Re: [gentoo-user] UEFI-fails to boot
Well, check if you have CONFIG_FB_EFI enabled in your config. > H30/01/26 0:28、Dan Johansson のメール: > > On 25.01.2018 16:24, Mick wrote: >> On Thursday, 25 January 2018 13:18:13 GMT Dan Johansson wrote: >>> Thanks for the advice, now grub starts and I get the menu (now I just >>> have to figure out why the kernel hangs after being loaded). >> >> Glad you got it loading. >> >> Check at what stage the kernel oops - this would point at what in the kernel >> configuration has gone awry. You must build the chipset and filesystem >> drivers in the kernel (not as loadable modules), which may even drop you in >> a >> command prompt to troubleshoot from there. > > The kernel did not oops, there were just the messages > Booting 'Gentoo GNU/Linux' > Loading Linux x86_64-4.9.76-gentoo-r1 > and here it was just hanging. > > I normally do not like to have an initramfs but to com further I > installed genkernel and let it build the kernel. > The genkernel generated kernel does boot and everything, except on thing > (at the moment), works - I can ssh into the box and continue configuration. > > The one thing that is not working is that I do not see ANY boot-messages > on the console (it just says Loading Linux and Loading initial ramdisk > ...), and I do not get a login-prompt and can not switch VT. > > But as I said, I can login with ssh and can now continue with the setup, > the "Console" has to wait. > > KR > > -- > Dan Johansson > *** > This message is printed on 100% recycled electrons! > *** >
Re: [gentoo-user] Using both Gnome and KDE Plasma?
On Thu, 2018-01-25 at 19:24 +, Neil Bothwick wrote: > On Thu, 25 Jan 2018 18:52:34 +0200, Mart Raudsepp wrote: > > > That said, the global USE flag manual copy to make.conf might get > > outdated; e.g. we plan to add USE=wayland as global default to > > gnome > > profile at some point soon(tm). > > Could you source it from make.conf rather than copying the contents? > That > way you'll keep up with the profile changes. Not sure, probably not a good idea, but the next sentence that you cut from the reply gave an (untested) option right there in that e-mail...
Re: [gentoo-user] Using both Gnome and KDE Plasma?
On Thu, 25 Jan 2018 18:52:34 +0200, Mart Raudsepp wrote: > That said, the global USE flag manual copy to make.conf might get > outdated; e.g. we plan to add USE=wayland as global default to gnome > profile at some point soon(tm). Could you source it from make.conf rather than copying the contents? That way you'll keep up with the profile changes. -- Neil Bothwick Q: What's the proper plural of a 'Net-connected Windows machine? A: A Botnet pgpjRZkW0QZZ8.pgp Description: OpenPGP digital signature
Re: [gentoo-user] UEFI-fails to boot
On Thu, 25 Jan 2018 17:28:49 +0100, Dan Johansson wrote: > The one thing that is not working is that I do not see ANY boot-messages > on the console (it just says Loading Linux and Loading initial ramdisk > ...), and I do not get a login-prompt and can not switch VT. > > But as I said, I can login with ssh and can now continue with the setup, > the "Console" has to wait. Does /dev/console exist on your root filesystem? -- Neil Bothwick Angular Momentum Makes The World Go 'Round pgpVtkP_kGX_5.pgp Description: OpenPGP digital signature
Re: [gentoo-user] Using both Gnome and KDE Plasma?
On Thu, 2018-01-25 at 18:48 +0200, Mart Raudsepp wrote: > On Thu, 2018-01-25 at 01:16 +0200, Nikos Chantziaras wrote: > > One user wants Gnome, the other wants Plasma. Is this doable? > > > > Currently, Plasma is installed, and the profile is: > > > > default/linux/amd64/17.0/desktop/plasma/systemd > > > > The profiles seem to be either-or. There's one for plasma, one for > > gnome. But I need both now :-/ > > They just provide ease of installation by enabling some USE flags > globally or per-package for you, to avoid having to add them manually > on a fresh install or so. > > You can create a local mix-in by having /etc/portage/make.profile as > a > directory with an eapi and parent file, but that has some caveats, so > given that I leave the details to find out elsewhere when desired. > Or you can just use one and add the stuff for the other manually, > verbatim or as-needed. > What they do can be seen in > profiles/targets/desktop/gnome > profiles/targets/desktop/plasma > subdirectories of your PORTDIR (probably /usr/portage/) > > The global USE flags each adds are in make.defaults file and the per- > package USE flag tweaks are in package.use files. > > So you could for example keep using the systemd plasma profile, add > the > global USE flags from > /usr/portage/profiles/targets/desktop/gnome/make.defaults to your > make.conf and symlink > /usr/portage/profiles/targets/desktop/gnome/package.use to a > gnome.use > file under /etc/portage/package.use/ directory or so and voilà, you > got > effectively both. That said, the global USE flag manual copy to make.conf might get outdated; e.g. we plan to add USE=wayland as global default to gnome profile at some point soon(tm). Though make.defaults could be symlinked as well, I think to /etc/portage/profile/make.defaults Though these changes are rare, and it's a convenience anyways, e.g. you might have decided to globally enable wayland earlier already; or once the change is made you might counter it with a global disable in make.conf anyways.
Re: [gentoo-user] Using both Gnome and KDE Plasma?
On Thu, 2018-01-25 at 01:16 +0200, Nikos Chantziaras wrote: > One user wants Gnome, the other wants Plasma. Is this doable? > > Currently, Plasma is installed, and the profile is: > > default/linux/amd64/17.0/desktop/plasma/systemd > > The profiles seem to be either-or. There's one for plasma, one for > gnome. But I need both now :-/ They just provide ease of installation by enabling some USE flags globally or per-package for you, to avoid having to add them manually on a fresh install or so. You can create a local mix-in by having /etc/portage/make.profile as a directory with an eapi and parent file, but that has some caveats, so given that I leave the details to find out elsewhere when desired. Or you can just use one and add the stuff for the other manually, verbatim or as-needed. What they do can be seen in profiles/targets/desktop/gnome profiles/targets/desktop/plasma subdirectories of your PORTDIR (probably /usr/portage/) The global USE flags each adds are in make.defaults file and the per- package USE flag tweaks are in package.use files. So you could for example keep using the systemd plasma profile, add the global USE flags from /usr/portage/profiles/targets/desktop/gnome/make.defaults to your make.conf and symlink /usr/portage/profiles/targets/desktop/gnome/package.use to a gnome.use file under /etc/portage/package.use/ directory or so and voilà, you got effectively both.
Re: [gentoo-user] UEFI-fails to boot
On 25.01.2018 16:24, Mick wrote: > On Thursday, 25 January 2018 13:18:13 GMT Dan Johansson wrote: >> Thanks for the advice, now grub starts and I get the menu (now I just >> have to figure out why the kernel hangs after being loaded). > > Glad you got it loading. > > Check at what stage the kernel oops - this would point at what in the kernel > configuration has gone awry. You must build the chipset and filesystem > drivers in the kernel (not as loadable modules), which may even drop you in a > command prompt to troubleshoot from there. The kernel did not oops, there were just the messages Booting 'Gentoo GNU/Linux' Loading Linux x86_64-4.9.76-gentoo-r1 and here it was just hanging. I normally do not like to have an initramfs but to com further I installed genkernel and let it build the kernel. The genkernel generated kernel does boot and everything, except on thing (at the moment), works - I can ssh into the box and continue configuration. The one thing that is not working is that I do not see ANY boot-messages on the console (it just says Loading Linux and Loading initial ramdisk ...), and I do not get a login-prompt and can not switch VT. But as I said, I can login with ssh and can now continue with the setup, the "Console" has to wait. KR -- Dan Johansson *** This message is printed on 100% recycled electrons! ***
Re: [gentoo-user] Questions about Pale Moon
On Wed, 2018-01-24 at 16:30 +, Peter Humphrey wrote: > On Tuesday, 23 January 2018 12:08:12 GMT Arve Barsnes wrote: > > > Maybe check command line output? > > I get eight of these: "undefined symbol: UCNV_TO_U_CALLBACK_STOP". > The full > list is attached. > (pale moon:2395): Gtk-WARNING **: Error loading theme icon 'go- > previous' for stock: Unable to load image-loading module: > /usr/lib64/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-svg.so: > /usr/lib64/libxml2.so.2: undefined symbol: UCNV_TO_U_CALLBACK_STOP Your libxml2 appears broken after some ICU update and no automatic rebuild of libxml2 or something (maybe something failed in the rebuilds earlier in the queue from what portage figured has to get rebuilt after icu upgrade). Rebuild libxml2 and that issue probably goes away.
Re: [gentoo-user] UEFI-fails to boot
On Thursday, 25 January 2018 13:18:13 GMT Dan Johansson wrote: > Thanks for the advice, now grub starts and I get the menu (now I just > have to figure out why the kernel hangs after being loaded). Glad you got it loading. Check at what stage the kernel oops - this would point at what in the kernel configuration has gone awry. You must build the chipset and filesystem drivers in the kernel (not as loadable modules), which may even drop you in a command prompt to troubleshoot from there. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] chromium build failure
On Thu, Jan 25 2018, Neil Bothwick wrote: > On Wed, 24 Jan 2018 22:29:59 -0600, Canek Peláez Valdés wrote: > >> > I have two laptops with *very* similar gentoo distributions. The >> > newer machine had a successful build first try. The second older (4 >> > years) machine had both failures. Each machine has profile >> > default/linux/amd64/17.0/desktop/gnome/systemd >> >> If the USE flags for chromium on both machines are the same, simply >> create a binary package from the machine that already built it. > > Good idea, as long as the CFLAGS are the same too. Thank you both. I will try it. allan
Re: [gentoo-user] UEFI-fails to boot
On 25.01.2018 12:35, Mick wrote: > On Thursday, 25 January 2018 10:54:28 GMT Dan Johansson wrote: >> I have bought me a shiny new Supermicro X10DRi-T motherboard with two >> Xenon-E5-2620-v3 CPUs for use as a server. >> >> I have configured the MB for UEFI-mode only and my rescuecd-USB-key >> boots find in UEFI-mode. >> >> Following the Handbook and the "EFI System Partition" handbook I have >> created the following GPT-disklayout: >> >> root@sysresccd /root % parted /dev/sda print >> Model: ATA ST1000DX002-2DV1 (scsi) >> Disk /dev/sda: 1000GB >> Sector size (logical/physical): 512B/4096B >> Partition Table: gpt >> Disk Flags: >> >> Number Start End SizeFile system Name Flags >> 1 1049kB 3146kB 2097kB fat32 grub bios_grub > > The above implies an MBR partition table approach to booting your OS, which a > non-UEFI (BIOS only) MoBo will need. However you are meant to be using UEFI > *only* and GPT ... > > >> 2 3146kB 137MB 134MB ext2boot boot, esp > > and this partition is what should be used for an UEFI MoBo, but the fs is > wrong. Change it to fat32 and check with gdisk that its partition code is > EF00, which according to your 'boot, esp' flags it should be. This is your > EFI System Partition (ESP). > > >> 3 137MB 4429MB 4292MB linux-swap(v1) swap >> 4 4429MB 5503MB 1074MB root >> 5 5503MB 1000GB 995GB vg >> >> Partition-1 was created like this: mkfs.fat -F 32 -n efi-boot /dev/sda1 >> Partition-2 was created like this: mkfs.ext2 -T small /dev/sda2 >> >> GRUB_PLATFORMS was set to "efi-64" in make.conf before emerging grub:2 >> >> /boot and /boot/efi is mounted like this >> # mount | grep boot >> /dev/sda2 on /boot type ext2 (rw,relatime,errors=continue,user_xattr,acl) >> /dev/sda1 on /boot/efi type vfat >> (rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mi >> xed,errors=remount-ro) >> >> I had to remount /sys/firmware/efi/efivars in rw-mode, otherwise >> grub-install would complain. > > Yes, this has been the case for some time now. You will always need to > remount it as rw before you change the contents of the ESP boot partition. > It > is also mentioned here: > > https://wiki.gentoo.org/wiki/GRUB2 > > >> grub-install was run like this >> "grub-install --target=x86_64-efi --efi-directory=/boot/efi" > > But ... /boot/efi is not your ESP. > > >> And "grub-mkconfig -o /boot/grub/grub.cfg" has been run without any errors. >> >> efibootmgr shows my gentoo as the first entry >> # efibootmgr >> Timeout: 1 seconds >> BootOrder: ,0001,0002,0003 >> Boot gentoo >> Boot0001 Hard Drive >> Boot0002 Network Card >> Boot0003 UEFI: Built-in EFI Shell > > Use 'efibootmgr -v' to check the path of the .efi image it tries to boot and > check the path is correct without any typos. What you show above is only a > label. > > >> But when I boot without the USB-key inserted I always "lands" in the >> Built-in EFI Shell - NO sign of GRUB. >> >> Any suggestions where I have gone wrong? >> >> KR > > The ESP needs to be formatted as vfat and the GRUB boot image grubx64.efi > should be installed there. Thanks for the advice, now grub starts and I get the menu (now I just have to figure out why the kernel hangs after being loaded). And just for reference, my partition-table now looks like this: Number Start End SizeFile system Name Flags 1 1049kB 137MB 136MB fat32 boot boot, esp 3 137MB 4429MB 4292MB linux-swap(v1) swap 4 4429MB 5503MB 1074MB root 5 5503MB 1000GB 995GB vg KR -- Dan Johansson *** This message is printed on 100% recycled electrons! ***
Re: [gentoo-user] Re: Kernel 4.14.14 has meltdown / spectre info in /sys
On Thu, Jan 25, 2018 at 1:39 AM, Adam Carter wrote: > > Seems to me like point versions of 4.9 and 4.14 are often released on the > same date, but fixes that have gone into 4.14 don't make it into 4.9 until > the subsequent release. Could be my imagination. One of the issues with Meltdown/Spectre in particular is that the affected parts of the kernel have undergone some change over the years, and the changes themselves are not trivial. For some of the much older kernels the fixes are basically complete rewrites, with their own quality issues and timelines. For 4.9 that probably isn't as much of a factor, but it wouldn't surprise me if the changes still migrate their way backwards in time. There have been regressions with some of these changes, and that being the case the maintainers might want to both reduce the number of people impacted and also test them first on the kernels most similar to mainline where the patches were developed. -- Rich
Re: [gentoo-user] UEFI-fails to boot
On Thursday, 25 January 2018 11:07:40 GMT Wols Lists wrote: > But what I think you're supposed to do is use UEFI to load the linux > kernel directly ... not sure how you do that yet :-) > > Cheers, > Wol If you do not need/want to use a boot loader like GRUB you can use the efibootmgr to set the kernel image to boot directly. For example: efibootmgr --create --disk /dev/sda --part 1 --label "gentoo-4.14.14_20-Jan" --loader "\EFI\BOOT\bootx64-4.14.14-gentoo.efi" Where \EFI\BOOT\bootx64-4.14.14-gentoo.efi is found under: # tree /boot /boot ├── EFI └── BOOT ├── System.map-4.14.12-gentoo ├── System.map-4.14.14-gentoo ├── System.map-4.14.8-gentoo-r1 ├── bootx64-4.14.12-gentoo.efi ├── bootx64-4.14.14-gentoo.efi ├── bootx64-4.14.8-gentoo-r1.efi ├── config-4.14.12-gentoo ├── config-4.14.14-gentoo └── config-4.14.8-gentoo-r1 You should also set CONFIG_EFI_STUB=y in your kernel. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] UEFI-fails to boot
On Thursday, 25 January 2018 10:54:28 GMT Dan Johansson wrote: > I have bought me a shiny new Supermicro X10DRi-T motherboard with two > Xenon-E5-2620-v3 CPUs for use as a server. > > I have configured the MB for UEFI-mode only and my rescuecd-USB-key > boots find in UEFI-mode. > > Following the Handbook and the "EFI System Partition" handbook I have > created the following GPT-disklayout: > > root@sysresccd /root % parted /dev/sda print > Model: ATA ST1000DX002-2DV1 (scsi) > Disk /dev/sda: 1000GB > Sector size (logical/physical): 512B/4096B > Partition Table: gpt > Disk Flags: > > Number Start End SizeFile system Name Flags > 1 1049kB 3146kB 2097kB fat32 grub bios_grub The above implies an MBR partition table approach to booting your OS, which a non-UEFI (BIOS only) MoBo will need. However you are meant to be using UEFI *only* and GPT ... > 2 3146kB 137MB 134MB ext2boot boot, esp and this partition is what should be used for an UEFI MoBo, but the fs is wrong. Change it to fat32 and check with gdisk that its partition code is EF00, which according to your 'boot, esp' flags it should be. This is your EFI System Partition (ESP). > 3 137MB 4429MB 4292MB linux-swap(v1) swap > 4 4429MB 5503MB 1074MB root > 5 5503MB 1000GB 995GB vg > > Partition-1 was created like this: mkfs.fat -F 32 -n efi-boot /dev/sda1 > Partition-2 was created like this: mkfs.ext2 -T small /dev/sda2 > > GRUB_PLATFORMS was set to "efi-64" in make.conf before emerging grub:2 > > /boot and /boot/efi is mounted like this > # mount | grep boot > /dev/sda2 on /boot type ext2 (rw,relatime,errors=continue,user_xattr,acl) > /dev/sda1 on /boot/efi type vfat > (rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mi > xed,errors=remount-ro) > > I had to remount /sys/firmware/efi/efivars in rw-mode, otherwise > grub-install would complain. Yes, this has been the case for some time now. You will always need to remount it as rw before you change the contents of the ESP boot partition. It is also mentioned here: https://wiki.gentoo.org/wiki/GRUB2 > grub-install was run like this > "grub-install --target=x86_64-efi --efi-directory=/boot/efi" But ... /boot/efi is not your ESP. > And "grub-mkconfig -o /boot/grub/grub.cfg" has been run without any errors. > > efibootmgr shows my gentoo as the first entry > # efibootmgr > Timeout: 1 seconds > BootOrder: ,0001,0002,0003 > Boot gentoo > Boot0001 Hard Drive > Boot0002 Network Card > Boot0003 UEFI: Built-in EFI Shell Use 'efibootmgr -v' to check the path of the .efi image it tries to boot and check the path is correct without any typos. What you show above is only a label. > But when I boot without the USB-key inserted I always "lands" in the > Built-in EFI Shell - NO sign of GRUB. > > Any suggestions where I have gone wrong? > > KR The ESP needs to be formatted as vfat and the GRUB boot image grubx64.efi should be installed there. HTH. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] UEFI-fails to boot
On 25/01/18 10:54, Dan Johansson wrote: > But when I boot without the USB-key inserted I always "lands" in the > Built-in EFI Shell - NO sign of GRUB. > > Any suggestions where I have gone wrong? Well, in your position I wouldn't be trying to load grub. I've got a new mobo (with a Ryzen 3 :-) but that won't POST at the moment, so I'm going to be in your position very soon building a new UEFI system. But what I think you're supposed to do is use UEFI to load the linux kernel directly ... not sure how you do that yet :-) Cheers, Wol
[gentoo-user] UEFI-fails to boot
I have bought me a shiny new Supermicro X10DRi-T motherboard with two Xenon-E5-2620-v3 CPUs for use as a server. I have configured the MB for UEFI-mode only and my rescuecd-USB-key boots find in UEFI-mode. Following the Handbook and the "EFI System Partition" handbook I have created the following GPT-disklayout: root@sysresccd /root % parted /dev/sda print Model: ATA ST1000DX002-2DV1 (scsi) Disk /dev/sda: 1000GB Sector size (logical/physical): 512B/4096B Partition Table: gpt Disk Flags: Number Start End SizeFile system Name Flags 1 1049kB 3146kB 2097kB fat32 grub bios_grub 2 3146kB 137MB 134MB ext2boot boot, esp 3 137MB 4429MB 4292MB linux-swap(v1) swap 4 4429MB 5503MB 1074MB root 5 5503MB 1000GB 995GB vg Partition-1 was created like this: mkfs.fat -F 32 -n efi-boot /dev/sda1 Partition-2 was created like this: mkfs.ext2 -T small /dev/sda2 GRUB_PLATFORMS was set to "efi-64" in make.conf before emerging grub:2 /boot and /boot/efi is mounted like this # mount | grep boot /dev/sda2 on /boot type ext2 (rw,relatime,errors=continue,user_xattr,acl) /dev/sda1 on /boot/efi type vfat (rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,errors=remount-ro) I had to remount /sys/firmware/efi/efivars in rw-mode, otherwise grub-install would complain. grub-install was run like this "grub-install --target=x86_64-efi --efi-directory=/boot/efi" And "grub-mkconfig -o /boot/grub/grub.cfg" has been run without any errors. efibootmgr shows my gentoo as the first entry # efibootmgr Timeout: 1 seconds BootOrder: ,0001,0002,0003 Boot gentoo Boot0001 Hard Drive Boot0002 Network Card Boot0003 UEFI: Built-in EFI Shell But when I boot without the USB-key inserted I always "lands" in the Built-in EFI Shell - NO sign of GRUB. Any suggestions where I have gone wrong? KR -- Dan Johansson, *** This message is printed on 100% recycled electrons! *** smime.p7s Description: S/MIME Cryptographic Signature
Re: [gentoo-user] chromium build failure
On Wed, 24 Jan 2018 22:29:59 -0600, Canek Peláez Valdés wrote: > > I have two laptops with *very* similar gentoo distributions. The > > newer machine had a successful build first try. The second older (4 > > years) machine had both failures. Each machine has profile > > default/linux/amd64/17.0/desktop/gnome/systemd > > If the USE flags for chromium on both machines are the same, simply > create a binary package from the machine that already built it. Good idea, as long as the CFLAGS are the same too. -- Neil Bothwick Like an atheist in a grave: all dressed up and no place to go. pgpWxemlAMKhH.pgp Description: OpenPGP digital signature
Re: [gentoo-user] Re: Using both Gnome and KDE Plasma?
On Thu, 25 Jan 2018 04:23:59 +0200, Nikos Chantziaras wrote: > > The differences between the profiles are mainly the USE flags set. As > > your system is already built with the plasma profiles, I'd add gnome > > to your USE flags and emerge gnome-base/gnome. Portage will shout at > > you if you need to enable any other USE flags. > > Don't know. There's use flags that seem to be set because it's > recommended to set them, not because they're needed. > > It seems to me that I have to choose either-or here in order to have a > well working desktop. Otherwise, if something doesn't seem work right, > I'll never know whether it's because of the profile or because of > something else. > > I suppose I have to switch to the gnome profile and then see if KDE is > still working as before. But it seems I'll have to manually replicate > whatever the plasma profile is setting by hand. Just compare the USE flags defaults for the two desktops. % cat /var/portage/profiles/targets/desktop/gnome/make.defaults # Copyright 1999-2015 Gentoo Foundation # Distributed under the terms of the GNU General Public License v2 USE="colord eds evo gnome gnome-keyring gnome-online-accounts gstreamer introspection libsecret nautilus pulseaudio tracker" % cat /var/portage/profiles/targets/desktop/plasma/make.defaults # Copyright 1999-2018 Gentoo Foundation # Distributed under the terms of the GNU General Public License v2 USE="activities consolekit declarative dri kde kipi kwallet phonon plasma policykit qml semantic-desktop widgets xcomposite" As long as you have all of these set you should be OK, whichever profile you start with. After all, a profile is only a convenient starting point. You could even create your own profile for both desktops, which could be useful if you want to replicate the setup on more than one computer. -- Neil Bothwick Synonym: a word you use when you can't spell the other one. pgpTEnwfBotpp.pgp Description: OpenPGP digital signature