Re: [gentoo-user] gentoo-sources-5.10.103 - will not boot
On 13/03/2022 22:26, the...@sys-concept.com wrote: On 3/13/22 14:34, Neil Bothwick wrote: On Sun, 13 Mar 2022 14:04:59 -0600, the...@sys-concept.com wrote: On 3/13/22 13:21, the...@sys-concept.com wrote: Upgraded to: gentoo-sources-5.10.103 and kernel will not boot, not even recovery mode. I'm getting some strange looping/scrolling message on the screen: Kscan: watching read 1 fsk983s I followed standard procedure: emerge -avq =sys-kernel/gentoo-sources-5.10.103 cd /usr/src/ eselect kernel set 3 (this is: linux -> linux-5.10.103-gentoo) cd linux cp ../linux-old_kernel/.config . mount /boot/ make oldconfig make make modules_prepare make modules_install make install grub-mkconfig -o /boot/grub/grub.cfg reboot ( did the same on my other boxes and this kernel is booting OK on other installations) Solved. Please post the solution. Otherwise anyone with a similar problem searching for an answer will find only the question and a tease that it can be fixed but not telling how. Simple human error :-/ When I did: cd linux cp ../linux-old_kernel/.config . mount /boot/ make oldconfig New entries showed up. Instead of pressing "enter" I made a mistake and press "Y" several times. This enabled some feature in the new kernel that shouldn't be there; example: "CONFIG_KCSAN = y" Redoing the process just by hitting "enter" soled the problem; new kernel boot as it should. make olddefconfig will the same as oldconfig while picking the defaults, this will save on hitting enter each time. Steve
Re: [gentoo-user] Any way to run multiple commands from single script in parallel?
On Monday, March 14, 2022 11:51:44 AM CET Björn Fischer wrote: > Hello Joost, > > > Is there a tool/method to execute multiple lines/commands > > simultaneously? Like having 3 or 4 run together and when 1 is > > finished, it will grab the next one in the list? > > probably, GNU Parallel is what you are looking for: > > https://www.gnu.org/software/parallel/parallel.html#examples > > The tool can handle most variants of batch processing scenarios without > the steep learning curve of fully fledged (clustered) job schedulers. Thanks, this seems to do what I need it to do.
Re: [gentoo-user] depclean wants to remove xf86-video-intel
On Mon, 14 Mar 2022 17:07:54 - (UTC), Grant Edwards wrote: > I was a bit startled thos morning when emerge --depclean wanted to > remove xf86-video-intel. I presume this is a result of the switch to > the "built in" modesetting driver? And there are no corresponding Xorg > config changes that need to be made? I bit the bullet, let it depclean and rebooted. The desktp came up as usual but there was a warning in the log about not being able to load the intel module [13.920] (==) Matched intel as autoconfigured driver 0 [13.920] (==) Matched modesetting as autoconfigured driver 1 [13.920] (==) Matched fbdev as autoconfigured driver 2 [13.920] (==) Matched vesa as autoconfigured driver 3 [13.920] (==) Assigned the driver to the xf86ConfigLayout [13.920] (II) LoadModule: "intel" [13.920] (WW) Warning, couldn't open module intel [13.920] (EE) Failed to load module "intel" (module does not exist, 0) [13.920] (II) LoadModule: "modesetting" [13.920] (II) Loading /usr/lib64/xorg/modules/drivers/modesetting_drv.so [13.935] (II) Module modesetting: vendor="X.Org Foundation" [13.935]compiled for 1.21.1.3, module version = 1.21.1 [13.935]Module class: X.Org Video Driver [13.935]ABI class: X.Org Video Driver, version 25.2 [13.935] (II) LoadModule: "fbdev" [13.935] (WW) Warning, couldn't open module fbdev [13.935] (EE) Failed to load module "fbdev" (module does not exist, 0) [13.935] (II) LoadModule: "vesa" [13.935] (II) Loading /usr/lib64/xorg/modules/drivers/vesa_drv.so [13.937] (II) Module vesa: vendor="X.Org Foundation" [13.937]compiled for 1.21.1.3, module version = 2.5.0 [13.937]Module class: X.Org Video Driver [13.937]ABI class: X.Org Video Driver, version 25.2 [13.937] (II) modesetting: Driver for Modesetting Kernel Drivers: kms [13.937] (II) VESA: driver for VESA chipsets: vesa [13.943] (II) modeset(0): using drv /dev/dri/card0 [13.943] (II) modeset(0): Creating default Display subsection in Screen section > > My video chipset is > > 00:02.0 VGA compatible controller: Intel Corporation IvyBridge GT2 > [HD Graphics 4000] (rev 09) I have: 00:02.0 VGA compatible controller: Intel Corporation HD Graphics 5500 (rev 09) > And the only card-selection configuration I've done was to set > VIDEO_CARDS="intel" in make.conf. ditto -- Neil Bothwick Old hitchhikers never die-they just throw in the towel. pgpCSqkSUmeDI.pgp Description: OpenPGP digital signature
RE: [gentoo-user] Any way to run multiple commands from single script in parallel?
If you don't want to do thread management yourself in bash then you can use something like GNU Parallel (in the repo) to handle forking and collating processes for you. Parallel in particular has the additional advantage that it's capable of shipping tasks off to other machines via SSH, so if you get to the point where you need a whole cluster to do your processing it's just a matter of adding a couple of arguments. LMP -Original Message- From: Ramon Fischer Sent: Monday, March 14, 2022 3:37 AM To: gentoo-user@lists.gentoo.org Subject: Re: [gentoo-user] Any way to run multiple commands from single script in parallel? Hello Joost, I suppose, that you are talking about Bash scripts. If so, you may put each individual command in a subshell by using an ampersand ("&") at the end of the line. This example[1] shows it nicely. -Ramon [1] 3. Parallelize running commands by grabbing PIDs.: https://will-keleher.com/posts/5-Useful-Bash-Patterns.html On 14/03/2022 11:13, J. Roeleveld wrote: > Hi, > > I often put multiple commands into a single file/script to be run in sequence. > (each line can be executed individually, there is no dependency) > > Is there a tool/method to execute multiple lines/commands > simultaneously? Like having 3 or 4 run together and when 1 is > finished, it will grab the next one in the list? > > I would prefer this over simply splitting the file as the different > lines/ commands will not take the same amount of time. > > Thanks, > > Joost > > > -- GPG public key: 5983 98DA 5F4D A464 38FD CF87 155B E264 13E6 99BF
[gentoo-user] depclean wants to remove xf86-video-intel
I was a bit startled thos morning when emerge --depclean wanted to remove xf86-video-intel. I presume this is a result of the switch to the "built in" modesetting driver? And there are no corresponding Xorg config changes that need to be made? My video chipset is 00:02.0 VGA compatible controller: Intel Corporation IvyBridge GT2 [HD Graphics 4000] (rev 09) And the only card-selection configuration I've done was to set VIDEO_CARDS="intel" in make.conf. -- Grant
Re: [gentoo-user] Any way to run multiple commands from single script in parallel?
On Mon, 14 Mar 2022 11:13:13 +0100 "J. Roeleveld" wrote: > Hi, > > I often put multiple commands into a single file/script to be run in sequence. > (each line can be executed individually, there is no dependency) > > Is there a tool/method to execute multiple lines/commands simultaneously? Like > having 3 or 4 run together and when 1 is finished, it will grab the next one > in > the list? > > I would prefer this over simply splitting the file as the different lines/ > commands will not take the same amount of time. > > Thanks, > > Joost > > > At the end there's a very rudimentary bash script to do this. I did not do much debugging (probably it fails already if max_jobs>#list_of_jobs). Anyway it's just making use of sending jobs to the background and "communicating" through a FIFO pipe (which you might want to delete at the end). #!/bin/bash list_of_jobs=("sleep 3" "sleep 5" "sleep 1" "sleep 10" "sleep 4") max_jobs=2 my_fifo=/tmp/my_job_fifo write_to_fifo="yes" function run_job () { eval "$@" if [[ $write_to_fifo == "yes" ]]; then echo "Writing to fifo ($@)" echo 1 > ${my_fifo} fi echo Finished job "$@" } function read_and_start_job() { next_job_idx=0 while [[ ${#list_of_jobs[@]} -gt $next_job_idx ]]; do while IFS= read -r line ; do echo "next_job_idx=${next_job_idx} total=${#list_of_jobs[@]}" if [[ $next_job_idx -lt ${#list_of_jobs[@]} ]] ; then job="${list_of_jobs[${next_job_idx}]}" echo "Executing: ${job}" run_job ${job} & let next_job_idx++ else echo "Set write_to_fifo=no" write_to_fifo="no" fi done < ${my_fifo} done write_to_fifo="no" wait } rm -Rf ${my_fifo} mkfifo ${my_fifo} read_and_start_job & while [[ ${max_jobs} -gt 0 ]] ; do let max_jobs-- echo 1 > ${my_fifo} done wait
Re: [gentoo-user] Any way to run multiple commands from single script in parallel?
Hello Joost, Is there a tool/method to execute multiple lines/commands simultaneously? Like having 3 or 4 run together and when 1 is finished, it will grab the next one in the list? probably, GNU Parallel is what you are looking for: https://www.gnu.org/software/parallel/parallel.html#examples The tool can handle most variants of batch processing scenarios without the steep learning curve of fully fledged (clustered) job schedulers. Cheers, Björn
Re: [gentoo-user] Any way to run multiple commands from single script in parallel?
Hello Joost, I suppose, that you are talking about Bash scripts. If so, you may put each individual command in a subshell by using an ampersand ("&") at the end of the line. This example[1] shows it nicely. -Ramon [1] 3. Parallelize running commands by grabbing PIDs.: https://will-keleher.com/posts/5-Useful-Bash-Patterns.html On 14/03/2022 11:13, J. Roeleveld wrote: Hi, I often put multiple commands into a single file/script to be run in sequence. (each line can be executed individually, there is no dependency) Is there a tool/method to execute multiple lines/commands simultaneously? Like having 3 or 4 run together and when 1 is finished, it will grab the next one in the list? I would prefer this over simply splitting the file as the different lines/ commands will not take the same amount of time. Thanks, Joost -- GPG public key: 5983 98DA 5F4D A464 38FD CF87 155B E264 13E6 99BF OpenPGP_signature Description: OpenPGP digital signature
[gentoo-user] Re: gentoo-sources-5.10.103 - will not boot
On 14/03/2022 00:26, the...@sys-concept.com wrote: Simple human error :-/ When I did: cd linux cp ../linux-old_kernel/.config . mount /boot/ make oldconfig New entries showed up. Instead of pressing "enter" I made a mistake and press "Y" several times. This enabled some feature in the new kernel that shouldn't be there; example: "CONFIG_KCSAN = y" Good old contradictory defaults in the kernel config. You press "?" to see the description of an option, and it tells you stuff like "say N unless you know what you're doing," but the default when pressing enter is Y... Go figure.