Mesa 7i92 Intermittent Read Errors

Hi,

I've checked the other posts on this issue "hm2/hm2_7i92.0: error finishing read!" with no luck so far employing the most straightforward solutions. The machine works fine until it disconnects from the error. Asking for help here before I dive into harder fixes (try new pc, reduce emi). Hal and ini files included!

error behaviour:
-intermittent (showing up after few seconds or hours)
-sometimes restarting the laptop makes the error take longer to show up
-certain restarts get the the error to appear in the first few minutes consistently when launching linuxcnc repeatedly
-the error still occurs when not operating the machine after initialising linuxcnc

solutions tried:
-changing eth cable to shielded
-jostling wires to look for poor connections/noise
-installing proprietary realtek eth drivers r8168
-checking 5v supply

laptop specs:
Release Linux Mint 19.2 Tina 64-bit
Kernel Linux 4.19.106-rt44-lcnc x86_64

CPU:
Topology: Dual Core model: Intel Core i5 M 430 bits: 64 type: MT MCP
L2 cache: 3072 KiB
Speed: 1852 MHz min/max: 1199/2267 MHz Core speeds (MHz): 1: 1862 2: 1862
3: 1862 4: 1862

Network:
Device-1: Broadcom and subsidiaries BCM43225 802.11b/g/n driver: wl
Device-2: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet
driver: r8168

Servo jitter is 81us (latency test)

halcmd show param *.tmax results:
Parameters:
Owner Type Dir Value Name
32 s32 RW 39938 debounce.0.tmax
26 s32 RW 0 hm2_7i92.0.read-request.tmax
26 s32 RW 1522316 hm2_7i92.0.read.tmax
26 s32 RW 946588 hm2_7i92.0.write.tmax
19 s32 RW 67660 motion-command-handler.tmax
19 s32 RW 119383 motion-controller.tmax
29 s32 RW 47012 pid.s.do-pid-calcs.tmax
29 s32 RW 52144 pid.x.do-pid-calcs.tmax
29 s32 RW 44124 pid.y.do-pid-calcs.tmax
29 s32 RW 49072 pid.z.do-pid-calcs.tmax
20 s32 RW 1656028 servo-thread.tmax

dmesg | tail -50 results:
[ 7.045766] ACPI Warning: SystemIO range 0x0000000000000530-0x000000000000053F conflicts with OpRegion 0x0000000000000500-0x0000000000000563 (\GPIO) (20180810/utaddress-213)
[ 7.045771] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
[ 7.045773] ACPI Warning: SystemIO range 0x0000000000000500-0x000000000000052F conflicts with OpRegion 0x0000000000000500-0x0000000000000563 (\GPIO) (20180810/utaddress-213)
[ 7.045778] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
[ 7.045779] lpc_ich: Resource conflict(s) found affecting gpio_ich
[ 7.049097] intel ips 0000:00:1f.6: CPU TDP doesn't match expected value (found 25, expected 29)
[ 7.049431] intel ips 0000:00:1f.6: failed to get i915 symbols, graphics turbo disabled until i915 loads
[ 7.049736] intel ips 0000:00:1f.6: IPS driver initialized, MCP temp limit 90
[ 7.553604] input: HP WMI hotkeys as /devices/virtual/input/input16
[ 7.913700] kvm: disabled by bios
[ 7.984707] kvm: disabled by bios
[ 8.041391] kvm: disabled by bios
[ 8.066940] Bluetooth: Core ver 2.22
[ 8.068836] NET: Registered protocol family 31
[ 8.068838] Bluetooth: HCI device and connection manager initialized
[ 8.068846] Bluetooth: HCI socket layer initialized
[ 8.068849] Bluetooth: L2CAP socket layer initialized
[ 8.068865] Bluetooth: SCO socket layer initialized
[ 8.106628] kvm: disabled by bios
[ 8.180980] usbcore: registered new interface driver btusb
[ 8.564106] enp3s0: 0xffffab7640a81000, 00:26:9e:ca:54:7e, IRQ 24
[ 8.639097] IPv6: ADDRCONF(NETDEV_UP): enp3s0: link is not ready
[ 8.861973] wl 0000:02:00.0 wlo1: renamed from wlan0
[ 8.969873] media: Linux media interface: v0.10
[ 9.005257] videodev: Linux video capture interface: v2.00
[ 9.129347] uvcvideo: Found UVC 1.00 device HP Webcam (064e:a127)
[ 9.174639] uvcvideo 2-1.5:1.0: Entity type for entity Extension 5 was not initialized!
[ 9.174646] uvcvideo 2-1.5:1.0: Entity type for entity Extension 4 was not initialized!
[ 9.174648] uvcvideo 2-1.5:1.0: Entity type for entity Processing 3 was not initialized!
[ 9.174651] uvcvideo 2-1.5:1.0: Entity type for entity Camera 1 was not initialized!
[ 9.201015] input: HP Webcam: HP Webcam as /devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.5/2-1.5:1.0/input/input17
[ 9.215078] usbcore: registered new interface driver uvcvideo
[ 9.215081] USB Video Class driver (1.1.1)
[ 9.635071] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[ 9.635075] Bluetooth: BNEP filters: protocol multicast
[ 9.635083] Bluetooth: BNEP socket layer initialized
[ 9.725630] IPv6: ADDRCONF(NETDEV_UP): wlo1: link is not ready
[ 10.703383] r8168: enp3s0: link up
[ 10.703403] IPv6: ADDRCONF(NETDEV_CHANGE): enp3s0: link becomes ready
[ 11.117826] random: crng init done
[ 11.117831] random: 7 urandom warning(s) missed due to ratelimiting
[ 11.123333] IPv6: ADDRCONF(NETDEV_UP): wlo1: link is not ready
[ 12.354031] IPv6: ADDRCONF(NETDEV_CHANGE): wlo1: link becomes ready
[ 15.229821] Bluetooth: RFCOMM TTY layer initialized
[ 15.229833] Bluetooth: RFCOMM socket layer initialized
[ 15.229843] Bluetooth: RFCOMM ver 1.11
[ 93.828103] perf: interrupt took too long (2527 > 2500), lowering kernel.perf_event_max_sample_rate to 79000
[ 121.937676] perf: interrupt took too long (3175 > 3158), lowering kernel.perf_event_max_sample_rate to 63000
[ 227.403295] perf: interrupt took too long (3973 > 3968), lowering kernel.perf_event_max_sample_rate to 50000
[ 315.181982] perf: interrupt took too long (4971 > 4966), lowering kernel.perf_event_max_sample_rate to 40000

26 s32 RW 1522316 hm2_7i92.0.read.tmax
26 s32 RW 946588 hm2_7i92.0.write.tmax

These times/1.8GHz total to more than 1 MS

My guess would be that like most laptops there is
power management that cannot be disabled

You might try a 2 ms servo thread

Thank you for the rapid response and i've made the change!

When calculating the times:
(read.tmax) -> 1522316 / 1.8GHz = ~0.846 ms
(write.tmax) -> 946588 / 1.8GHz = ~0.526 ms
(servo-thread.tmax) -> 1656028 / 1.8GHz = ~0.920 ms

Though the times are not explicitly over 1MS im guessing its close enough to matter? I'll report back later on the results.

Cheers

Quick update, haven't been able to do extensive tests yet ,but the watchdog did bite followed by a read error.

Previously the watchdog was set to 5ms and has now been bumped up to 10ms.

p.s ran the machine again and the read error has shown itself again

The laptop is a HP Pavilion dv6-2120tx
support.hp.com/au-en/document/c02033034
I believe it is stock besides replacing the hdd for a ssd.

In terms of bios it is minimalist and disabled everything I could. Didn't have power saving options I could disable.

I was able to get a hold of a dell optiplex 780 SFF. Going to swap in the ssd and install linuxcnc.

Running linuxcnc on the new computer I have not experienced the read error once since the last post. It seems the laptop was not up to spec.

Thank you for the help