* [gentoo-user] Anybody got a Gentoo system working under uclibc? @ 2016-04-27 14:12 waltdnes 2016-04-27 17:25 ` Corbin 2016-05-01 5:41 ` waltdnes 0 siblings, 2 replies; 10+ messages in thread From: waltdnes @ 2016-04-27 14:12 UTC (permalink / raw To: Gentoo Users List I'm testing uclibc in a 32-bit QEMU Gentoo VM. I want to eventually install on my ancient 32-bit-only netbook. I'm running into problems. See https://forums.gentoo.org/viewtopic-p-7912494.html To summarize... * static IP works for eth0, but dhcpcd fails, which is bad news for a laptop or netbook * when building x-rg-server, the builds for xf86-input-keyboard and xf86-video-intel fail to compile * www-client/w3m fails to build, so I can't even get decent textmode web browsing -- Walter Dnes <waltdnes@waltdnes.org> I don't run "desktop environments"; I run useful applications ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [gentoo-user] Anybody got a Gentoo system working under uclibc? 2016-04-27 14:12 [gentoo-user] Anybody got a Gentoo system working under uclibc? waltdnes @ 2016-04-27 17:25 ` Corbin 2016-04-27 21:31 ` waltdnes 2016-05-01 5:41 ` waltdnes 1 sibling, 1 reply; 10+ messages in thread From: Corbin @ 2016-04-27 17:25 UTC (permalink / raw To: gentoo-user On 04/27/2016 09:12 AM, waltdnes@waltdnes.org wrote: > I'm testing uclibc in a 32-bit QEMU Gentoo VM. I want to eventually > install on my ancient 32-bit-only netbook. I'm running into problems. > See https://forums.gentoo.org/viewtopic-p-7912494.html > > To summarize... > * static IP works for eth0, but dhcpcd fails, which is bad news for a > laptop or netbook > * when building x-rg-server, the builds for xf86-input-keyboard and > xf86-video-intel fail to compile > * www-client/w3m fails to build, so I can't even get decent textmode > web browsing > You already found my test load :) Glad it helped someone. Did you have/add "evdev" in "make.conf" for INPUT_DEVICES? "evdev" is built-into-the-kernel or a kernel module. The newer input function, augmenting/replacing the AT keyboard and /dev/psaux. ( Seems "evdev" is almost replacing the functions of "xf86-input" libraries for Xorg.) What CPU flags did you set/unset for the build? My reason for asking ... a lot of software has an unmarked dependency on "sse/sse2". If the info on the Atom CPU I found was correct ... it doesn't have sse/sse2 instruction sets. Useful Info : https://wiki.gentoo.org/wiki/Safe_CFLAGS Hope this helps. ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [gentoo-user] Anybody got a Gentoo system working under uclibc? 2016-04-27 17:25 ` Corbin @ 2016-04-27 21:31 ` waltdnes 2016-04-28 0:30 ` Corbin 0 siblings, 1 reply; 10+ messages in thread From: waltdnes @ 2016-04-27 21:31 UTC (permalink / raw To: gentoo-user On Wed, Apr 27, 2016 at 12:25:41PM -0500, Corbin wrote > > On 04/27/2016 09:12 AM, waltdnes@waltdnes.org wrote: > > I'm testing uclibc in a 32-bit QEMU Gentoo VM. I want to eventually > > install on my ancient 32-bit-only netbook. I'm running into problems. > > See https://forums.gentoo.org/viewtopic-p-7912494.html > > > > To summarize... > > * static IP works for eth0, but dhcpcd fails, which is bad news for a > > laptop or netbook > > * when building x-rg-server, the builds for xf86-input-keyboard and > > xf86-video-intel fail to compile > > * www-client/w3m fails to build, so I can't even get decent textmode > > web browsing > > > > You already found my test load :) > Glad it helped someone. > > Did you have/add "evdev" in "make.conf" for INPUT_DEVICES? > "evdev" is built-into-the-kernel or a kernel module. The newer input > function, augmenting/replacing the AT keyboard and /dev/psaux. > ( Seems "evdev" is almost replacing the functions of "xf86-input" > libraries for Xorg.) No. Do I have to set something in the kernel? Will the keyboard work with evdev and without xf86-input-keyboard? My netbook uses framebuffer video driver, so I might try using that in the uclibc VM. > What CPU flags did you set/unset for the build? > > My reason for asking ... a lot of software has an unmarked dependency on > "sse/sse2". If the info on the Atom CPU I found was correct ... it > doesn't have sse/sse2 instruction sets. My Atom netbook has them, according to /proc/cpuinfo flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc arch_perfmon pebs bts aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 xtpr pdcm movbe lahf_lm dtherm tpr_shadow vnmi flexpriority Note also the "movbe" flag. *I CANNOT TEST RUN MY NETBOOK-OPTIMIZED CODE I BUILD IN THE QEMU VM*. QEMU claims to emulate "Atom" CPUs. There are over 100 different atoms in the periodic table known to physicists, and Intel is just as bad, with all their "Atom" variants. QEMU's "Atom" emulation does not support the "movbe" flag. So I simply run with the host CPU in a 32-bit VM, and set "-march=native". When building for the netbook, I build with "-march=bonnell". Important distinction... different "-march" values != crossdev As long as the "CHOST" variable is identical, you can build for different "-march" variants, even those that won't run in the build environment. > Useful Info : https://wiki.gentoo.org/wiki/Safe_CFLAGS See also http://gentoo-en.vfose.ru/wiki/Safe_Cflags#-march.3Dnative gcc -march=native -E -v - </dev/null 2>&1 | grep cc1 Ignore all the fancy flags. Both of those pages come up with output that shows "-march=bonnell" on my Atom https://en.wikipedia.org/wiki/Bonnell_%28microarchitecture%29 I have a standard glibc 32-bit VM which I use for Gentoo distcc for the Atom, as well as manually building Pale Moon web browser. I use the following settings for both... CFLAGS="-O2 -march=bonnell -mfpmath=sse -pipe -fomit-frame-pointer \ -fno-unwind-tables -fno-asynchronous-unwind-tables" CCFLAGS="${CFLAGS}" "-march=bonnell" implies all the other flags, so no need to enter them into the CFLAGS variable. Thanks for that initial post. I had almost given up on even a minimal install, before I found it. -- Walter Dnes <waltdnes@waltdnes.org> I don't run "desktop environments"; I run useful applications ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [gentoo-user] Anybody got a Gentoo system working under uclibc? 2016-04-27 21:31 ` waltdnes @ 2016-04-28 0:30 ` Corbin 2016-04-28 3:16 ` waltdnes 0 siblings, 1 reply; 10+ messages in thread From: Corbin @ 2016-04-28 0:30 UTC (permalink / raw To: gentoo-user On 04/27/2016 04:31 PM, waltdnes@waltdnes.org wrote: > On Wed, Apr 27, 2016 at 12:25:41PM -0500, Corbin wrote >> >> On 04/27/2016 09:12 AM, waltdnes@waltdnes.org wrote: >>> I'm testing uclibc in a 32-bit QEMU Gentoo VM. I want to eventually >>> install on my ancient 32-bit-only netbook. I'm running into problems. >>> See https://forums.gentoo.org/viewtopic-p-7912494.html >>> >>> To summarize... >>> * static IP works for eth0, but dhcpcd fails, which is bad news for a >>> laptop or netbook >>> * when building x-rg-server, the builds for xf86-input-keyboard and >>> xf86-video-intel fail to compile >>> * www-client/w3m fails to build, so I can't even get decent textmode >>> web browsing >>> >> >> You already found my test load :) >> Glad it helped someone. >> >> Did you have/add "evdev" in "make.conf" for INPUT_DEVICES? >> "evdev" is built-into-the-kernel or a kernel module. The newer input >> function, augmenting/replacing the AT keyboard and /dev/psaux. >> ( Seems "evdev" is almost replacing the functions of "xf86-input" >> libraries for Xorg.) > > No. Do I have to set something in the kernel? Will the keyboard work > with evdev and without xf86-input-keyboard? My netbook uses framebuffer > video driver, so I might try using that in the uclibc VM. > >> What CPU flags did you set/unset for the build? >> >> My reason for asking ... a lot of software has an unmarked dependency on >> "sse/sse2". If the info on the Atom CPU I found was correct ... it >> doesn't have sse/sse2 instruction sets. > > My Atom netbook has them, according to /proc/cpuinfo > > flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge > mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx > constant_tsc arch_perfmon pebs bts aperfmperf pni dtes64 monitor ds_cpl > vmx est tm2 ssse3 xtpr pdcm movbe lahf_lm dtherm tpr_shadow vnmi > flexpriority > > Note also the "movbe" flag. *I CANNOT TEST RUN MY NETBOOK-OPTIMIZED > CODE I BUILD IN THE QEMU VM*. QEMU claims to emulate "Atom" CPUs. > There are over 100 different atoms in the periodic table known to > physicists, and Intel is just as bad, with all their "Atom" variants. > QEMU's "Atom" emulation does not support the "movbe" flag. So I simply > run with the host CPU in a 32-bit VM, and set "-march=native". When > building for the netbook, I build with "-march=bonnell". Important > distinction... > > different "-march" values != crossdev > > As long as the "CHOST" variable is identical, you can build for > different "-march" variants, even those that won't run in the build > environment. > >> Useful Info : https://wiki.gentoo.org/wiki/Safe_CFLAGS > > See also http://gentoo-en.vfose.ru/wiki/Safe_Cflags#-march.3Dnative > > gcc -march=native -E -v - </dev/null 2>&1 | grep cc1 > > Ignore all the fancy flags. Both of those pages come up with output > that shows "-march=bonnell" on my Atom > https://en.wikipedia.org/wiki/Bonnell_%28microarchitecture%29 > > I have a standard glibc 32-bit VM which I use for Gentoo distcc for > the Atom, as well as manually building Pale Moon web browser. I use the > following settings for both... > > CFLAGS="-O2 -march=bonnell -mfpmath=sse -pipe -fomit-frame-pointer \ > -fno-unwind-tables -fno-asynchronous-unwind-tables" > > CCFLAGS="${CFLAGS}" > > "-march=bonnell" implies all the other flags, so no need to enter them > into the CFLAGS variable. > > Thanks for that initial post. I had almost given up on even a minimal > install, before I found it. > Your Welcome. Link for "evdev" : https://en.wikipedia.org/wiki/Evdev The default kernel config has "evdev" built into the kernel. On my desktop, Nvidia drivers do look for and use "evdev" without Wayland support in Xorg. ( XFCE ) This might be pedantic ... ... add "sse sse2 sse3 ssse3 mmx acpi -mmxext" to your USE flags in "make.conf" ( if not already present. ) Some packages look for those flags in strange ways. Using a one-problem-at-a-time approach ... < Proposed Test / Xorg Fix > ... adding "libinput" to the USE flags in your "make.conf". ... setting INPUT_DEVICES="evdev" in "make.conf". Desktops that should work with "libinput/evdev" ONLY are QT4, QT5, XFCE. XFCE provides its own keyboard library and config applets/plugins. Don't know about QT4 / QT5. If you have not already done this ... might want to set VIDEO_CARDS="fbdev vesa" or just "fbdev" to save compile time. This is getting very interesting. Please let us know how this works out :) ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [gentoo-user] Anybody got a Gentoo system working under uclibc? 2016-04-28 0:30 ` Corbin @ 2016-04-28 3:16 ` waltdnes 2016-04-28 13:33 ` Corbin 0 siblings, 1 reply; 10+ messages in thread From: waltdnes @ 2016-04-28 3:16 UTC (permalink / raw To: gentoo-user On Wed, Apr 27, 2016 at 07:30:12PM -0500, Corbin wrote > > > > Your Welcome. > > Link for "evdev" : https://en.wikipedia.org/wiki/Evdev > > The default kernel config has "evdev" built into the kernel. > > On my desktop, Nvidia drivers do look for and use "evdev" without > Wayland support in Xorg. ( XFCE ) > > This might be pedantic ... > ... add "sse sse2 sse3 ssse3 mmx acpi -mmxext" to your USE flags in > "make.conf" ( if not already present. ) Some packages look for those > flags in strange ways. > > Using a one-problem-at-a-time approach ... > > < Proposed Test / Xorg Fix > > ... adding "libinput" to the USE flags in your "make.conf". > ... setting INPUT_DEVICES="evdev" in "make.conf". > > Desktops that should work with "libinput/evdev" ONLY are QT4, QT5, XFCE. > XFCE provides its own keyboard library and config applets/plugins. Don't > know about QT4 / QT5. > > If you have not already done this ... might want to set > VIDEO_CARDS="fbdev vesa" or just "fbdev" to save compile time. > > This is getting very interesting. > Please let us know how this works out :) It's not the cpu flags. I have an identical glibc-based VM where the drivers build just fine. Also, this is a code problem... both with and without "evdev"... > libtool: compile: i686-gentoo-linux-uclibc-gcc -std=gnu99 -DHAVE_CONFIG_H -I. > -I/var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src > -I.. -fvisibility=hidden -I/usr/include/xorg -I/usr/include/pixman-1 > -I/usr/include/X11/dri -I/usr/include/libdrm -Wall -Wpointer-arith > -Wmissing-declarations -Wformat=2 -Wstrict-prototypes > -Wmissing-prototypes -Wnested-externs -Wbad-function-cast > -Wold-style-definition -Wdeclaration-after-statement -Wunused > -Wuninitialized -Wshadow -Wmissing-noreturn -Wmissing-format-attribute > -Wredundant-decls -Wlogical-op -Werror=implicit -Werror=nonnull > -Werror=init-self -Werror=main -Werror=missing-braces > -Werror=sequence-point -Werror=return-type -Werror=trigraphs > -Werror=array-bounds -Werror=write-strings -Werror=address > -Werror=int-to-pointer-cast -Werror=pointer-to-int-cast > -fno-strict-aliasing -O2 -march=native -mfpmath=sse -fomit-frame-pointer > -pipe -fno-unwind-tables -fno-asynchronous-unwind-tables -c > /var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src/at_scancode.c > -fPIC -DPIC -o .libs/at_scancode.o > /var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src/lnx_kbd.c: > In function 'OpenKeyboard': > /var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src/lnx_kbd.c:194:8: > error: implicit declaration of function 'getpgid' > [-Werror=implicit-function-declaration] > rc = tcsetpgrp(pInfo->fd, getpgid(0)); > ^ > /var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src/lnx_kbd.c:194:8: > warning: nested extern declaration of 'getpgid' [-Wnested-externs] -- Walter Dnes <waltdnes@waltdnes.org> I don't run "desktop environments"; I run useful applications ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [gentoo-user] Anybody got a Gentoo system working under uclibc? 2016-04-28 3:16 ` waltdnes @ 2016-04-28 13:33 ` Corbin 2016-04-28 16:14 ` [gentoo-user] " James 2016-04-28 23:07 ` [gentoo-user] " waltdnes 0 siblings, 2 replies; 10+ messages in thread From: Corbin @ 2016-04-28 13:33 UTC (permalink / raw To: gentoo-user On 04/27/2016 10:16 PM, waltdnes@waltdnes.org wrote: > On Wed, Apr 27, 2016 at 07:30:12PM -0500, Corbin wrote >> >> >> >> Your Welcome. >> >> Link for "evdev" : https://en.wikipedia.org/wiki/Evdev >> >> The default kernel config has "evdev" built into the kernel. >> >> On my desktop, Nvidia drivers do look for and use "evdev" without >> Wayland support in Xorg. ( XFCE ) >> >> This might be pedantic ... >> ... add "sse sse2 sse3 ssse3 mmx acpi -mmxext" to your USE flags in >> "make.conf" ( if not already present. ) Some packages look for those >> flags in strange ways. >> >> Using a one-problem-at-a-time approach ... >> >> < Proposed Test / Xorg Fix > >> ... adding "libinput" to the USE flags in your "make.conf". >> ... setting INPUT_DEVICES="evdev" in "make.conf". >> >> Desktops that should work with "libinput/evdev" ONLY are QT4, QT5, XFCE. >> XFCE provides its own keyboard library and config applets/plugins. Don't >> know about QT4 / QT5. >> >> If you have not already done this ... might want to set >> VIDEO_CARDS="fbdev vesa" or just "fbdev" to save compile time. >> >> This is getting very interesting. >> Please let us know how this works out :) > > It's not the cpu flags. I have an identical glibc-based VM where the > drivers build just fine. Also, this is a code problem... both with and > without "evdev"... > >> libtool: compile: i686-gentoo-linux-uclibc-gcc -std=gnu99 -DHAVE_CONFIG_H -I. >> -I/var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src >> -I.. -fvisibility=hidden -I/usr/include/xorg -I/usr/include/pixman-1 >> -I/usr/include/X11/dri -I/usr/include/libdrm -Wall -Wpointer-arith >> -Wmissing-declarations -Wformat=2 -Wstrict-prototypes >> -Wmissing-prototypes -Wnested-externs -Wbad-function-cast >> -Wold-style-definition -Wdeclaration-after-statement -Wunused >> -Wuninitialized -Wshadow -Wmissing-noreturn -Wmissing-format-attribute >> -Wredundant-decls -Wlogical-op -Werror=implicit -Werror=nonnull >> -Werror=init-self -Werror=main -Werror=missing-braces >> -Werror=sequence-point -Werror=return-type -Werror=trigraphs >> -Werror=array-bounds -Werror=write-strings -Werror=address >> -Werror=int-to-pointer-cast -Werror=pointer-to-int-cast >> -fno-strict-aliasing -O2 -march=native -mfpmath=sse -fomit-frame-pointer >> -pipe -fno-unwind-tables -fno-asynchronous-unwind-tables -c >> /var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src/at_scancode.c >> -fPIC -DPIC -o .libs/at_scancode.o >> /var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src/lnx_kbd.c: >> In function 'OpenKeyboard': >> /var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src/lnx_kbd.c:194:8: >> error: implicit declaration of function 'getpgid' >> [-Werror=implicit-function-declaration] >> rc = tcsetpgrp(pInfo->fd, getpgid(0)); >> ^ >> /var/tmp/portage/x11-drivers/xf86-input-keyboard-1.8.1/work/xf86-input-keyboard-1.8.1/src/lnx_kbd.c:194:8: >> warning: nested extern declaration of 'getpgid' [-Wnested-externs] > Did some searching on "getpgid" ... looks to be related to GNU glibc. > https://www.gnu.org/software/libc/manual/html_node/Process-Group-Functions.html > error: implicit declaration of function 'getpgid' Maybe they used an implicit glibc getpgid call instead of the specific posix_getpgid call? ( guessing ) ( Is Xorg dumping POSIX compliance? Or is this a bug? ) Questions ... if you will permit : Are you saying that in "make.conf" you set INPUT_DEVICES="evdev" and did a test compile run? The emerge you tried ... was it "xorg-base/xorg-x11"? Or did you try a meta package for a desktop? If a hard dependency link between Xorg server -> xf86-input-keyboard exists, this will never work. I have no idea at this point if this is true. What I have been reading suggests that the xf86-input keyboard and mouse libs are being phased out. With that call ?error? ... Xorg may be an impossible goal / waste of time on uClibc. Thank you for sharing this info. ^ permalink raw reply [flat|nested] 10+ messages in thread
* [gentoo-user] Re: Anybody got a Gentoo system working under uclibc? 2016-04-28 13:33 ` Corbin @ 2016-04-28 16:14 ` James 2016-04-28 23:07 ` [gentoo-user] " waltdnes 1 sibling, 0 replies; 10+ messages in thread From: James @ 2016-04-28 16:14 UTC (permalink / raw To: gentoo-user Corbin <corbinbird <at> charter.net> writes: > If a hard dependency link between Xorg server -> xf86-input-keyboard > exists, this will never work. I have no idea at this point if this is > true. What I have been reading suggests that the xf86-input keyboard and > mouse libs are being phased out. > With that call ?error? ... Xorg may be an impossible goal / waste of > time on uClibc. Hmmmm. Anthony (Blueness@gentoo.org) is extraordinarily knowledgable with uclibc and related issues. So that is where I would seek clarity, advice and pathways for what you seek [1]. hth, James [1] https://wiki.gentoo.org/wiki/Project:Hardened_uClibc ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [gentoo-user] Anybody got a Gentoo system working under uclibc? 2016-04-28 13:33 ` Corbin 2016-04-28 16:14 ` [gentoo-user] " James @ 2016-04-28 23:07 ` waltdnes 2016-04-29 3:52 ` Corbin 1 sibling, 1 reply; 10+ messages in thread From: waltdnes @ 2016-04-28 23:07 UTC (permalink / raw To: gentoo-user On Thu, Apr 28, 2016 at 08:33:01AM -0500, Corbin wrote > > > Questions ... if you will permit : > > Are you saying that in "make.conf" you set INPUT_DEVICES="evdev" and did > a test compile run? > The emerge you tried ... was it "xorg-base/xorg-x11"? > Or did you try a meta package for a desktop? xorg-server, intending to add ICEWM later. > If a hard dependency link between Xorg server -> xf86-input-keyboard > exists, this will never work. I have no idea at this point if this is > true. What I have been reading suggests that the xf86-input keyboard and > mouse libs are being phased out. > > With that call ?error? ... Xorg may be an impossible goal / waste of > time on uClibc. I tried INPUT_DEVICES="evdev mouse keyboard", which probably caused the problem. With INPUT_DEVICES="evdev" and VIDEO_CARDS="vesa fbdev", things build OK. Whether they'll work, I don't know, but at least it builds. ICEWM appears to be building too. Actually, ICEWM has a "(uclibc)" USE flag which is automatically hard-invoked or hard-masked depending on whether or not the system uses uclibc. James points to https://wiki.gentoo.org/wiki/Project:Hardened_uClibc which mentions "Lilblue"... > Continued developments in uClibc have made it increasingly suitable > for systems like Lilblue, our security-enhanced, fully featured > XFCE4 desktop, amd64 system built on uClibc. WHEEEEEEEEEE! IT WOIKS! Whilst I was typing away, the ICEWM and xterm builds finished. I've brought up the basic ICEWM window with 4 work areas. The fixes to my problem were... INPUT_DEVICES="evdev" VIDEO_CARDS="vesa fbdev" enable "udev" flag in make.conf (actually it runs on eudev, but...) I've never used "evdev" before, so I was not familiar with how to set it up. It's a bit disappointing, because evdev *DEMANDS* udev, so I can forget about switching in busybox's mdev for udev. BTW, I stumbled over "1 weird little tip" on the internet, to make debugging easier for many bootup problems, not just uclibc. Add the "--noclear" option to the first console in /etc/inittab. With this option, the initial login prompt does *NOT* clear away late bootup output, including useful error messages... # TERMINALS c1:12345:respawn:/sbin/agetty --noclear 38400 tty1 linux c2:2345:respawn:/sbin/agetty 38400 tty2 linux c3:2345:respawn:/sbin/agetty 38400 tty3 linux c4:2345:respawn:/sbin/agetty 38400 tty4 linux c5:2345:respawn:/sbin/agetty 38400 tty5 linux c6:2345:respawn:/sbin/agetty 38400 tty6 linux -- Walter Dnes <waltdnes@waltdnes.org> I don't run "desktop environments"; I run useful applications ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [gentoo-user] Anybody got a Gentoo system working under uclibc? 2016-04-28 23:07 ` [gentoo-user] " waltdnes @ 2016-04-29 3:52 ` Corbin 0 siblings, 0 replies; 10+ messages in thread From: Corbin @ 2016-04-29 3:52 UTC (permalink / raw To: gentoo-user On 04/28/2016 06:07 PM, waltdnes@waltdnes.org wrote: > On Thu, Apr 28, 2016 at 08:33:01AM -0500, Corbin wrote >> >> >> Questions ... if you will permit : >> >> Are you saying that in "make.conf" you set INPUT_DEVICES="evdev" and did >> a test compile run? >> The emerge you tried ... was it "xorg-base/xorg-x11"? >> Or did you try a meta package for a desktop? > > xorg-server, intending to add ICEWM later. > >> If a hard dependency link between Xorg server -> xf86-input-keyboard >> exists, this will never work. I have no idea at this point if this is >> true. What I have been reading suggests that the xf86-input keyboard and >> mouse libs are being phased out. >> >> With that call ?error? ... Xorg may be an impossible goal / waste of >> time on uClibc. > > I tried INPUT_DEVICES="evdev mouse keyboard", which probably caused > the problem. With INPUT_DEVICES="evdev" and VIDEO_CARDS="vesa fbdev", > things build OK. Whether they'll work, I don't know, but at least it > builds. ICEWM appears to be building too. Actually, ICEWM has a > "(uclibc)" USE flag which is automatically hard-invoked or hard-masked > depending on whether or not the system uses uclibc. > > James points to https://wiki.gentoo.org/wiki/Project:Hardened_uClibc > which mentions "Lilblue"... > >> Continued developments in uClibc have made it increasingly suitable >> for systems like Lilblue, our security-enhanced, fully featured >> XFCE4 desktop, amd64 system built on uClibc. > > WHEEEEEEEEEE! IT WOIKS! Whilst I was typing away, the ICEWM and > xterm builds finished. I've brought up the basic ICEWM window with 4 > work areas. The fixes to my problem were... > > INPUT_DEVICES="evdev" > VIDEO_CARDS="vesa fbdev" > enable "udev" flag in make.conf (actually it runs on eudev, but...) > > I've never used "evdev" before, so I was not familiar with how to set > it up. It's a bit disappointing, because evdev *DEMANDS* udev, so I can > forget about switching in busybox's mdev for udev. > > BTW, I stumbled over "1 weird little tip" on the internet, to make > debugging easier for many bootup problems, not just uclibc. Add the > "--noclear" option to the first console in /etc/inittab. With this > option, the initial login prompt does *NOT* clear away late bootup > output, including useful error messages... > > # TERMINALS > c1:12345:respawn:/sbin/agetty --noclear 38400 tty1 linux > c2:2345:respawn:/sbin/agetty 38400 tty2 linux > c3:2345:respawn:/sbin/agetty 38400 tty3 linux > c4:2345:respawn:/sbin/agetty 38400 tty4 linux > c5:2345:respawn:/sbin/agetty 38400 tty5 linux > c6:2345:respawn:/sbin/agetty 38400 tty6 linux > Great! Glad to hear it. Might want to update the forum posting ... or I can update the one I made with this info, if you want. It might save someone else a lot of head scratching :) ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [gentoo-user] Anybody got a Gentoo system working under uclibc? 2016-04-27 14:12 [gentoo-user] Anybody got a Gentoo system working under uclibc? waltdnes 2016-04-27 17:25 ` Corbin @ 2016-05-01 5:41 ` waltdnes 1 sibling, 0 replies; 10+ messages in thread From: waltdnes @ 2016-05-01 5:41 UTC (permalink / raw To: Gentoo Users List On Wed, Apr 27, 2016 at 10:12:53AM -0400, waltdnes@waltdnes.org wrote > I'm testing uclibc in a 32-bit QEMU Gentoo VM. I want to eventually > install on my ancient 32-bit-only netbook. I'm running into problems. > See https://forums.gentoo.org/viewtopic-p-7912494.html > > To summarize... > * static IP works for eth0, but dhcpcd fails, which is bad news for a > laptop or netbook I finally figured out dhcp. 1) net-misc/dhcp needs to be emerged, because it provides /sbin/dhclient 2) /etc/conf.d/net has to read smilar to... config_eth0="dhclient" Replace "eth0" with whatever the actual interface name is. -- Walter Dnes <waltdnes@waltdnes.org> I don't run "desktop environments"; I run useful applications ^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2016-05-01 5:41 UTC | newest] Thread overview: 10+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2016-04-27 14:12 [gentoo-user] Anybody got a Gentoo system working under uclibc? waltdnes 2016-04-27 17:25 ` Corbin 2016-04-27 21:31 ` waltdnes 2016-04-28 0:30 ` Corbin 2016-04-28 3:16 ` waltdnes 2016-04-28 13:33 ` Corbin 2016-04-28 16:14 ` [gentoo-user] " James 2016-04-28 23:07 ` [gentoo-user] " waltdnes 2016-04-29 3:52 ` Corbin 2016-05-01 5:41 ` waltdnes
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox