inxi/pinxi + --recommends and Slackware package names, help complete list
 

I decided to finally add slackpkg / installpkg type package names to the --recommends. I also added Zenwalk's netpkg, having just learned that there was a 2022 zenwalk release after long dormancy, so I figured why not?

You can check these if you have pinxi installed (pinxi -U) or if not:
I used mainly https://slackbuilds.org and https://slackware.pkgs.org to find the package names.

I also took this as an excuse to refactor the CheckRecommends feature, which was well past overdue.

I think I've found most of the slackware package names, but there's a few I either can't find, or which I'm not clear about (currently in pinxi shows '?' for ones I am not sure about:

Programs:
modinfo

I found references to modinfo in module-init-tools in earlier slackware package pools, but nothing I think past 13.0. modinfo now appears to belong to kmod package. Is this correct? Was that previously in module-init-tools?

Perl Modules:
Note that JSON::PP entered Core Modules at 5.014, that one is only really relevant to get the full module package name listing like inxi has for apt, pacman, rpm. Or for very old Slackwares, which would probably have had that packaged separately, but I can't find that name.

Also note that if I assign the value '', then pinxi/inxi will show 'N/A', which would be the correct result if the package is not available in any reasonably standard slackware/slackpkg repository.

There is also a big USB upgrade, which provides interesting information for USB 3 and 4. The USB upgrade fixed a lot of things, and added some:

I have no access to USB 4 data currently, apparently it's trickling out now, a manjaro user's laptop shipped with theoretical ryzen support for usb 4 over pcie, but he was getting only USB 1 speeds because Asus or AMD have not shipped the firmware update yet to enable the PCIe based USB4 lanes, which is when I decided that a USB enhancement was probably in order, since it was and is not easy to explain what the USB bus is really doing.

This example is interesting because you can see the parent hub on bus 6-0 is USB 3.0 at 5 Gbps, 1 lane, but the child hub is actually capable of USB 3.2, 2 lanes, and 10-20Gbps, depending, so it's working in a slowed mode. I just realized that right now, I've been working with the USB stuff so much I didn't realize that the data was actually explaining a weird behavior I'd never understood before.

Note that USB 3.2 and 4 only get all their lanes (which work like PCIe lanes, 1 up, 1 down channel per lane, usually, except with USB 4 120 Gbps) with a USB C connector.

Very nice h2-1 !

I've got a Discrete Thunderbolt 4 Chip in my Laptop that still doesn't work quite right.

Below is some data for you.

I've got a USB Mouse and an Anker Thundrbolt 4 Hub Plugged Directly into the Laptop.

Then there is an external NVMe Enclosure and a Thumb Drive in the Anker

HTH

-- kjh

Indeed.

If you use https://packages.slackware.com/ you can search by package and by content.


you can also use to find the name of the package that contains a file.

All updates to docs and pinxi now live in 3.3.26-21

Slax-Dude, thanks for confirming. I found the slackpkg file-search [file] last night when trying to complete the last slackware names. Obviously would have been much faster if I'd found it earlier, that's convenient. I was using an old distrowatch slackware package manager command table, but it's out of date significantly, and did not have that option for slackpkg. I'm updating my docs for this now as well.

I view a nice side-effect of working with this slackware package manager stuff that it familiarizes me with the tools.

https://packages.slackware.com/ is a good one, I started a docs/inxi-repo-pakcage-manager.txt doc, so I'll update that with this info.

As an aside, I noticed this thread on slpkg https://www.linuxquestions.org/quest...er-4175720937/ and decided to read through it to see if I'd missed anything related to slpkg, and found a slackware type repo file that inxi does not yet support, the slpkg repo file, there was a sample in the thread so I can probably add support for that syntax in --repos.

As a further surprise, I found that zenwalk has released a new version, 15.0, after I believe a very long period of dormancy, so I added support for netpkg and it's repo file format, and for zenwalk system base feature, that was easy. I thought ZenWalk was long gone, but it's apparently still alive and kicking.

kjhambrick, you managed to find an unsupported mode, I thought I'd found all possible variants, thanks.

I thought only USB 3.2 and greater supported 10 Gbps lanes. However, the 'mode:' is the official usb consortium terminology, but this combination was not listed anywhere I researched. Mode is basically a function of rev, lanes, and speed. To call the USB consortium explanations of this... difficult, opaque, etc, is kind. I also added another possible mode, for rev 3.2 at 40Gbps, 2 lanes, which would be mode: 3.2 gen-3x2. That should never appear, but neither should yours.

https://en.wikipedia.org/wiki/USB4

Do you know the actual USB revision of that hub? As noted, there does appear to be firmware issues for USB 4, do you know which of the hubs/devices are supposed to be USB 4? The thunderbolt 4 is usb 4, but appears to be connected to a USB 2 hub, though that's how the other system I saw that had USB4 over pcie showed itself without the firmware upgrade, which must be coming soon, it's odd they shipped these laptops without the support. My understanding is that Linux kernel has had USB 4 support for quite a while now, somewhere in the mid 5.x series if I remember right..

One last package remaining, this one doesn't show up with any type of search:

Perl module: XML::Dumper

I assume if for example: slackpkg file-search XML/Dumper returns nothing, that package doesn't exist.

I tested that syntax with other modules, like Data::Dumper: slackpkg file-search Data/Dumper and it correctly returned core Perl, since it's a core module.

I believe its' safe to assume this module is not packaged? At least not in slackware, slackpkgplus, or slackonly.

h2-1 --

I wouldn't be surprised if this isn't a Kernel Bug.

I have been running a Kernel Patch for months now hoping to get my Discrete Thunderbolt 4 Ports to work.

This is the Kernel Bugzilla Thread: https://bugzilla.kernel.org/show_bug.cgi?id=214259

Thunderbolt 4 still does not work correctly on my Laptop but at least Mika's most recent patch eliminated the scary long hangs at boot when my Discrete Thunderbolt Chip was enabled in the Insyde BIOS :)

The Hub is this: Anker Docking Station, 777 Thunderbolt Docking Station (Apex, 12-in-1, Thunderbolt 4)

Specs are on that Page.

HTH

-- kjh

officially my laptop should have one thunderbolt 4.
pinxi say 2 one thunderbolt and one thunderbolt 3 (if i m reading right the output)
since I never look for it... now i look it more close and took a photo.
I really dont understand ... that symbols....

Edit. Lapto papers said

This is corrected in pinxi 3.3.26-21 (pinxi -U to get new version)

rizitis, thanks for showing another sample of that exact mode issue.

It will now show the correct mode, which is gen-2x2, aka, 2 lanes at 10Gbps/lane.

I'll check back later.

The research I did came up with AMD saying that the support for usb4/pcie was in the ryzen 6xxx cpu, but it needed some kind of firmware update, which had not been released as of last summer.

I did not find more on it, but that's very recent relatively speaking, and I also don't know when the update was actually released, or if it's a kernel bug, I didn't find much on it.

pinxi 3.3.26-22 fixes I think the last set of errors:

was the wrong mode, that's actually gen-2x1.

I simplified the usb 3 rev mode string generator to use only speeds and lanes, and also to support case where system reports a true usb 4 device/hub as 3.2, which would be speeds of 40, 80 Gbps, which have to be usb 4 aka thunderbird 4. There's also a 3 lane 120 Gbps in the works, that's two dedicated transmit lanes, and I think 1 dedicated receive lanes.

Note that in that instance, you'd see inxi show:

I think that's what it would show anyway, haven't seen a real data set for how that is seen by linux in /sys. 1 lane is usually 1 tx (transmit) and 1 rx (receive), similar to PCIe lanes. At least that's how I understand it.

I believe usb 1, 2 gets sent down a separate wire, which then is further sent to the nearest usb 2 hub, not the usb 3 hub it's actually inserted into. You can verify this by checking the bus IDs of the hubs/devices when you insert usb 3, 4, or 1, 2, into a usb 3 or 4 hub.

I believe the effective mode is always determined by the slowest link in the chain of hubs and devices on that bus.

This also explains why you will see a usb 2 or 1 device inserted into a usb 3 or 4 hub/port show up either not at all, or on another usb 2 hub on the system, which was something I'd been wondering about.

I found these two helpful to understand conceptually how the wiring works.

Very good color coded usb connector types diagrams:
https://fabiensanglard.net/usbcheat/index.html

Good usb c connector and usb 3, 4 info:
https://tripplite.eaton.com/products...ypes-standards

h2-1 --

Below is the USB / ThunderBolt for my Laptop with the latest pinxi

-- kjh

Hmm, maybe I'll mull over this a bit.

I believe technically the 'mode' for USB 1 and 2 is 'Legacy', though I only saw that in one place.

I wasn't originally going to show 'lanes:' for USB 1, 2, but accidentally enabled that during a change, and decided I liked it, but there's an issue, if you follow the color code link, it notes that 4 wire usb is only half duplex, unlike 8 wire. And it says USB 3.2 and 4 only work at full speed with USB C connector type, since that has I think 4 tx and 4 rx channels. It's confusing.

this is my output:

very good! thank you!
the only mistake is thisbut its not mistake of pinxi...
just for information:
if ponce=true in toml then
sbo=false by default from slpkg
else it will break system...
if you can fix that then its 100% ok!

The safest way here was to do all the blocks, then grep the array for start with 'ponce ~', and if yes, grep it again for not start with 'sbo ~'. This makes the position of the items irrelevant since it runs after the array of repos is built.

This seems fairly reliable, and not too fragile, assuming it doesn't change.

Note also that pinxi is using the prefix for the variable names to determine when a new block starts, not the empty lines, that's safer long term, since it looks like those are just setting some internal variable names, and the empty lines don't necessarily matter so it's ignoring comment lines and empty lines and lines starting with [...

Thanks for explaining it, I can see why slpkg would automatically disable one if the other is true since it would be damaging to the system, easy user mistake to make so makes sense to protect against it.

This is running now in pinxi 3.3.26-26

ponce false:
ponce true:

Ok!
I can give you some more work :D
If you want to support sbopkg in pinxi
then pinxi must search if exist
and if exist == and not empty; then
print thsese 2 lines
elif if not exist or if exist but empty
then search in
and print these
else
No active sbopkg

sbopkg first looking to /root/.sbopkg.conf to see which of the 2 SlackBuilds repo is active
if /root/.sbopkg.conf not exist or if its empty the it looking in
REPO_NAME= (should be SBo or SBo-git)
REPO_BRANCH=(15.0 or current)

:)