Of course, my hardware config is too new to work clean.
There’s a strong feeling that, all in all, every new kernel is making a significantly better job than the kernels before.
In fact, kernels before 4.11 shouldn’t be used anymore for hardware like mine.
But there’s a big problem:
Specially cold boot is having more or less retries ending in “kernel panic”.
But there is a lot of unnecessary messages extra slowing down the boot process - and at uptime often one processor is working 100% on junk error messages (mainly endless repetitions of “unexpected IRQ trap at vector 07”), trying to write them to the log file, and writing that this isn’t possible (no logging inititialised - anyway: a logfile would be flooded with junk).
Is it possible to minimize these messages to a tolerable quantity?
What’s the cause for “unexpected IRQ trap at vector 07” messages? (I think, it’s because of the not working Polaris driver - provisorially VGA interface instead of HDMI/DP)
Is it possible to minimize the number of boot tries?
(I’ve seen a 500 kernel panics - 100 times as much than in the last 30 years together)
When restrictions like using a surrogate screen driver are necessary, it would be a good idea to give an unmissable hint in the course of installation together with the successions
I’m using the last ‘stable kernel’, now 4.11.2-1.g03903d8-default.
It seems to work nice - when it’s up.
But it’s having much more boot problems - even at warm boot! - e.g than kernel 4.11.0-3.
Every update to a new kernel is an adventure!
The update to 4.11.2-1-1 … was at first leading to disaster: it was not starting at all even after a lot of retries, all working kernels were purged. The last chance was an “update” from the original install DVD, but then the net wasn’t reachable - the net connection wasn’t initialized, why ever. I had to set them manually, but the router with the addresses wasn’t reachable … in former time I would have had notices on paper …
It’s a good idea to make such notices, e.g. screen dumps of the settings.
I also want to make a Ryzen build, but these messages are not inspiring
Not only people with Linux have problems but also people who use that other (surrogate!) OS
Most times it seems that the memory is the culprit.
Read a lot about memory problems speed/type memory/size of
Do you have these problems with default bios timings?
Newly it’s said(e.g. http://www.nextized.net/2017/04/160 ), an error in kernel module „AMD_PINCTRL“ shall be responsible for „Unexpected IRQ Trap at Vector 07“.
This error is mainly seen on mainboards from Gigabyte using internally GPIO Pins from the mainboard.
My mainboard is indeed from Gigabyte. There are rumors, it’s for steering LED’s on the mainboard, of course a crazy junk feature. If it’s so: shame also on Gigabyte!
I don’t use other than standard EFI/Bios parameters, especially timings.
I think, it’s too early to give a concrete answer, but I’m guessing, that my PC is having no memory problems.
Indeed, I’ve seen different kernels showing up different symptoms looking like different hardware problems, e.g. hardware error messages, disappearing with a later kernel.
I got the message "Unexpected IRQ Trap at Vector 07“ on all kernels, but in a very different way - from only slowing up the boot procedure, making cold boot impossible or nearly impossible; only at boot time or using one processor to 100% for error handling, or at random up to take away up to over 99% of processing power.
Today for some hours I wasn’t able to login here - got no answer to login tries.
Maybe that was because of a lot of „Unexpected IRQ Trap at Vector 07“ messages with Kernel 4.11.2.1.
Only with the install DVD I was able to upgrade to kernel 4.11.2.2, but with this kernel my PC is completely unable to boot.
Now I’m using my ‘fall back system’ - 42.2 with kernel 11.0.3 - of all kernels it’s having the least number of „Unexpected IRQ Trap at Vector 07“.messages.
But when booting was successful and beside of update problems, Tumbleweed with kernel 4.11.2.1 was running more smooth.
The number of „Unexpected IRQ Trap at Vector 07“ seems to be typical for a system version, but also random.
All is possible.
Nearly fifty years ago I had to work on a very sporadic compiler failure. I found an unbelievable heavy error, corrected it - and the compiler was failing completely. I had to correct another unbelievable heavy error. Both errors were there since the first delivery. Two of us were working a day long and didn’t find any hint, why our compiler was able to translate the least program before correcting both errors.
Using Kernel 4.11.3-1.g7262353-vanilla.
Haven’t checked 4.11.3-1.g7262353-default ("“don’t change a running kernel”;))
First fast booting linux on my PC (not taking up to 120 min as before), running mainly stable with up to 16 hard working parallel processes.
A remainig problem: perhaps due to insane unsuccessful mouse clicks 2 times suddenly frozen halts - no reaction to monkey grips ( Ctrl-Alt-F5 … ), no logging.
It’s the very first really usable kernel for my PC.
