kernel-generic 5.12.16 built on the affected machine allows to use all its 6 USB ports for input devices (at least keyboard and mouse). I plan to upload it today to the main Slint repository with its siblings: kernel-modules, kernel-headers and kernel-source.

so we can deduce that the issue probably appeared with commits of 5.12.17, right ?

EDIT :
I think we will remain in the dark until we have a trace of what's going on at the boot

Yes, but... I tried openSUSE Tumbleweed live KDE with 5.13.4 and all my USB ports worked. The difference in behavior could be due to the configuration and/or to a patched kernel. Will build a vanilla 5.13.5 with their config to check.

You can diff these files ;-)

The Opensuse one came from :
https://download.opensuse.org/reposi...1.2.x86_64.rpm

I will, but only if my attempt is successful, which I should know in about one hour ;)

PS I copied the config file from the live system.

Build failed. It looks seems that openSUSE patches their kernel, so their config file doesn't match a vanilla one. Now trying to build from the default Arch config file. If that doesn't succeed, I'll have at least learnt something today.

They also provide a vanilla one

https://download.opensuse.org/reposi...1.2.x86_64.rpm

Thanks for the info. I'll first see what I come up with the Arch config.

Mr. Didier, I suggest you to look for the openSUSE Tumbleweed sources.

https://download.opensuse.org/source.../repo/oss/src/

https://download.opensuse.org/source....4-1.2.src.rpm

This one will be more or less along with our slackware-current. The other variant of openSUSE is for enterprise or development.

And I suggest you to do your experiments with -current.

Your issues may be produced by the really old GCC of Slackware 14.2.

In other hand, yes. They patch the kernel heavily. For example, those are the openSUSE patches from that .src.rpm

Also, please note that they start with the base release, i.e. 5.13.0 then they patches up the sources, until they arrive to wanted version.

I for one, I suspect that they also skip some of patches from kernel.org.

What I find interesting for your issue is this patch:

Revert-usb-renesas-xhci-Fix-handling-of-unknown-ROM-state.patch
The Renesas chips drives tons of cheap USB 3.0 adapters.

Thanks LuckyCyborg. The build failed too with the config from Arch. The age of gcc is something I also have in mind, although according to the kernel documentation the minimal requirements to compile the kernel mention a minimum version of 4.9 for GCC (withy a caveat: "The gcc version requirements may vary depending on the type of CPU in your computer".

Yes I could experiment on Slackware64-current, but if I am successful there and not in Slint that won't help much.

Which leads to two newbie questions:

• Can I upgrade gcc from 5.5 to the version shipped in -current (10.3.0 at time of writing) without rebuilding or upgrading any other package? Else what should be rebuilt or upgraded (glibc?) ?
• Can I build only the kernel with a new gcc?

In my humble opinion, a heavy upgrade of GCC like you describe may lead in a need for a mass rebuild.

In other hand, regarding building the kernel (and even the external modules) with a dedicated GCC is a question to which response makes me also interested.

I did in the past some experiments with using a x86_64 kernel from Slackware64 under the i586 Slackware. Yes, it's possible, at least in theory.

Thanks for your answers. I will provide 5.12.16 for Slint which is ready for upload for now, then explore my options going further. Maybe someone more skilled than me (not a big challenge, that) will solve the issue for Slackware -current in a way I can apply to Slint, which would be ideal. In any case I am of course ready to test a proposed solution in Slackware64-current.

Yes, you can install a newer gcc in /opt and use it to compile the kernel. But then you also have to remember to compile any kernel modules with the same compiler - that's why it's best to just use the system gcc to compile the kernel.

On my work work computer I have /opt/gcc-{4.9.4,6.3.0,7.3.0,8.4.0,10.2.0} At least a few of the versions are for MATLAB.

Good news
 

With the kernel 5.13.6 released by Patrick used in Slackware64-current as well as the one built for and installed in Slint my mouse and keyboard are usable when connected to any of the six USB ports of my tower.

Chris[1] and anyone having suffered from this issue, please test to confirm that it is also solved for you.

Then I will mark this thread as [SOLVED].

[1]I know it's 3 AM in Brisbane. So do it when you can...

Still having the same behavior here unfortunately.

Applied update to 5.13.6. Rebooted without shutting down to verify no issues - mouse and keyboard worked fine. Then shut down completely and restarted - no response from keyboard and mouse. Hit reset button to restart and mouse and keyboard worked fine again.

The only time things seemed "fixed" for me was with the 5.13.5 upgrade, but that broke with the subsequent -current update even though no kernel update was included and I can't see anything else in the update that might have caused the change in behavior.

It's very odd and seems random with few other folks reporting similar issues. Maybe it's a random hardware problem? I'm just happy I can get things to work easily, but would interested to figure out why this is happening. Curious minds want to know ...

Oh well.

The issue you describe doesn't look like the one reported in this thread as behaviors differ. Still eager to get reports to know if now (with kernel 5.13.6) all USB ports work, from folks for whom it was not the case.

k5.13.6
 

upgraded kernel 5.13.6 and it dropped the (usb) mouse and keyboard during the installation. Had to reinstall 5.13.5 from a live install to recover. Installation is on a usb sdd 500G external drive.

Well, you may be correct that it's not the same issue. It is odd though that we are both having keyboard and mouse USB problems that seem to have started at the same time after a kernel update. I'm not convinced we aren't seeing different symptoms of the same cause. I'm just not technical enough to know where to begin with respect to diagnosing.


I will continue follow this thread and refrain from chiming in so as not to gum up the discussion in case it is truly a separate problem.

Not sure it's the same issue either. What do you mean exactly writing "it dropped the (usb) mouse and keyboard during installation?" Do you really mean that you lost the keyboard and mouse before rebooting? In this case this is indeed another issue as the new kernel won't be used before you reboot. Or do you mean that you lost the keyboard and mouse after reboot? Then it could be because the package kernel-modules-5.13.6 has not been installed. How did you install the 5.13.6 packages: manually or using slackpkg?

Yes, good news. The update to 5.13.6 has fixed the problem for me.

I also tried the warm/cold boot scenario but it made no difference here - all USB2 and 3 ports now work correctly in both cases.

chris

Thanks for the report. I mark this thread as [SOLVED].

I really lost the keyboard and mouse before rebooting. In fact, it was while slackpkg was installing the kernel modules package after the kernel was installed. But you're right, it could be an old problem which used to occur (intel_idle.max_cstate) similarly, but coincidentally crashing my box.