driver-boards/4691...ead?start=130#256114

No failure in 24 hours of running.

Please Log in to join the conversation.

More
3 years 11 months ago #256240 by rodw
I gave up on getting the kernel compiled and upgraded to bookworm 6.0. I have a win!
Discussed here
forum.linuxcnc.org/9-installing-linuxcnc...9-13?start=10#256235

I still don't understand why ping times are way high with isolcpus=2,3 but the tmax values are awesome. No finishing read errors after an hour or so. and no real time delays.

Its clearly to do with network drivers.


 

Please Log in to join the conversation.

More
3 years 11 months ago #256246 by rodw

I didn't have IRQ coalescing disabled because I thought that was only for Intel ethernet controllers.

Code:
john@cave:~$ lspci | egrep -i 'network|ethernet' 07:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 06) 09:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8125 2.5GbE Controller (rev 04) [code][code]JT
[/code][/code]
The cards are always the same old suspects.
What drivers are installed for them?
Which one is the mesa card? sudo lshw -class networkorethtool -i eth0I know I've broken stuff for you before but the Odroid guys say to build the R8125 driver from the realtek source
The other one should be using the r8168-dkms driver
I wonder if we should be black listing existing drivers to get the R8168-dkm driver installed?
 

Please Log in to join the conversation.

More
3 years 11 months ago #256266 by rodw
My Brix was left on all day while I was at work and not one error finishing read showed up!

So I am trying again to compile the 6.1 kernel on i3 but its failing very early in the build. I am starting from the 6.0 kernel non-realtime and preempt_rt both fail,

First I needed to add a few more dependencies
Code:
sudo apt install build-essential bin86 #libncurses5-dev pkg-config libssl-dev flex bison qtbase5-dev openssl libelf-dev
Code:
Then I got
Code:
  CC      /home/debian/devel/rtlinux/linux-6.1-rc3/tools/bpf/resolve_btfids/str_error_r.o   LD      /home/debian/devel/rtlinux/linux-6.1-rc3/tools/bpf/resolve_btfids/resolve_btfids-in.o   LINK     resolve_btfids make: *** [Makefile:1254: scripts] Error 2 make: *** Waiting for unfinished jobs....
Code:
This article says its possible to build a deb with
Code:
make -j $(nproc) deb-pkg
Perhaps somebody how has got this working could try building the deb. That would be more useful in the long term. The article says how to install it too
Code:
[code]for pkg in ./*.deb; do sudo dpkg -i $pkg; done
[/code]

It would be cool if we could offer a packaged deb. Hosted in Google Drive or something.

Please Log in to join the conversation.

More
3 years 11 months ago #256267 by rodw

Please Log in to join the conversation.

More
3 years 11 months ago #256268 by arvidb

I am starting from the 6.0 kernel non-realtime and preempt_rt both fail,
Re-run the kernel build without the -j parameter (i.e. single threaded) to get a clearer error message. You might also have to scroll up a bit to find the actual error.

This article
says its possible to build a deb with
Code:
make -j $(nproc) deb-pkg
Perhaps somebody how has got this working could try building the deb. That would be more useful in the long term.
I built mine using the bindeb-pkg target and it worked fine. I used a localmodconfig though (a kernel configured specifically for my machine) so it won't be useful for others; it would be pointless to share it. But yes, it should be possible to build a debian package that could be shared.

Please Log in to join the conversation.

More
3 years 11 months ago #256271 by rodw
Thanks, add  
Code:
libssl-dev and rsync
to the dependencies. I thought I'd fixed the ssl one. Using 1 make thread certainly showed the issue.

I think it would be rare for a user to need to build a custom kerel with features other than prempt_rt so a Deb would be cool. I'll give it a go this time...
well on a second attempt now we have rsync installed...

See what its like by morning...

Please Log in to join the conversation.

More
3 years 11 months ago #256282 by JT

I didn't have IRQ coalescing disabled because I thought that was only for Intel ethernet controllers.

Code:
john@cave:~$ lspci | egrep -i 'network|ethernet' 07:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 06) 09:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8125 2.5GbE Controller (rev 04)
JT
The cards are always the same old suspects.
What drivers are installed for them?
Which one is the mesa card? sudo lshw -class networkorethtool -i eth0I know I've broken stuff for you before but the Odroid guys say to build the R8125 driver from the realtek source
The other one should be using the r8168-dkms driver
I wonder if we should be black listing existing drivers to get the R8168-dkm driver installed?
 
 

I get this:
Code:
john@cave:~$ sudo lshw -class networkorethtool -i eth0I [sudo] password for john: Hardware Lister (lshw) - usage: lshw [-format] [-options ...]        lshw -version     -version        print program version () format can be     -html           output hardware tree as HTML     -xml            output hardware tree as XML     -json           output hardware tree as a JSON object     -short          output hardware paths     -businfo        output bus information options can be     -class CLASS    only show a certain class of hardware     -C CLASS        same as '-class CLASS'     -c CLASS        same as '-class CLASS'     -disable TEST   disable a test (like pci, isapnp, cpuid, etc. )     -enable TEST    enable a test (like pci, isapnp, cpuid, etc. )     -quiet          don't display status     -sanitize       sanitize output (remove sensitive information like serial numbers, etc.)     -numeric        output numeric IDs (for PCI, USB, etc.)     -notime         exclude volatile attributes (timestamps) from output

JT

Please Log in to join the conversation.

More
3 years 11 months ago #256283 by JT
Running lshw -short to get a list of class names I get this:
Code:
john@cave:~$ sudo lshw -short H/W path Device Class Description ================================================================ system System Product Name (SKU) /0 bus TUF GAMING B550-PRO /0/0 memory 64KiB BIOS /0/31 memory 32GiB System Memory /0/31/0 memory [empty] /0/31/1 memory 16GiB DIMM DDR4 Synchronous Unbuffered (Unregist /0/31/2 memory [empty] /0/31/3 memory 16GiB DIMM DDR4 Synchronous Unbuffered (Unregist /0/34 memory 384KiB L1 cache /0/35 memory 3MiB L2 cache /0/36 memory 32MiB L3 cache /0/37 processor AMD Ryzen 5 5600X 6-Core Processor /0/100 bridge Starship/Matisse Root Complex /0/100/0.2 generic Starship/Matisse IOMMU /0/100/1.1 bridge Starship/Matisse GPP Bridge /0/100/1.1/0 /dev/nvme0 storage Samsung SSD 980 1TB /0/100/1.1/0/0 hwmon0 disk NVMe disk /0/100/1.1/0/2 /dev/ng0n1 disk NVMe disk /0/100/1.1/0/1 /dev/nvme0n1 disk 1TB NVMe disk /0/100/1.1/0/1/1 /dev/nvme0n1p1 volume 930GiB EXT4 volume /0/100/1.1/0/1/2 /dev/nvme0n1p2 volume 976MiB Extended partition /0/100/1.1/0/1/2/5 /dev/nvme0n1p5 volume 976MiB Linux swap volume /0/100/1.2 bridge Starship/Matisse GPP Bridge /0/100/1.2/0 bus 500 Series Chipset USB 3.1 XHCI Controller /0/100/1.2/0/0 usb1 bus xHCI Host Controller /0/100/1.2/0/0/6 input AURA LED Controller /0/100/1.2/0/0/7 bus USB2.0 Hub /0/100/1.2/0/0/7/3 input USB Receiver /0/100/1.2/0/0/7/3/0 input14 input Logitech Wireless Mouse /0/100/1.2/0/0/8 input2 input USB Keyboard Consumer Control /0/100/1.2/0/1 usb2 bus xHCI Host Controller /0/100/1.2/0.1 scsi3 storage 500 Series Chipset SATA Controller /0/100/1.2/0.1/0 /dev/cdrom disk DVD-ROM TS-H353B /0/100/1.2/0.1/1 /dev/sda disk 2TB WDC WD20EZRX-00D /0/100/1.2/0.1/1/1 /dev/sda1 volume 1863GiB EXT4 volume /0/100/1.2/0.2 bridge 500 Series Chipset Switch Upstream Port /0/100/1.2/0.2/0 bridge Advanced Micro Devices, Inc. [AMD] /0/100/1.2/0.2/1 bridge Advanced Micro Devices, Inc. [AMD] /0/100/1.2/0.2/2 bridge Advanced Micro Devices, Inc. [AMD] /0/100/1.2/0.2/3 bridge Advanced Micro Devices, Inc. [AMD] /0/100/1.2/0.2/3/0 enp7s0 network RTL8111/8168/8411 PCI Express Gigabit Ethernet C /0/100/1.2/0.2/8 bridge Advanced Micro Devices, Inc. [AMD] /0/100/1.2/0.2/9 bridge Advanced Micro Devices, Inc. [AMD] /0/100/1.2/0.2/9/0 enp9s0 network RTL8125 2.5GbE Controller /0/100/3.1 bridge Starship/Matisse GPP Bridge /0/100/3.1/0 bridge Navi 10 XL Upstream Port of PCI Express Switch /0/100/3.1/0/0 bridge Navi 10 XL Downstream Port of PCI Express Switch /0/100/3.1/0/0/0 /dev/fb0 display Navi 23 [Radeon RX 6600/6600 XT/6600M] /0/100/3.1/0/0/0.1 card0 multimedia Navi 21/23 HDMI/DP Audio Controller /0/100/3.1/0/0/0.1/0 input17 input HDA ATI HDMI HDMI/DP,pcm=3 /0/100/3.1/0/0/0.1/1 input18 input HDA ATI HDMI HDMI/DP,pcm=7 /0/100/3.1/0/0/0.1/2 input19 input HDA ATI HDMI HDMI/DP,pcm=8 /0/100/3.1/0/0/0.1/3 input20 input HDA ATI HDMI HDMI/DP,pcm=9 /0/100/3.1/0/0/0.1/4 input21 input HDA ATI HDMI HDMI/DP,pcm=10 /0/100/7.1 bridge Starship/Matisse Internal PCIe GPP Bridge 0 to b /0/100/7.1/0 generic Starship/Matisse PCIe Dummy Function /0/100/8.1 bridge Starship/Matisse Internal PCIe GPP Bridge 0 to b /0/100/8.1/0 generic Starship/Matisse Reserved SPP /0/100/8.1/0.1 generic Starship/Matisse Cryptographic Coprocessor PSPCP /0/100/8.1/0.3 bus Matisse USB 3.0 Host Controller /0/100/8.1/0.3/0 usb3 bus xHCI Host Controller /0/100/8.1/0.3/1 usb4 bus xHCI Host Controller /0/100/8.1/0.4 card1 multimedia Starship/Matisse HD Audio Controller /0/100/8.1/0.4/0 input22 input HDA Digital PCBeep /0/100/8.1/0.4/1 input23 input HD-Audio Generic Front Mic /0/100/8.1/0.4/2 input24 input HD-Audio Generic Rear Mic /0/100/8.1/0.4/3 input25 input HD-Audio Generic Line /0/100/8.1/0.4/4 input26 input HD-Audio Generic Line Out Front /0/100/8.1/0.4/5 input27 input HD-Audio Generic Line Out Surround /0/100/8.1/0.4/6 input28 input HD-Audio Generic Line Out CLFE /0/100/8.1/0.4/7 input29 input HD-Audio Generic Front Headphone /0/100/14 bus FCH SMBus Controller /0/100/14.3 bridge FCH LPC Bridge /0/100/14.3/0 system PnP device PNP0c01 /0/100/14.3/1 system PnP device PNP0c02 /0/100/14.3/2 system PnP device PNP0b00 /0/100/14.3/3 system PnP device PNP0c02 /0/100/14.3/4 communication PnP device PNP0501 /0/100/14.3/5 system PnP device PNP0c02 /0/101 bridge Starship/Matisse PCIe Dummy Host Bridge /0/102 bridge Starship/Matisse PCIe Dummy Host Bridge /0/103 bridge Starship/Matisse PCIe Dummy Host Bridge /0/104 bridge Starship/Matisse PCIe Dummy Host Bridge /0/105 bridge Starship/Matisse PCIe Dummy Host Bridge /0/106 bridge Starship/Matisse PCIe Dummy Host Bridge /0/107 bridge Starship/Matisse PCIe Dummy Host Bridge /0/108 bridge Matisse/Vermeer Data Fabric: Device 18h; Functio /0/109 bridge Matisse/Vermeer Data Fabric: Device 18h; Functio /0/10a bridge Matisse/Vermeer Data Fabric: Device 18h; Functio /0/10b bridge Matisse/Vermeer Data Fabric: Device 18h; Functio /0/10c bridge Matisse/Vermeer Data Fabric: Device 18h; Functio /0/10d bridge Matisse/Vermeer Data Fabric: Device 18h; Functio /0/10e bridge Matisse/Vermeer Data Fabric: Device 18h; Functio /0/10f bridge Matisse/Vermeer Data Fabric: Device 18h; Functio /1 input0 input Power Button /2 input1 input Power Button /3 input15 input PC Speaker /4 input16 input Eee PC WMI hotkeys

JT

Please Log in to join the conversation.

More
3 years 11 months ago #256296 by rodw
@JT, so it looks like you are using the RT8125 for your mesa card. Have you tried swapping it to the other NIC?

There is something really screwy going on with the kernel drivers. Re the 8125 on an Odroid H2+,  Rob Murphy (aka BeagleBrainz) said to mewhen I was using a 5.x RT kernel with the H2+ I was getting following errors when jogging. What it came down to is the Realtek drivers that come with the kernel have issues. All I did was download the driver source from Realtek, as is required with the 4.x kernel, built and installed that. No issues since. This was a good few months back. ATM I'm running Mint 21 with a Debian RT kernel. The way I have it setup I can update Linuxcnc via apt-get update using the Debian repos. Debian calls it apt pinning.This explains it:  wiki.odroid.com/odroid-h3/hardware/insta...net_driver_on_h3plus "There is an r8169 driver for the RTL8125B series in the Linux kernel 5.9 and above, but the stock driver in the Linux kernel tree has 15~30% lower performance than the vendor's custom driver (Realtek) on a 2.5GbE subnet."
The link to the driver is on the odroid page above.

So I think it really confirms the kernel drivers for Realtek are crappy! RTL8125 support was added in kernel 5.9.Re the debs, I failed to build them as I got tangled up with certifiates and keys which I don't know enough about. I just found some Debian docs, that explain how to fix that, I might try again.  wiki.debian.org/BuildADebianKernelPackage

Please Log in to join the conversation.

Time to create page: 0.281 seconds
Powered by Kunena Forum