Adapting F-96-CE Puppy Linux for a Very Recent Laptop (3/2024)

[vtpup]

HP Envy Laptop 17t-cr100
Fossapup F-96 CE rev 4
Huge kernel: huge-6.1.8-fossapup64

I get a boot error message every time I start. It's not fatal, and boot continues after a 60 second pause. But I'd like to know what it means and what to do to correct it.

The white text says:

"Adding module intel_ishtp_hidmodprobe: can't load module intel_ishtp_hid (kernel/drivers/hid/intel-ish-hid/intel-ishtp-hid.ko): No such device"

Last line of red text says:

"hid_sensor_hub: module verification failed: signature and/or required key missing - tainting kernel"

When I check after booting, the following exists:

/lib/modules/6.1.8/kernel/drivers/hid/intel-ish-hid/intel-ishtp-hid.ko

When I check Boot Manager after booting, I find that "intel_ishtp_hid" module IS loaded.

I do notice that in all of the above there is a mix of underscores and dashes. I'm not sure if that's normal??

If I check my main Fossapup directory I find a file called fossapup64initmodules.txt with the following contents:

Code: Select all

hid_sensor_custom,hid_multitouch,hid_sensor_hub,intel_ishtp_hid,i2c_hid_acpi,i2c_hid,intel_ish_ipc,intel_ishtp

Below is a photo of the screen:

BootError.jpg (75.6 KiB) Viewed 1616 times

[ozsouth]

@vtpup - you need to edit the ... initmodules.txt file, commenting out (or removing) module lines causing problem (any 'not found' module). If changing kernels, one must remove this file & it will be recreated next boot, hopefully more appropriately. I only noticed this behaviour late last year.

fixboot.jpg (24.26 KiB) Viewed 1568 times

[vtpup]

That worked! Thanks ozsouth.

[vtpup]

@ozsouth Oops not so fast. That touchpad stopped responding. I had to modprobe intel_ishtp_hid to get the cursor to move again.

[rockedge]

maybe the code to build the module wasn't "signed" with the proper keys so when the kernel is loading modules this one does not have the right hash keys and is considered "tainted".

I had read something to this affect and was worried a few years ago that problems when woof-CE generates kernels with the kernel-kit would appear because we had no mechanism to "sign" with valid keys the modules. We didn't seem to encounter any problems caused by not signing the kernel so it's most likely not the problem.

So F96-CE_4 will pause after spitting errors then proceeds to boot "normally"?

[vtpup]

Thanks Rockedge. I don't know. I've been having intermittent problems with this touchpad, so I wonder if the drivers and kernel are mature enough even with this 6.1.8 huge kernel swap. I'm looking forward to a huge kernel past 6.8, because I also don't have speaker sound, which I believe gets included after that point. I don't actually care about the speaker sound -- headphones are fine. But the touchpad is a different matter. I have to have that working perfectly.

The cursor will freeze at odd times. I think related to possibly pressing the touchpad's (real) left button while dragging something. Seems more frequent when a browser (Seamonkey or Firefox) is running, but I'm not able to pin it definitely on something. I've tried a million adjustments in Psynclient, and have got it a little better, but it's still maddening at times.

A usb mouse always works to rescue me then.

I do wonder if my kernel, etc is late enough. This laptop was built at HP in February. That's how new it is. I ordered direct from them, and with the options I chose. 15-13500H proc, nvme 1 tb hd etc. It's just too damn new, I think.

[ozsouth]

@vtpup - I made a 6.8.1 low-latency kernel (ubuntu derivative flavour) for rockedge. Both of us confirmed it works. I pm'd you the link if you'd like to try it. Remember to delete the ... initmodules.txt before 1st reboot if you use it.

As for your needed module, you can late-load it by puttting it in a script in /etc/init.d . Touchpad would activate up to 5 sec after desktop shows.

[Clarity]

Hello @ozsouth
Is there any instructions or documentation on use of the kernel 'late-load' ability somewhere I can review?

Curious

[ozsouth]

@Clarity - no, this is just a fairly basic workaround. I've had to use this occasionally with difficult drivers.

I to add to the linux/kernel line in grub/syslinux cfg: modprobe.blacklist=some_module
Then make a script to load it after bootup, like this:

#!/bin/sh
modprobe some_module

Script is composed in texteditor, saved (as say start_some_module) in /etc/init.d , then made executable - in terminal:
chmod 755 /etc/init.d/start_some_module

Next bootup it would be stopped from loading at boot, then loaded after boot.
Note that could still cause a hangup for some rogue modules.

[mikewalsh]

@vtpup :-

I'm a-wonderin', Steve. I agree that your hardware is SO new that the kernel hasn't had a chance to catch up yet.....but is your touchpad actually a "touchpad", or is it one of these new-fangled "clickpads" - buttons under the surface of the touchpad - with physical buttons duplicated below it?

Wouldn't be the first time I've seen similar; the old Latitude has the standard touchpad + buttons where you'd expect to find them, then the buttons are duplicated immediately above the touchpad for the trackstick pointing device in the middle of the keyboard.....

Perhaps "hidden", sub-surface buttons might explain the erratic behaviour of your cursor. Probably nowt to do with it, so if this suggestion is irrelevant feel free to ignore it!

Just a thought, FWIW.

Mike.

[vtpup]

Sorry to take a bit to answer all of your posts, guys, thanks so much for them:

maybe the code to build the module wasn't "signed" with the proper keys so when the kernel is loading modules this one does not have the right hash keys and is considered "tainted".

I had read something to this affect and was worried a few years ago that problems when woof-CE generates kernels with the kernel-kit would appear because we had no mechanism to "sign" with valid keys the modules. We didn't seem to encounter any problems caused by not signing the kernel so it's most likely not the problem.

So F96-CE_4 will pause after spitting errors then proceeds to boot "normally"?

My understanding of "tainted" kernels is that it's only a matter of not being able to enter kernel bug reports because the kernel was compiled with something unsupported (or unrecognized). That's only from something I read. And yes F-96-CE_4 (with the huge kernel swap) will pause for 60 seconds, but then proceed to boot normally for me.

@vtpup - I made a 6.8.1 low-latency kernel (ubuntu derivative flavour) for rockedge. Both of us confirmed it works. I pm'd you the link if you'd like to try it. Remember to delete the ... initmodules.txt before 1st reboot if you use it.

As for your needed module, you can late-load it by puttting it in a script in /etc/init.d . Touchpad would activate up to 5 sec after desktop shows.

That's great! Thanks kindly ozsouth, I'll try it.

@vtpup :-

I'm a-wonderin', Steve. I agree that your hardware is SO new that the kernel hasn't had a chance to catch up yet.....but is your touchpad actually a "touchpad", or is it one of these new-fangled "clickpads" - buttons under the surface of the touchpad - with physical buttons duplicated below it?

Wouldn't be the first time I've seen similar; the old Latitude has the standard touchpad + buttons where you'd expect to find them, then the buttons are duplicated immediately above the touchpad for the trackstick pointing device in the middle of the keyboard.....

Perhaps "hidden", sub-surface buttons might explain the erratic behaviour of your cursor. Probably nowt to do with it, so if this suggestion is irrelevant feel free to ignore it!

Just a thought, FWIW.

Mike.

Thanks Mike. Yes it is the type of touchpad where there are two physical switches beneath the lower portion of the touchpad. I use these like right and left mouse buttons, For instance, to drag an app's icon across the desktop, I would push down on the lower left corner of the touchpad with my left index finger, and slide my right finger over the touchpad to position the icon where I wanted it, and then let up with my left finger.

This is a habit picked up from my older notebook, which actually had two physical mouse buttons separate from the touchpad, and just below it. The difference now is thephysical switches are located under the touchpad surface. The touchpad corners move physically up and down for these switches.

And BTW, the cursor freeze seems to happen most often when I do this left press and drag maneuver. And often when a browser window is open. I'm not absolutely certain these are the only times.

Another symptom is that I can sometimes get partial cursor function back by holding that left button down. BUT this often causes all kinds of unwanted behavior, like dragging things I don't want, or activating thiings I don't want, etc. It can be crazy-making to say the least!

Rebooting usually straightens the touchpad out again. Restarting X, does not seem to help.

I have read elsewhere that some touchpads now have their own dedicated processor, to reduce load and power usage on the main CPU. Amazing how complicated things are getting. Obviously if you don't have the right , firmware, drivers. communication protocols, optiions timing and settings for whatever version touchpad was installed by the manufacturer, things can get weird.

[rockedge]

I made a 6.8.1 low-latency kernel (ubuntu derivative flavour) for rockedge. Both of us confirmed it works.

I am seeing excellent performance with this 6.8.1 low-latency kernel in F96-CE_4. Also noticed good performance running QEMU virtual machines. Lately have been using the -cpu host parameter on the QEMU start command line.

[Chelsea80]

@vtpup

Further to post by @mikewalsh

Found this:

Keyboard & Input
Keyboard with dedicated number pad

There's a separate numeric pad located on the right side of the keyboard.

Keyboard backlight

Keyboard backlight makes typing in dark environments easier.

Touchpad with click buttons built into the surface

The left and right click buttons are built into the surface of the touchpad, as opposed to separate physical buttons.

Here:

https://laptoping.com/specs/product/hp-envy-17t-cr100/

May help

Chelsea80

[vtpup]

@ozsouth and @rockedge sorry but the new kernel didn't work with touchpad, wifi, or speaker sound. The procedure I used was a new install on USB using f2stickpup, then replacing vmlinuz and fdrv_fossapup64_9.6.sfs with the two new files, and renaming them to what was removed. I left everything else as stock new install.

Using a usb mouse I tried to use the mouse/keyboard manager's touchpad manager, but it reported "no synaptics touchpad found"

I tried a few of the network managers, but none could find an interface. I tried forcing iwlwifi, which I know to be the correct overall interface, but the device still wasn't detected. The wifi card is an intel wifi 6e ax211 -- pretty standard card.

For speakers, I did find that there were a lot new available high fidelity digital and analog choices for the sound system, and headphones worked well, and speaker controls were found, and even the VU meters showed sound happening in the speaker controls sections, but no actual sound was coming out of the speakers themselves. (I don't really care about this, though, just reporting for feedback)

Maybe additional drivers/firmware/modules are required. The main interest is the touchpad. I'm pretty sure I can find the missing puzzle piece for wifi. Sound is acceptable if headphone only, as long as I get a sane touchpad.

[mikewalsh]

@vtpup :-

@ozsouth and @rockedge sorry but the new kernel didn't work with touchpad, wifi, or speaker sound. The procedure I used was a new install on USB using f2stickpup, then replacing vmlinuz and fdrv_fossapup64_9.6.sfs with the two new files, and renaming them to what was removed. I left everything else as stock new install.

Umm.....what about the zdrv? If you've replaced vmlinuz, then this is a must.....because that's where the modules live that are compiled specifically to work with that kernel release version. The frdv is for firmware, and CAN in some cases be optional.....but there's nothing optional about the zdrv, unless my understanding of the kernel-swap procedure in these more modern Puppies is all out of whack!

Since I've performed many successful kernel-swaps over the last few years, I don't think I'm SO far out....

(*shrug...*)

Again, just a thought, FWIW.

Mike.

[vtpup]

@vtpup :-

@ozsouth and @rockedge sorry but the new kernel didn't work with touchpad, wifi, or speaker sound. The procedure I used was a new install on USB using f2stickpup, then replacing vmlinuz and fdrv_fossapup64_9.6.sfs with the two new files, and renaming them to what was removed. I left everything else as stock new install.

Umm.....what about the zdrv? If you've replaced vmlinuz, then this is a must.....because that's where the modules live that are compiled specifically to work with that kernel release version. The frdv is for firmware, and CAN in some cases be optional.....but there's nothing optional about the zdrv, unless my understanding of the kernel-swap procedure in these more modern Puppies is all out of whack!

Since I've performed many successful kernel-swaps over the last few years, I don't think I'm SO far out....

(*shrug...*)

Again, just a thought, FWIW.

Mike.

Mike there was no zdrv specifically in the 6.8 kernel pkg. The zdrv in the installation is just the one that came packaged in F96-CE 4

Maybe I did it wrong, and should have replaced zdrv not fdrv with what was supplied : kernel-modules.sfs-6.8.1-64oz-ao-LOLAT?

[rockedge]

Maybe I did it wrong, and should have replaced zdrv not fdrv with what was supplied : kernel-modules.sfs-6.8.1-64oz-ao-LOLAT?

Yes!!! During every swap of a kernel the most important components are the vmlinuz and zdrv_fossapup64_9.6.sfs that are changed. It is important to use the zdrv that matches the kernel version vmlinuz.

The fdrv is more interchangeable as long as the name matches the zdrv, like fdrv_fossapup64_9.6.sfs.

Rename kernel-modules.sfs-6.8.1-64oz-ao-LOLAT to the same exact name of the zdrv, in this case zdrv_fossapup64_9.6.sfs

This is what the change_kernels script does. Rename and copies the huge kernel components into the frugal installation after renaming the original.

Try in a terminal change_kernels or Menu->Setup->quickpet fossa for automated kernel swapping.

[vtpup]

Made a brand new installation again from ISO on thumbdrive and this time replaced zdrv (and renamed as indicated) and surprisingly there was no change when I tested the functions.

[rockedge]

@vtpup So the devices are still not working and is there a delay in the boot cycle?

Might be a fdrv issuel......I can make a Void Linux borrowed fdrv (200 M) to try out. I am starting to think the firmware is the problem now.

[vtpup]

Rockedge, just to be absolutely sure, I did a new download of F96-CE 4 ISO, then did an f2stickpup new installation. Rebooted to that. Then I did the Quickpet method and pointed to a fresh download of huge-6.8.1-ubun64oz-ao-LOLAT.tar.bz2 .

When I started up after the kernel change, touchpad wasn't working, wifi wasn't working, and the speaker wasn't working. This was also the same as before the kernel change.

If I remember correctly, I think Bookworm 10.0.4 might have had touchpad and wifi working OOTB, if that's any help. I'm going to check that to be sure with a fresh install of that to USB right now.

[vtpup]

Yes, Bookworm (now version 10.0.6) in a fresh install to USB drive does recognize both touchlpad and wifi OOTB. Kernel is, I believe 6.1.76. Speaker sound does not work, and, as I said earlier, I believe kernel support for that was supposed to arrive after 6.8.

Okay back to the original topic which was the error messages on boot-up on this laptop w F-96-CE and the 6.1.8 Fossa Huge kernel:

The error message is now gone. The only difference I'm aware of is, instead of removing intel_ishtp_hid entry from the file fossapup64initmodules.txt, that file itself has been removed altogether. Touchpad function is still present, though still subject to glitches at certain times. But the error messages at boot are definitely gone now. So I guess this issue is solved, sort of.

I hate to give up on getting the touchpad functioning well, and possible solutions to that with a later kernel -- mods, do you want me to mark this solved?

[rockedge]

mods, do you want me to mark this solved?

would be better if it was like 90% or better to say solved. Does Bookworm have a fdrv to borrow?

[vtpup]

Yes it has a,b,f,and zdrvs. It's possible to work F96 with a borrowed fdrv?

Later edit:

I tired using the fdrv from Bookworm 10.0.6 in a new instance of F96-CE v4 using the original kernel and the suggested huge kernel (6.8.1) and there was no change. Touchpad, wifi, and speaker were not working.

I did notice that Fossa Quickpet Change Kernel mechanism didn't seem to work. When booting the kernel version is visible, and although I had specified the new kernel, the old kernel number was displayed after reboot. Also the old kernel version was visible in Conky.

Method used: I dragged and dropped huge-6.8.1-ubun64oz-ao-LOLAT.tar.bz2 to the entry box on page 1 of the Change Kernel app. and hit OK. The app reported it was unpacking the kernel, and then did nothing more.

So after that failure, I also did the manual method of replacing vmlinuz and zdrv, and renaming them to suit fossapup64, using the same exact filenames they replaced. And added the fdrv from Bookworm. As I said, no improvement.

Some points of interest. If you look on page 2 (Kernel Files tab) of the Change Kernel app It claims that firmware goes in zdrv and modules go in fdrv. This is why I had originally put the modules in fdrv (going back to Mike Walsh's post above). So I'm still confused about these things. Is the Change Kernel app working correctly?

[ozsouth]

@vtpup - you can't use bookwormpup's fdrv without conversion - it is usrmerge format, f96ce4 isn't.
Seems your current issue is firmware, because bookworm worked.
You could try the fdrv linked below (59mb, rename to same as your current f96ce4's filename), or check your pm in about half an hour for a converted (to non-usrmerge) version of bookworm's fdrv (93mb).

Later: 59mb fdrv didn't help.

[vtpup]

thanks ozsouth, will do.

note my edits above btw re problems with change kernels app, and confusion over where to put modules (zdrv or fdrv)

[ozsouth]

@vtpup

Alternatively, you could keep 6.1.8-fossapup64 kernel, removing fossapup64initmodules.txt, and:

- then add to the linux/kernel line in grub/syslinux cfg: modprobe.blacklist=intel_ishtp_hid
- then make a script to load it after bootup, like this:

#!/bin/sh
modprobe intel_ishtp_hid

Script is composed in texteditor, saved (as start_intel_ishtp_hid) in /etc/init.d , then made executable - in terminal:
chmod 755 /etc/init.d/start_intel_ishtp_hid

Next bootup, module would be stopped from loading at boot, then loaded after boot.
Note that could still cause a hangup for some rogue modules.

[mikewalsh]

@vtpup :-

AFAIK, Steve - and from what I've gathered over the last couple of years - fdrvs are kernel-agnostic. In other words, they're NOT tied to any specific kernel compilation......because if that were the case, gyrog wouldn't have been able to release firmware SFS packages like he does...

viewtopic.php?t=2302

(You might get some joy from this thread.....and there again, if your hardware IS that new, you might still have to be patient and wait a while longer. This is the perpetual snag with brand-new hardware under Linux; unlike Windoze, where manufacturers go out of their way to ensure drivers are available for every new item the instant it hits the shelves, we DO have to wait until the kernel gets around to supporting it......which can be anywhere from a few weeks to several months, depending on how much free time the person who maintains that particular bit of the kernel is able to devote to it.)

Mike.

[vtpup]

Ozsouth, tried your first linked fdrv and no change in capabilities.

I believe that to successfully run the wifi card the firmware must have present the intel iwlwifi-so-a0-gf-a0-72.ucode and iwlwifi-so-a0-gf-a0-71.ucode pieces, since those were mentioned as loading in my logs when wifi was successful. I looked inside the fdrvs so far and have not seen those included in lib firmware.

I'm not sure on the synaptics touchpad driver yet. here's the device type..

Code: Select all

# xinput list
snip....
⎜   ↳ SYNA32E6:00 06CB:CE7D Mouse             	id=10	[slave  pointer  (2)]
⎜   ↳ SYNA32E6:00 06CB:CE7D Touchpad          	id=11	[slave  pointer  (2)]
snip.....

kernel messages for touchpad when succesfully initialized:

Code: Select all

Apr  3 22:27:08 puppypc32144 user.info kernel: [   12.139508] input: SYNA32E6:00 06CB:CE7D Mouse as /devices/pci0000:00/0000:00:15.0/i2c_designware.0/i2c-9/i2c-SYNA32E6:00/0018:06CB:CE7D.0006/input/input10
Apr  3 22:27:08 puppypc32144 user.info kernel: [   12.139644] input: SYNA32E6:00 06CB:CE7D Touchpad as /devices/pci0000:00/0000:00:15.0/i2c_designware.0/i2c-9/i2c-SYNA32E6:00/0018:06CB:CE7D.0006/input/input11
Apr  3 22:27:08 puppypc32144 user.info kernel: [   12.139712] hid-generic 0018:06CB:CE7D.0006: input,hidraw3: I2C HID v1.00 Mouse [SYNA32E6:00 06CB:CE7D] on i2c-SYNA32E6:00

[ozsouth]

@vtpup the ...72.ucode firmware file is in the fdrv I pm'd you - so wifi should work.

@mikewalsh - sadly, this usrmerge thing affects fdrv's, as well as kernels & kernel sources. Also, bookwormpup's fdrv has a bunch of microcode stuff included, which may not play well with other pups. I removed all that in the rebuild I pm'd him.

[vtpup]

Thanks ozsouth, I tried the new fdrv out. It nearly ran wifi (identified the wrong interface). I noticed that the 71 iwlwifi wasn't present and I had one, so added it and instantly wifi worked.

The touchpad didn't work so I did take a shot at modprobing designware I2s module, but that didn't fix it. Even after a reboot.