Re: [gentoo-user] [SOLVED] Kernel build failing on new install
On Tue, Oct 13, 2020 at 06:06:51PM +0100, Peter Humphrey wrote > On Tuesday, 13 October 2020 15:13:34 BST Walter Dnes wrote: > > > It seems that deleting stuff in "make menuconfig" is 90% of the work in a > > kernel config. All etherenet drivers are default enabled. What bugs me is > > that with every kernel upgrade, "make oldconfig" shows a whole bunch of new > > ethernet drivers, and they're all default enabled. > > You can avoid that by copying your .config file from the previous version. That's what "make oldconfig" is about. Copy over the previous kernel .config and use that as a starting point. Any additional ethernet drivers in the new kernel are defaulted to enabled. -- Walter Dnes I don't run "desktop environments"; I run useful applications
Re: [gentoo-user] tried desktop profile
On Tuesday, October 13, 2020 6:52:05 PM CEST Jude DaShiell wrote: > Let's see if this is a correct bottom post. I've never seen anything in > this life. Eyes never developed enough for me to see. > The desktop profile emerge failed, and no more has been done with the old > machine yet. I could change to basic profile then emerge @world then > emerge the basic profile @world then emerge --depclean to clear the > unnecessary desktop packages from the machine. If I'm not mistaken, the > new machine has 8 cores on it but all the ssd disks I have for it are > 120MB disks so that might be a bit small for gentoo. I'll read through > the hardware requirements again and make sure. I really hope you mistyped and meant those SSDs are 120GB (Gigabyte, instead of Megabyte) 120GB is enough. I have an older laptop with that and it runs KDE/Plasma just fine and even manages some games from steam. Most important is the memory requirement. 2GB Ram can be made to work, but if you want a desktop, you need to forget about the likes of KDE and Gnome and go with something that needs less memory. I have no idea how well they support accessibility tools like speakup though. Hope someone else with more knowledge about this can add their input. Kind regards, Joost
Re: [gentoo-user] X Forwarding from virtual host
On Wednesday, October 14, 2020 4:20:46 AM CEST Dan Egli wrote: > Okay, this is I HOPE a simple enough question. I have a virtual server > running on my Win10 Host (not my ideal O/S!) that has a full X > environment on it. I usually connect via Putty(ssh) using VirtualBox's > Host Only network. That's great for text, but how do I set things up so > that I can run X programs on the virtual box and have them show on my > Win host? I have an implimentation of X for Windows (Xming)running, and > I set putty to forward X connections, but when I try something as silly > as xeyes, it fails. I've notice that the DISPLAY environment isn't being > set, but setting it myself doesn't seem to help. The Virtual Server's IP > is 192.168.56.25 and the Host automatically gets .1, so I tried setting > DISPLAY=129.168.56.1:0 and it doesn't work. I get a message "No protocol > specified" followed by the error "Error: Can't open display: > 192.168.56.1:0.0" > > Putty is set to forward X connections, and uses the same destination. > What am I doing wrong? http://www.geo.mtu.edu/geoschem/docs/putty_install.html
[gentoo-user] X Forwarding from virtual host
Okay, this is I HOPE a simple enough question. I have a virtual server running on my Win10 Host (not my ideal O/S!) that has a full X environment on it. I usually connect via Putty(ssh) using VirtualBox's Host Only network. That's great for text, but how do I set things up so that I can run X programs on the virtual box and have them show on my Win host? I have an implimentation of X for Windows (Xming)running, and I set putty to forward X connections, but when I try something as silly as xeyes, it fails. I've notice that the DISPLAY environment isn't being set, but setting it myself doesn't seem to help. The Virtual Server's IP is 192.168.56.25 and the Host automatically gets .1, so I tried setting DISPLAY=129.168.56.1:0 and it doesn't work. I get a message "No protocol specified" followed by the error "Error: Can't open display: 192.168.56.1:0.0" Putty is set to forward X connections, and uses the same destination. What am I doing wrong? -- Dan Egli On my Test server OpenPGP_0xF8A7B3F2AAB08F9D.asc Description: application/pgp-keys OpenPGP_signature Description: OpenPGP digital signature
[gentoo-user] Re: [SOLVED] Kernel build failing on new install
On 2020-10-13, Neil Bothwick wrote: > On Tue, 13 Oct 2020 10:13:34 -0400, Walter Dnes wrote: > >> > Try disabling CONFIG_IKHEADERS in your kernel config. >> > I have it disabled on my system. >> >> Thanks. That fixed my problem. It seems that deleting stuff in >> "make menuconfig" is 90% of the work in a kernel config. I can't even remember the last time I configured a kernel from scratch. I always start with a .config file from an older version or a different machine. >> All etherenet drivers are default enabled. What bugs me is that >> with every kernel upgrade, "make oldconfig" shows a whole bunch of >> new ethernet drivers, and they're all default enabled. > > AFAIR that option only enables the brand of card, the individual models > still have to be selected. That's how it has always worked for me. By default it will _show_ you new drivers, but they're disabled by default. If you're paying attention, you can disable the a whole submenu full of new-but-disabled drivers and save a bunch of time. > But it is annoying when configuring a new version of a working kernel. Yep. -- Grant
Re: [gentoo-user] tried desktop profile
On 13/10/2020 17:52, Jude DaShiell wrote: Let's see if this is a correct bottom post. I've never seen anything in this life. Eyes never developed enough for me to see. Perfect, great! Cheers, Wol
Re: [gentoo-user] [SOLVED] Kernel build failing on new install
On Tue, Oct 13, 2020 at 06:06:51PM +0100, Peter Humphrey wrote: > On Tuesday, 13 October 2020 15:13:34 BST Walter Dnes wrote: > > > It seems that deleting stuff in "make menuconfig" is 90% of the work in a > > kernel config. All etherenet drivers are default enabled. What bugs me is > > that with every kernel upgrade, "make oldconfig" shows a whole bunch of new > > ethernet drivers, and they're all default enabled. > > You can avoid that by copying your .config file from the previous version. Even then, running a `make syncconfig` on your old .config in a new kernel sources directory will ask all the same questions regarding new features. Some defaults are sensible, some are not. -- Ashley Dixon suugaku.co.uk 2A9A 4117 DA96 D18A 8A7B B0D2 A30E BF25 F290 A8AA signature.asc Description: PGP signature
Re: [gentoo-user] [SOLVED] Kernel build failing on new install
On Tuesday, 13 October 2020 15:13:34 BST Walter Dnes wrote: > It seems that deleting stuff in "make menuconfig" is 90% of the work in a > kernel config. All etherenet drivers are default enabled. What bugs me is > that with every kernel upgrade, "make oldconfig" shows a whole bunch of new > ethernet drivers, and they're all default enabled. You can avoid that by copying your .config file from the previous version. -- Regards, Peter.
Re: [gentoo-user] Re: [gentoo-user] UEFI booting again
On Monday, 12 October 2020 17:43:04 BST Michael wrote: > The UEFI firmware contains a number of variables in key/value pairs, stored > on NVRAM. One of these is a table containing a Boot Menu within an > editable area of the firmware, which can be manipulated with the EFI shell > (efibootmgr) to set, rename, delete bootable .efi images. [...etc] Thanks Michael; more-or-less what I thought. -- Regards, Peter.
Re: [gentoo-user] [SOLVED] Kernel build failing on new install
On Tue, 13 Oct 2020 10:13:34 -0400, Walter Dnes wrote: > > Try disabling CONFIG_IKHEADERS in your kernel config. > > I have it disabled on my system. > > Thanks. That fixed my problem. It seems that deleting stuff in "make > menuconfig" is 90% of the work in a kernel config. All etherenet > drivers are default enabled. What bugs me is that with every kernel > upgrade, "make oldconfig" shows a whole bunch of new ethernet drivers, > and they're all default enabled. AFAIR that option only enables the brand of card, the individual models still have to be selected. But it is annoying when configuring a new version of a working kernel. -- Neil Bothwick Nothing is illegal if one hundred businessmen decide to do it. pgpyNQPcR7w4z.pgp Description: OpenPGP digital signature
Re: [gentoo-user] tried desktop profile
On Tue, 13 Oct 2020, Wols Lists wrote: > Date: Tue, 13 Oct 2020 12:40:23 > From: Wols Lists > Reply-To: gentoo-user@lists.gentoo.org > To: gentoo-user@lists.gentoo.org > Subject: Re: [gentoo-user] tried desktop profile > > On 13/10/20 16:37, Jude DaShiell wrote: > > Since the profile without the desktop emerged @world correctly after the > > make.conf file got adjusted correctly for this computer I think all it > > will run will be the basic profile. Even if it were possible to emerge > > the compiles necessary for a desktop and get the rest of the system built > > I'd probably run into difficulties later. I'll do the basic profile on > > this machine and maybe do a desktop on the newer machine at some later > > date. > > No, you can add any programs you want and it'll just pull in what you need. > > HOWEVER, now you've fixed it for the basic profile, I'd just change that > to the desktop profile and re-emerge, and it'll pull in all the new > stuff to go to the desktop profile. > > NB - you're still putting your post below the .sig marker. If you can't > see clearly, we accept that it's difficult, but it helps us if you can > get it right (like with language and foreigners, there's a big > difference between people who have difficulty doing the right thing, and > people who can't be bothered to do the right thing ...) Let's see if this is a correct bottom post. I've never seen anything in this life. Eyes never developed enough for me to see. The desktop profile emerge failed, and no more has been done with the old machine yet. I could change to basic profile then emerge @world then emerge the basic profile @world then emerge --depclean to clear the unnecessary desktop packages from the machine. If I'm not mistaken, the new machine has 8 cores on it but all the ssd disks I have for it are 120MB disks so that might be a bit small for gentoo. I'll read through the hardware requirements again and make sure. > > Cheers, > Wol > > --
Re: [gentoo-user] tried desktop profile
On 13/10/20 16:37, Jude DaShiell wrote: > Since the profile without the desktop emerged @world correctly after the > make.conf file got adjusted correctly for this computer I think all it > will run will be the basic profile. Even if it were possible to emerge > the compiles necessary for a desktop and get the rest of the system built > I'd probably run into difficulties later. I'll do the basic profile on > this machine and maybe do a desktop on the newer machine at some later > date. No, you can add any programs you want and it'll just pull in what you need. HOWEVER, now you've fixed it for the basic profile, I'd just change that to the desktop profile and re-emerge, and it'll pull in all the new stuff to go to the desktop profile. NB - you're still putting your post below the .sig marker. If you can't see clearly, we accept that it's difficult, but it helps us if you can get it right (like with language and foreigners, there's a big difference between people who have difficulty doing the right thing, and people who can't be bothered to do the right thing ...) Cheers, Wol
Re: [gentoo-user] tried desktop profile
On Tue, 13 Oct 2020, Michael wrote: > Date: Tue, 13 Oct 2020 08:04:30 > From: Michael > Reply-To: gentoo-user@lists.gentoo.org > To: gentoo-user@lists.gentoo.org > Subject: Re: [gentoo-user] tried desktop profile > > On Tuesday, 13 October 2020 09:30:00 BST Jude DaShiell wrote: > > > I'm trying -j1 first. This machine has 50% of its maximum ram capacity > > in use and only has 2gb of ram capacity so yes this is a low memory > > machine. Why I'm using it at all is since it has available a 3tb hard > > drive. As long as /tmp directories under the $HOME directory structure > > have better system protection than /tmpfs and /var/tmp if the -j1 build > > fails I'll try pointing the memory to a safer place. I need to buy some > > decent sized ssd drives since that way I can do this on my new machine > > with 14GB of ram. > > OK, your RAM is not enough to build most large packages today. I'm thinking > rust, chromium, gcc, firefox, libreoffice, as examples. Most of these will > refuse to emerge right at the start as they check for adequate /var/tmp space. > In any case, the solution for low RAM PCs is to set up adequate space on your > disk for those packages only. Also to increase your swap, or add a swapfile > just for these larger packages. > > Have a read at 'Example 2' here: > > https://wiki.gentoo.org/wiki//etc/portage/package.env > > If your / fs partition has inadequate free space, you can set up /var/tmp/ > notmpfs to a different partition as long as you remember to mount it with > 'mount -o exec' and activate any swapfile(s) in advance. > > If you only set up more swap and leave PORTAGE_TMPDIR on RAM, the swapping > from your RAM to disk will inevitably incur I/O bottleneck conditions and will > start thrashing the disk, which could slow everything down to a crawl, > potentially for days. Therefore, it is worth switching to BFQ scheduler when > heavy swappage is expected: > > # echo bfq > /sys/block/sda/queue/scheduler > > where sda is the disk on which the swapfile or partition is set. -- Since the profile without the desktop emerged @world correctly after the make.conf file got adjusted correctly for this computer I think all it will run will be the basic profile. Even if it were possible to emerge the compiles necessary for a desktop and get the rest of the system built I'd probably run into difficulties later. I'll do the basic profile on this machine and maybe do a desktop on the newer machine at some later date.
[gentoo-user] [SOLVED] Kernel build failing on new install
On Tue, Oct 13, 2020 at 09:08:33AM +0200, J. Roeleveld wrote > On Tuesday, October 13, 2020 3:04:33 AM CEST Walter Dnes wrote: > > I'm near the tail-end of an install, trying to build the kernel. > > "make" gets an error as follows. Any ideas? > > > > (chroot) livecd /usr/src/linux # make > > CALLscripts/checksyscalls.sh > > CALLscripts/atomic/check-atomics.sh > > DESCEND objtool > > CHK include/generated/compile.h > > CHK kernel/kheaders_data.tar.xz > > GEN kernel/kheaders_data.tar.xz > > make[1]: *** [kernel/Makefile:133: kernel/kheaders_data.tar.xz] Error 127 > > make: *** [Makefile:1729: kernel] Error 2 > > Try disabling CONFIG_IKHEADERS in your kernel config. > I have it disabled on my system. Thanks. That fixed my problem. It seems that deleting stuff in "make menuconfig" is 90% of the work in a kernel config. All etherenet drivers are default enabled. What bugs me is that with every kernel upgrade, "make oldconfig" shows a whole bunch of new ethernet drivers, and they're all default enabled. -- Walter Dnes I don't run "desktop environments"; I run useful applications
Re: [gentoo-user] tried desktop profile
On Tuesday, 13 October 2020 09:30:00 BST Jude DaShiell wrote: > I'm trying -j1 first. This machine has 50% of its maximum ram capacity > in use and only has 2gb of ram capacity so yes this is a low memory > machine. Why I'm using it at all is since it has available a 3tb hard > drive. As long as /tmp directories under the $HOME directory structure > have better system protection than /tmpfs and /var/tmp if the -j1 build > fails I'll try pointing the memory to a safer place. I need to buy some > decent sized ssd drives since that way I can do this on my new machine > with 14GB of ram. OK, your RAM is not enough to build most large packages today. I'm thinking rust, chromium, gcc, firefox, libreoffice, as examples. Most of these will refuse to emerge right at the start as they check for adequate /var/tmp space. In any case, the solution for low RAM PCs is to set up adequate space on your disk for those packages only. Also to increase your swap, or add a swapfile just for these larger packages. Have a read at 'Example 2' here: https://wiki.gentoo.org/wiki//etc/portage/package.env If your / fs partition has inadequate free space, you can set up /var/tmp/ notmpfs to a different partition as long as you remember to mount it with 'mount -o exec' and activate any swapfile(s) in advance. If you only set up more swap and leave PORTAGE_TMPDIR on RAM, the swapping from your RAM to disk will inevitably incur I/O bottleneck conditions and will start thrashing the disk, which could slow everything down to a crawl, potentially for days. Therefore, it is worth switching to BFQ scheduler when heavy swappage is expected: # echo bfq > /sys/block/sda/queue/scheduler where sda is the disk on which the swapfile or partition is set. signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] tried desktop profile
On Tue, 13 Oct 2020, Andreas Fink wrote: > Date: Tue, 13 Oct 2020 02:28:01 > From: Andreas Fink > Reply-To: gentoo-user@lists.gentoo.org > To: gentoo-user@lists.gentoo.org > Subject: Re: [gentoo-user] tried desktop profile > > On Tue, 13 Oct 2020 02:10:04 -0400 > Jude DaShiell wrote: > > > x86_64-pc-linux-gnu-g++: fatal error: Killed signal terminated > > program cc1plus compilation terminated. > > These two lines stongly suggest that you ran out of memory while > compilation. You could try to build it with only one job (-j1), > currently you are using -j2. > Another option would be to buy more RAM ;) > And last but not least, you could increase your swap memory, but be > prepared that your system becomes unresponsive and compilation will > probably take forever. > > Cheers > Andreas > > -- With -j1 in MAKEOPTS in make.conf, I'm coming up on 75% compiled Last fraction I read was [1495/2082].
Re: [gentoo-user] Kernel build failing on new install
On Tuesday, 13 October 2020 08:08:33 BST J. Roeleveld wrote: > On Tuesday, October 13, 2020 3:04:33 AM CEST Walter Dnes wrote: > > I'm near the tail-end of an install, trying to build the kernel. > > > > "make" gets an error as follows. Any ideas? > > > > (chroot) livecd /usr/src/linux # make > > > > CALLscripts/checksyscalls.sh > > CALLscripts/atomic/check-atomics.sh > > DESCEND objtool > > CHK include/generated/compile.h > > CHK kernel/kheaders_data.tar.xz > > GEN kernel/kheaders_data.tar.xz > > > > make[1]: *** [kernel/Makefile:133: kernel/kheaders_data.tar.xz] Error 127 > > make: *** [Makefile:1729: kernel] Error 2 > > Try disabling CONFIG_IKHEADERS in your kernel config. > I have it disabled on my system. > > -- > Joost In kernel headers are typically used on OSs which do not have kernel headers in their filesystem, like Android on embedded devices. I'm not sure it would be required on a desktop/server PC, unless some special requirements dictate it to be so. signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] tried desktop profile
On Tue, 13 Oct 2020, Wols Lists wrote: > Date: Tue, 13 Oct 2020 03:54:48 > From: Wols Lists > Reply-To: gentoo-user@lists.gentoo.org > To: gentoo-user@lists.gentoo.org > Subject: Re: [gentoo-user] tried desktop profile > > On 13/10/20 07:48, J. Roeleveld wrote: > > On Tuesday, October 13, 2020 8:28:01 AM CEST Andreas Fink wrote: > >> On Tue, 13 Oct 2020 02:10:04 -0400 > >> > >> Jude DaShiell wrote: > >>> x86_64-pc-linux-gnu-g++: fatal error: Killed signal terminated > >>> program cc1plus compilation terminated. > >> > >> These two lines stongly suggest that you ran out of memory while > >> compilation. You could try to build it with only one job (-j1), > >> currently you are using -j2. > >> Another option would be to buy more RAM ;) > >> And last but not least, you could increase your swap memory, but be > >> prepared that your system becomes unresponsive and compilation will > >> probably take forever. > > > > One more suggestion, your PORTAGE_TMPDIR is set to "/var/tmp". > > By default, this is a tmpfs, which means it's all kept in RAM. > > If that's true then somebody has ed up! > > /tmp is specified as "files may disappear at any time" > > /var/tmp is specified as "temporary storage that should survive a reboot" > > Okay, that's not the exact wording, but that is the effect. > > > > If your system is that low on RAM, you might want to change that to a > > different location that is backed by a real disk. > > > The other thing about default tmpfs, is it defaults to half of ram. So > it could actually be quite small. I explicitly set mine to be big, > because I configure oodles of swap. > > Cheers, > Wol > I'm trying -j1 first. This machine has 50% of its maximum ram capacity in use and only has 2gb of ram capacity so yes this is a low memory machine. Why I'm using it at all is since it has available a 3tb hard drive. As long as /tmp directories under the $HOME directory structure have better system protection than /tmpfs and /var/tmp if the -j1 build fails I'll try pointing the memory to a safer place. I need to buy some decent sized ssd drives since that way I can do this on my new machine with 14GB of ram. > > --
Re: [gentoo-user] kernel bad scheduling from idle thread
On Tue, 13 Oct 2020 03:39:48 -0400, J. Roeleveld wrote: > > On Tuesday, October 13, 2020 9:26:08 AM CEST John Covici wrote: > > Hi. I am getting a very bad problem on my new kernel that I did after > > my last update. It gave me the following sequence over and over for a > > while, and then stopped with the messages, but the system crashed. I > > then went back to my old kernel which had been working for more than a > > month and got a couple of the same traces, but I hope its stopped, so > > I can write this message. A google search finds nothing relevant. > > > > Here is what I get: > > > > Oct 13 02:34:40 ccs.covici.com kernel: bad: scheduling from the idle > > thread! > > Oct 13 02:34:40 ccs.covici.com kernel: CPU: 5 PID: 0 Comm: swapper/5 > > Tainted: PW O 5.4.69-gentoo #1 > > Oct 13 02:34:40 ccs.covici.com kernel: Hardware name: Supermicro Super > > Server/X11SCA-W, BIOS 1.2 12/05/2019 > > Oct 13 02:34:40 ccs.covici.com kernel: Call Trace: > > Oct 13 02:34:40 ccs.covici.com kernel: dump_stack+0x71/0x98 > > Oct 13 02:34:40 ccs.covici.com kernel: dequeue_task_idle+0x1f/0x28 > > Oct 13 02:34:40 ccs.covici.com kernel: __schedule+0x167/0x5d6 > > Oct 13 02:34:40 ccs.covici.com kernel: schedule_idle+0x2a/0x33 > > Oct 13 02:34:40 ccs.covici.com kernel: do_idle+0x1d0/0x1f2 > > Oct 13 02:34:40 ccs.covici.com kernel: cpu_startup_entry+0x1d/0x1f > > Oct 13 02:34:40 ccs.covici.com kernel: start_secondary+0x17b/0x198 > > Oct 13 02:34:40 ccs.covici.com kernel: secondary_startup_64+0xa4/0xb0 > > Oct 13 02:34:40 ccs.covici.com kernel: bad: scheduling from the idle > > thread! > > Oct 13 02:34:40 ccs.covici.com kernel: CPU: 5 PID: 0 Comm: swapper/5 > > Tainted: PW O 5.4.69-gentoo #1 > > Oct 13 02:34:40 ccs.covici.com kernel: Hardware name: Supermicro Super > > Server/X11SCA-W, BIOS 1.2 12/05/2019 > > Oct 13 02:34:40 ccs.covici.com kernel: Call Trace: > > Oct 13 02:34:40 ccs.covici.com kernel: dump_stack+0x71/0x98 > > Oct 13 02:34:40 ccs.covici.com kernel: dequeue_task_idle+0x1f/0x28 > > Oct 13 02:34:40 ccs.covici.com kernel: __schedule+0x167/0x5d6 > > Oct 13 02:34:40 ccs.covici.com kernel: schedule_idle+0x2a/0x33 > > Oct 13 02:34:40 ccs.covici.com kernel: do_idle+0x1d0/0x1f2 > > Oct 13 02:34:40 ccs.covici.com kernel: cpu_startup_entry+0x1d/0x1f > > Oct 13 02:34:40 ccs.covici.com kernel: start_secondary+0x17b/0x198 > > Oct 13 02:34:40 ccs.covici.com kernel: secondary_startup_64+0xa4/0xb0 > > > > > > How would I even be able to find out what is happening here? > > > > Any thoughts appreciated. > > Which kernel versions have you tried? > And which additional kernel modules have you added? > > I myself am using 5.4.66 as previous versions were causing kernel panics. > This is the first 5.4 kernel I have tried. I did a make oldconfig and added a number of modules, but I would think the kernel would only load the ones that are needed. I could not tell you the exact ones, as I try to make the kernel somewhat general, so I may add things which may not be necessary for my particular situation. -- Your life is like a penny. You're going to lose it. The question is: How do you spend it? John Covici wb2una cov...@ccs.covici.com
Re: [gentoo-user] tried desktop profile
On 13/10/20 07:48, J. Roeleveld wrote: > On Tuesday, October 13, 2020 8:28:01 AM CEST Andreas Fink wrote: >> On Tue, 13 Oct 2020 02:10:04 -0400 >> >> Jude DaShiell wrote: >>> x86_64-pc-linux-gnu-g++: fatal error: Killed signal terminated >>> program cc1plus compilation terminated. >> >> These two lines stongly suggest that you ran out of memory while >> compilation. You could try to build it with only one job (-j1), >> currently you are using -j2. >> Another option would be to buy more RAM ;) >> And last but not least, you could increase your swap memory, but be >> prepared that your system becomes unresponsive and compilation will >> probably take forever. > > One more suggestion, your PORTAGE_TMPDIR is set to "/var/tmp". > By default, this is a tmpfs, which means it's all kept in RAM. If that's true then somebody has ed up! /tmp is specified as "files may disappear at any time" /var/tmp is specified as "temporary storage that should survive a reboot" Okay, that's not the exact wording, but that is the effect. > > If your system is that low on RAM, you might want to change that to a > different location that is backed by a real disk. > The other thing about default tmpfs, is it defaults to half of ram. So it could actually be quite small. I explicitly set mine to be big, because I configure oodles of swap. Cheers, Wol
Re: [gentoo-user] kernel bad scheduling from idle thread
On Tuesday, October 13, 2020 9:26:08 AM CEST John Covici wrote: > Hi. I am getting a very bad problem on my new kernel that I did after > my last update. It gave me the following sequence over and over for a > while, and then stopped with the messages, but the system crashed. I > then went back to my old kernel which had been working for more than a > month and got a couple of the same traces, but I hope its stopped, so > I can write this message. A google search finds nothing relevant. > > Here is what I get: > > Oct 13 02:34:40 ccs.covici.com kernel: bad: scheduling from the idle > thread! > Oct 13 02:34:40 ccs.covici.com kernel: CPU: 5 PID: 0 Comm: swapper/5 > Tainted: PW O 5.4.69-gentoo #1 > Oct 13 02:34:40 ccs.covici.com kernel: Hardware name: Supermicro Super > Server/X11SCA-W, BIOS 1.2 12/05/2019 > Oct 13 02:34:40 ccs.covici.com kernel: Call Trace: > Oct 13 02:34:40 ccs.covici.com kernel: dump_stack+0x71/0x98 > Oct 13 02:34:40 ccs.covici.com kernel: dequeue_task_idle+0x1f/0x28 > Oct 13 02:34:40 ccs.covici.com kernel: __schedule+0x167/0x5d6 > Oct 13 02:34:40 ccs.covici.com kernel: schedule_idle+0x2a/0x33 > Oct 13 02:34:40 ccs.covici.com kernel: do_idle+0x1d0/0x1f2 > Oct 13 02:34:40 ccs.covici.com kernel: cpu_startup_entry+0x1d/0x1f > Oct 13 02:34:40 ccs.covici.com kernel: start_secondary+0x17b/0x198 > Oct 13 02:34:40 ccs.covici.com kernel: secondary_startup_64+0xa4/0xb0 > Oct 13 02:34:40 ccs.covici.com kernel: bad: scheduling from the idle > thread! > Oct 13 02:34:40 ccs.covici.com kernel: CPU: 5 PID: 0 Comm: swapper/5 > Tainted: PW O 5.4.69-gentoo #1 > Oct 13 02:34:40 ccs.covici.com kernel: Hardware name: Supermicro Super > Server/X11SCA-W, BIOS 1.2 12/05/2019 > Oct 13 02:34:40 ccs.covici.com kernel: Call Trace: > Oct 13 02:34:40 ccs.covici.com kernel: dump_stack+0x71/0x98 > Oct 13 02:34:40 ccs.covici.com kernel: dequeue_task_idle+0x1f/0x28 > Oct 13 02:34:40 ccs.covici.com kernel: __schedule+0x167/0x5d6 > Oct 13 02:34:40 ccs.covici.com kernel: schedule_idle+0x2a/0x33 > Oct 13 02:34:40 ccs.covici.com kernel: do_idle+0x1d0/0x1f2 > Oct 13 02:34:40 ccs.covici.com kernel: cpu_startup_entry+0x1d/0x1f > Oct 13 02:34:40 ccs.covici.com kernel: start_secondary+0x17b/0x198 > Oct 13 02:34:40 ccs.covici.com kernel: secondary_startup_64+0xa4/0xb0 > > > How would I even be able to find out what is happening here? > > Any thoughts appreciated. Which kernel versions have you tried? And which additional kernel modules have you added? I myself am using 5.4.66 as previous versions were causing kernel panics. -- Joost
Re: [gentoo-user] app-shells/gentoo-zsh-completions no longer working
On Mon, 12 Oct 2020 18:28:44 -0400, John Covici wrote: > > % autoload -U compinit promptinit > > % compinit > > % promptinit; prompt gentoo > > > > [1] > > https://gitweb.gentoo.org/proj/zsh-completion.git/tree/src/_gentoo_repos > > [2] > > https://gitweb.gentoo.org/proj/zsh-completion.git/tree/src/_gentoo_repos_conf > > [3] > > https://gitweb.gentoo.org/repo/gentoo.git/tree/app-shells/zsh/zsh-5.8.ebuild#n197 > > > Thanks for the hint -- I had compinit already, but not promptinit and > that seems to have fixed it. Oddly though, I don't have promptinit and all works for me. -- Neil Bothwick Being politically correct means always having to say you're sorry. pgp3JWdFQC2M2.pgp Description: OpenPGP digital signature
[gentoo-user] kernel bad scheduling from idle thread
Hi. I am getting a very bad problem on my new kernel that I did after my last update. It gave me the following sequence over and over for a while, and then stopped with the messages, but the system crashed. I then went back to my old kernel which had been working for more than a month and got a couple of the same traces, but I hope its stopped, so I can write this message. A google search finds nothing relevant. Here is what I get: Oct 13 02:34:40 ccs.covici.com kernel: bad: scheduling from the idle thread! Oct 13 02:34:40 ccs.covici.com kernel: CPU: 5 PID: 0 Comm: swapper/5 Tainted: PW O 5.4.69-gentoo #1 Oct 13 02:34:40 ccs.covici.com kernel: Hardware name: Supermicro Super Server/X11SCA-W, BIOS 1.2 12/05/2019 Oct 13 02:34:40 ccs.covici.com kernel: Call Trace: Oct 13 02:34:40 ccs.covici.com kernel: dump_stack+0x71/0x98 Oct 13 02:34:40 ccs.covici.com kernel: dequeue_task_idle+0x1f/0x28 Oct 13 02:34:40 ccs.covici.com kernel: __schedule+0x167/0x5d6 Oct 13 02:34:40 ccs.covici.com kernel: schedule_idle+0x2a/0x33 Oct 13 02:34:40 ccs.covici.com kernel: do_idle+0x1d0/0x1f2 Oct 13 02:34:40 ccs.covici.com kernel: cpu_startup_entry+0x1d/0x1f Oct 13 02:34:40 ccs.covici.com kernel: start_secondary+0x17b/0x198 Oct 13 02:34:40 ccs.covici.com kernel: secondary_startup_64+0xa4/0xb0 Oct 13 02:34:40 ccs.covici.com kernel: bad: scheduling from the idle thread! Oct 13 02:34:40 ccs.covici.com kernel: CPU: 5 PID: 0 Comm: swapper/5 Tainted: PW O 5.4.69-gentoo #1 Oct 13 02:34:40 ccs.covici.com kernel: Hardware name: Supermicro Super Server/X11SCA-W, BIOS 1.2 12/05/2019 Oct 13 02:34:40 ccs.covici.com kernel: Call Trace: Oct 13 02:34:40 ccs.covici.com kernel: dump_stack+0x71/0x98 Oct 13 02:34:40 ccs.covici.com kernel: dequeue_task_idle+0x1f/0x28 Oct 13 02:34:40 ccs.covici.com kernel: __schedule+0x167/0x5d6 Oct 13 02:34:40 ccs.covici.com kernel: schedule_idle+0x2a/0x33 Oct 13 02:34:40 ccs.covici.com kernel: do_idle+0x1d0/0x1f2 Oct 13 02:34:40 ccs.covici.com kernel: cpu_startup_entry+0x1d/0x1f Oct 13 02:34:40 ccs.covici.com kernel: start_secondary+0x17b/0x198 Oct 13 02:34:40 ccs.covici.com kernel: secondary_startup_64+0xa4/0xb0 How would I even be able to find out what is happening here? Any thoughts appreciated. -- Your life is like a penny. You're going to lose it. The question is: How do you spend it? John Covici wb2una cov...@ccs.covici.com
Re: [gentoo-user] Kernel build failing on new install
On Tuesday, October 13, 2020 3:04:33 AM CEST Walter Dnes wrote: > I'm near the tail-end of an install, trying to build the kernel. > "make" gets an error as follows. Any ideas? > > (chroot) livecd /usr/src/linux # make > CALLscripts/checksyscalls.sh > CALLscripts/atomic/check-atomics.sh > DESCEND objtool > CHK include/generated/compile.h > CHK kernel/kheaders_data.tar.xz > GEN kernel/kheaders_data.tar.xz > make[1]: *** [kernel/Makefile:133: kernel/kheaders_data.tar.xz] Error 127 > make: *** [Makefile:1729: kernel] Error 2 Try disabling CONFIG_IKHEADERS in your kernel config. I have it disabled on my system. -- Joost
Re: [gentoo-user] tried desktop profile
On Tuesday, October 13, 2020 8:28:01 AM CEST Andreas Fink wrote: > On Tue, 13 Oct 2020 02:10:04 -0400 > > Jude DaShiell wrote: > > x86_64-pc-linux-gnu-g++: fatal error: Killed signal terminated > > program cc1plus compilation terminated. > > These two lines stongly suggest that you ran out of memory while > compilation. You could try to build it with only one job (-j1), > currently you are using -j2. > Another option would be to buy more RAM ;) > And last but not least, you could increase your swap memory, but be > prepared that your system becomes unresponsive and compilation will > probably take forever. One more suggestion, your PORTAGE_TMPDIR is set to "/var/tmp". By default, this is a tmpfs, which means it's all kept in RAM. If your system is that low on RAM, you might want to change that to a different location that is backed by a real disk. -- Joost
Re: [gentoo-user] tried desktop profile
On Tue, 13 Oct 2020 02:10:04 -0400 Jude DaShiell wrote: > x86_64-pc-linux-gnu-g++: fatal error: Killed signal terminated > program cc1plus compilation terminated. These two lines stongly suggest that you ran out of memory while compilation. You could try to build it with only one job (-j1), currently you are using -j2. Another option would be to buy more RAM ;) And last but not least, you could increase your swap memory, but be prepared that your system becomes unresponsive and compilation will probably take forever. Cheers Andreas