Show newer

testing cpu bugs now:

Topology: Quad Core model: AMD GX-412TC SOC bits: 64 type: MCP
family: 16 (22) model-id: 30 (48) stepping: 1 microcode: 7030105
L2 cache: 2048 KiB
Speed: 600 MHz ....
Vulnerabilities: Type: l1tf status: Not affected
Type: meltdown status: Not affected
Type: spec_store_bypass
mitigation: Speculative Store Bypass disabled via prctl and seccomp
Type: spectre_v1 mitigation: __user pointer sanitization
Type: spectre_v2 mitigation: Full AMD retpoline

- feature request from maintainer, request for enhanced --admin -C cpu errata data.

This is a good idea, and should not be too difficult to carry out.

Look for it in inxi 3.0.23. Will give type and status of each cpu issue, like meltdown, spectre_v1 and spectre_v2.

This feature only works with newer Linux kernels.

smxi boosted

Some people give me playful grief because I have a pocket notebook that is never more than 10 feet away from me 24 hours a day and I frequently write down any random random thoughts that occur to me, lists, things I need to remember, interesting quotes, etc etc.

I must say, though, that despite being someone you could call “fully digital” for pretty much everything, starting to carry this notebook a few years ago has been one of the most life-changing oh great decisions I’ve ever made.

3 just came out, so I decided to test the system base feature on it, and discovered that has yet again changed the structure of their /etc/os-release file, adding in some more items.

Rather than play guessing games for proper LMDE system base detection, and other Debian based distro system bases, I decided to just add in a Debian base data generator. This is working in , though it probably won't hit master branch for a few weeks.

smxi boosted

3.0.22 also, more system base added, deepin and blankon, tedious process to match those.

3.0.22 now live. Features -Gxxx compositor version (if found, not many show version data), many more compositors, added more system tray type devices to -Sxxx info item

Fixed a big bug in networking where devices on the same bus id would show the same IF identifier because it wasn't checking the busid in the matching tool. Thanks anonymous debugger data sender!

Added more disk vendor IDs, the number of no name Asian vendors appears to be endless.

New desktop found, , from .

thanks anonymous debugger dataset, found a glitch in IF to Device match in network, in cases where a complex NIC with multiple BusIDs is running (4 in this case), inxi was showing only the first IF for all the cards, now it matches to bus id, vendor id, and chip id, for non arm/mips/sparc/ppc devices, to make sure it gets the right IF address per bus id.

Good timing on that debugger dataset (thanks skynet), just squeezing in that last patch and testing of it before inxi 3.0.22 goes live.

I thought I'd found all the desktops, but apparently tweaking Gnome until it's a newish desktop is more popular than I thought. 's is another one. Added to for upcoming new version, along with a bunch of small goodies, like more docks and system tray items for -Sxxx info: item.

dev branch , no significant problems beyond very little data with tests, retested device, things back to working again after a few small breaks in the soc device parser caused by fixes.

All in all about as easy to implement and extend support to new untested architectures as I'd hoped when developing the initial ARM logic, easier than I expected, in fact.

All help is appreciated if anyone has these types of devices, or can give access to one.

I came across some qemu images that are built to run on x86 for , , , and . I fired up some, and now have a rough PPC running in pinxi. This is in addition to ARM and MIPS which were already supported to some degree. Will test on SPARC soon as well, and work those kinks out.

Note that virtualized ARM, MIPS, PPC, and SPARC tests are no substitute for testing on the real hardware, so anyone who wants to donate either datasets from latest pinxi, or give ssh access, would help.

in dev version pinxi, fixing issue with MATE desktop version ambiguity, someone in the project is neglecting to actually consistently update mate-about --version to reflect, well, the actual version.

New fix is to take version from mate-about and mate-session and pick the larger/higher one, and show that, which makes it dynamic. I wish fixes like this were not needed, but they are..

Added a few more truly obscure window managers as well.

Thanks gm10 for the MATE version glitching issue report.

update. Added 396.54 driver, and extended Linux kernel tests to 4.18 for nvidia.

After giving it some thought, I decided to give i3 tiling wm a real try. So far, it seems particularly well suited for laptops, particularly wide screen laptops, which I find annoying in that if you work with code or text, the wide screen is too wide, but half width is just about right for consoles or code editors. These keyboard driven window managers are not for everyone, but if your workflow involves really knowing roughly what you want to do and it's consistent, i3wm is nice. Easy to configure too.

inxi 3.0.21 turned out to be an epic release.

New stuff: Moksha desktop, qtile wm added. USB sub system fully refactored, added usb drivers, expanded usb types, added serial numbers and interfaces for -xxx. Added Xvesa and tce package manager/repo support (for tinycore).

Fixed issues with busybox ps, now doesn't spray errors if busybox ps, which means all systems running busybox for tools will now 'just work'. Dynamically tests and assigns proper columns now.

Added a few more distro / base ids.

Added tce repo data for tinycore, this new inxi is getting more features than I expected! The installed version required adding perl5 and pci-utilz tce packages, plus the wget.tcz package to get https support.

I may approach tinycore to see if they want to package inxi, it seems to now work fine with almost no further programs required beyond perl and lspci, both of which are in the big tinycore live version, but not on the installed version as far as I could tell.

Prompted by a recent @distrotube on tinycore, I decided to give it a spin to see how well inxi runs on it.

Answer: just a few bug fixes, (tinycore uses something like busybox, ps aux in this case only output to 3, not 11 or more, columns), tweaking the @ps_aux / @ps_cmd builders, and the --processes item, to handle varying column counts for ps output. No cpu/mem use data since it's not there. Even flwm / wbar detection working.

This should fix all busybox based systems so they run without errors.

I decided to refactor the entire USB log and output handlers, now it's no longer using the hacks I put in for lsusb -v parsing, and runs only clean arrays, predictable indexes, so I don't have to loop through them any more to find the right parts in each USB items. Debugged and working on audio, network, and main USB feature. I'll have to delay inxi 3.0.21 until tomorrow to make sure there aren't any obvious bugs or oversights. Tested on 8 year old system, found a few changes required, all seems good now.

Very close to ready, AntiX guys as usual are helpful and exposed a few corner cases I had not handled. Now using USB type data from /sys not lsusb always, it's more reliable because I can modify it or hack it as required. Hub ports count extended to BSD usbdevs parser, and BSD logic adjusted to fit new logic. Probably will do 3.0.21 tomorrow if no new corner cases appear. Since this logic produces much better results than the lsusb only version, it's easy to dump the old methods.

Show older
Mastodon for Tech Folks

This Mastodon instance is for people interested in technology. Discussions aren't limited to technology, because tech folks shouldn't be limited to technology either!