My build scripts use su to install chapter 6, using the package user management
system.
Not now they don't!
coreutils no longer installs su. Instead util-linux-2.22-? will provide su.
But util-linux will need a PAM header file to build su.
So is this end of su as we know (and love?) it in LFS
>
> LFS uses the shadow version of su.
>
Yes, of course - I'd forgotten that shadow provides it.
I changed my script so that at the end of chapter 5 it compiles shadow and then
instead of make install, it does:
cp src/su /tools/bin/
I compiled shadow after xz in chapter 5 and it configures
>> I already understand the problem. Thank.
>>
>As do I now that I know that the su-tools I thought may have come packaged
>with the helper >scripts didn't, it was from coreutils and is now awol
>well, like I said it's been a while, but I have acquired a couple of old Dell
>d600 laptops so w
>True, but easy with a script, something like
> install $(which su) /tools/bin/su
Yes. I'm going to hang on to the coreutils-8.17 version for the moment because
I know that one works.
>Shot in the dark,.but are you using login flag?
> su -l user
Not as far as I'm aware.
>> What is l
Worth noting I think that util-linux-2.22 also installs /bin/kill and
/usr/share/man/man1/kill.1.
These are subsequently overwritten by procps.
May be should do --disable-kill in util-linux, if that's possible.
I haven't checked to see.
jb.
--
http://l
util-linux-2.22 installs /usr/bin/utmpdump and /usr/share/man/man1/utmpdump.1
These are subsequently overwritten by sysvinit.
jb.
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the abov
I'm building SVN-20100905
On boot, I get lots of complaints that
/run/var/bootlog doesn't exist when running /lib/lsb/init-functions
Should init-functions create bootlog or am I missing its installation somewhere?
jb.
--
http://linuxfromscratch.org/ma
>
> Did you create /run in Section 6.5?
Yes
> Does /etc/fstab have
>
> tmpfs /run tmpfsdefaults0 0
Yes
>
> Do you have CONFIG_TMPFS=y in your kernel configuration?
>
No, that's the culprit. I don't usually look in pseudo filesystems.
I think in the pas
Not sure if this is the right place for this - I'm sure you'll let me know if
it isn't.
I had the following problems with a systemd setup:
glibc-2.17-fhs-1.patch
is in http://www.linuxfromscratch.org/patches/downloads/glibc/
and not
http://www.linuxfromscratch.org/patches/lf
>
> The libcap2_2.22-orig.tar.gz link is broken.
>
> You have libcap_2.22.orig.tar.gz
>
> It should be:
>
> http://ftp.de.debian.org/debian/pool/main/libc/libcap2/libcap2_2.22.orig.tar.gz
> as in blfs.
>
In fact it would be very helpful if you renamed libcap to be libcap2 to make it
consistent w
I installed lfs with systemd and encountered no problems at all - to my mild
surprise.
Encouraged by this, I've been adding blfs packages and have now come up against
the question of starting various daemons.
For example, I installed mysql and used the mysql.service file from Arch-Linux
as a g
>
> $ cat /lib/systemd/system/mysqld.service
> [Unit]
> Description=MySQL Server
> After=network.target
>
> [Service]
> User=mysql
> Group=mysql
> ExecStart=/usr/bin/mysqld --pid-file=/var/run/mysqld/mysqld.pid
> Restart=always
> PrivateTmp=true
>
> [Install]
> WantedBy=multi-user.target
>
> $ cat
I've built lfs-systemd-20130324 successfully and, encouraged by this, I went on
to add some blfs packages.
When I login I don't appear to have a session set up.
When I do:
loginctl list-seats
I get seat0 - 1 seat listed
but when I do:
loginctl list-sessions
it says 0 sessions listed.
Which
>
> You need to add:
> session optional pam_systemd.so
> to the /etc/pam.d/system-session.
Excellent - thank you, That single addition fixed it. I'm writing this using
Firefox, having successfully logged-in.
>
> Also if you are going to use a graphical login manager, you need to modify
> thei
> > > You need to add:
> > > session optional pam_systemd.so
> > > to the /etc/pam.d/system-session.
> >
> > Excellent - thank you, That single addition fixed it. I'm writing this
> > using Firefox, having successfully logged-in.
>
> Also, as done with ConsoleKit, you need the pam_loginuid module,
In the systemd branch, you have
LFS-Network-Scripts (20130518) -
22 KB:
but it doesn't appear to be mentioned at all in the networking section (may be
I missed it)
which is probably just as well as the link:
http://www.linuxfromscratch.org/lfs/downloads/systemd/lfs-network-scrip
In Chapter 6 in systemd, glibc has this patch:
patch -Np1 -i ../glibc-2.17-fhs-1.patch
This is not in the svn version.
Is the patch necessary or can it be removed to be consistent with svn?
In chapter 6, the configure for systemd-204 needs this addition:
--with-dbuspolicydir=/etc/dbus-1/system
I've been playing with the systemd version of lfs, mainly as a learning
exercise.
For some reason the interface wlp3s0 takes an age to appear.
It seems that the kernel is not asking for the iwlwifi firmware for about 2
minutes 10 seconds after the start of the boot.
The only possibly relevant
> Subject: [lfs-dev] lfs-systemd: interface wlp3s0 not available
>
> I've been playing with the systemd version of lfs, mainly as a learning
> exercise.
>
> For some reason the interface wlp3s0 takes an age to appear.
>
> It seems that the kernel is not asking for the iwlwifi firmware for about 2
>
> Am, --with-firmware-path="/lib/firmware/updates:/lib/firmware" configure
> switch should enable udev firmware loader iirc. But your log says:
> "systemd-udevd[1579]: IMPORT{builtin}: 'firmware' unknown", which
> confuses me. From what I know, for newer kernel versions, systemd/udevd
> won't use
>
> The relevant bit of the .config file:
>
> CONFIG_FW_LOADER=y
> CONFIG_FIRMWARE_IN_KERNEL=y
> CONFIG_EXTRA_FIRMWARE="iwlwifi-6000-4.ucode"
> CONFIG_EXTRA_FIRMWARE_DIR="/lib/firmware"
> CONFIG_FW_LOADER_USER_HELPER=y
> CONFIG_DEBUG_DRIVER=y
> CONFIG_DEBUG_DEVRES=y
>
> So now I have to play around
>
>> I put the firmware file in the firmware dir of the kernel src tree
>> and changed
>> CONFIG_EXTRA_FIRMWARE_DIR accordingly and recompiled.
>
>
> As the example:
>
> CONFIG_FIRMWARE_IN_KERNEL=y
> CONFIG_EXTRA_FIRMWARE='iwlwifi-6000-4.ucode'
> CONFIG_EXTRA_FIRMWARE_DIR='/lib/firmware'
>
> Anoth
>>
> Relase notes are at
> http://article.gmane.org/gmane.comp.lib.glibc.alpha/34237
>
> The pt_chown program is now disabled by default (CVE-2013-2207,
> glibc bugzilla #15755 - incorrectly granting access to another
> user's pseudo-terminal). I assume we can just remove it from the
> list of prog
>Not going to be just konsole; it'll affect any program that tries to
>grantpt() as part of the PTY allocation handshake. xterm, aterm, sshd,
>expect, you name it.
That's right - for e.g.
if you chroot into lfs and try and open a teminal in the host, grantpt will
fail.
So if you get this 'featu
shadow installs /sbin/nologin
util-linux-2.24 now also installs a version of nologin
Do you want the shadow version or the util-linux version?
There is a --disable-nologin in the configure script of util-linux if you want
the shadow version.
As the book stands now, and if you install as root,
Subject: [lfs-dev] util-linux-2.24 and nologin
shadow installs /sbin/nologin
util-linux-2.24 now also installs a version of nologin
Do you want the shadow version or the util-linux version?
There is a --disable-nologin in the configure script of util-linux if you want
the shadow version.
> Date: Wed, 20 Nov 2013 14:25:16 -0600
> John Burrell wrote:
> > Subject: [lfs-dev] util-linux-2.24 and nologin
> >
> > shadow installs /sbin/nologin
> >
> > util-linux-2.24 now also installs a version of nologin
> >
> > Do you want the shadow ver
I'm trying to create some package archives.
I've started with linux_headers. I then did glibc and used
--with-headers= so it picked up the correct
headers.
Now I need to build binutils and make sure that it sees the correct toolchain -
I created the specs file and edited it so it points to the
.
>>
> I do not quite understand what you have done. Your first build was linux
> headers then glibc? Without doing anything with gcc?
Well I'm building the archive files using a machine running LFS so gcc is
version 4.8.2
>If so, you cannot expect
> gcc to know about the glibc you built. All t
.
> As how to automate the setting of rpath under gcc, I guess you can with
> the specs, but I have never done it.
these lines in the specs file:
*link_libgcc:
%D
can be changed to:
*link_libgcc:
%D -rpath /lib/%M
Apparently %M expands to either ../lib or ../lib64 depending on 32 or 64 bit
ar
>> Now I need to build binutils and make sure that it sees the correct
>> toolchain -
>
> You have it backwards. Binutils, then gcc, then glibc. Not glibc,
> then binutils.
>
> What exactly are you going after.
It's an experiment. I wondered whether one could build the four core packages
in the
> Do not know how you progressed along this way. Maybe, you could try
> first to see whether you can get gcc to do what you want by passing
> flags to the command line, then try to automate passing those flags
> using specs.
>
> Maybe the -L flag is enough, but you have also to specify the "crtxx"
When I try and update the dynamic linker, /usr/lib/ld-2.18.so in my case,
I get the message 'Text file busy'
and when I then access the chroot window I get a seg fault.
If I close this window, open a new window, unmount the lfs disks, remount them
and enter chroot, everything works fine and the
>
>> When I try and update the dynamic linker, /usr/lib/ld-2.18.so in my
>> case,
>>
>> I get the message 'Text file busy'
>>
>> and when I then access the chroot window I get a seg fault.
>
> You aren't clear,
As mud - sorry, I'll try harder next time.
>
> /usr/lib being your host system that yo
>
>> When I try and update the dynamic linker, /usr/lib/ld-2.18.so in my
>> case,
>
> When why where and why?
2 whys, so why is really important :))
and there is no answer to that. The universe just is.
More seriously though, this is part of my previous post on updating the 4 core
packages in L
>...Also, it is *not* possible to replace either /lib/ld-linux.so.2 or
>/lib64/ld-linux-x86-64.so.2 (...and yes, I use those paths intentionally,
>since those are the ABI standard) on a running system unless you replace both
>it and /lib{,64}/libc.so.6 from the same process.
They are from the same
>
> My secret is that I use /tools toolchain for upgrading Glibc when
> necessarrry. Just point the PATH to /tools/bin before starting the
> upgrade and it should be done. Overwriting a file should get you the
> same result for many libraries which are used by the utility replacing
> the linker (cp
>
> As Pierre mentioned, try creating tarballs for the packages and
> installing with tar instead of cp or install.
>
The reason it seg faults is because I've chrooted to LFS, so the solution is
not to chroot.
I mount the LFS disks and the required file systems and update the packages
from the t
>>> (And you really want to "replace" them with an unlink() and copy, or copy
>>> to a temp file and unlink()/link(), sequence.)
>>
>> So the sequence I should follow is:
>>
>> -rwxr-xr-x 1 root root 156051 Dec 11 08:40 ld-2.18.so
>> lrwxrwxrwx 1 root root 10 Dec 11 08:40 ld-linux-x86-64.so.2 -
> From: john_burr...@hotmail.com
> To: lfs-dev@linuxfromscratch.org
> Date: Wed, 18 Dec 2013 17:49:31 +
> Subject: Re: [lfs-dev] Updating the dynamic linker
>
(And you really want to "replace" them with an unlink() and copy, or copy
to a temp f
> Hello there,
>
> After many months since lfs systemd branch was created, I am pleased to
> announce that I consider it being usable for everyone that wants to use it.
>
> I've managed to add missing bits to the branch today and doing that I've
> finished what it was necessarry to bring it up to p
.
>>>
>>> You can clone lfs-systemd book using svn co
>>> svn://svn.linuxfromscratch.org/LFS/branches/systemd
>>
>> A small request re. the svn cloning -
>>
>> When the book is downloaded using svn, it creates systemd/BOOK
>> which contains all the files except one dir.
>>
>> The directory .svn doe
I'm building the systemd version of LFS
When I configure systemd-208 with libcap-2.23 I get:
checking sys/capability.h usability... no
checking sys/capability.h presence... no
checking for sys/capability.h... no
configure: error: *** POSIX caps headers not found
The file /usr/include/sys/capabil
>
> Systemd branch doesn't yet have libcap-2.23 because of that reason. I
> was waiting for fix to land in BLFS before I could merge the changes.
> Revert to 2.22 or use patch from BLFS to fix the issue.
>
> Also, Follow book, book good.
>
Yes, but have to go off-piste for the real excitement - an
I periodically recreate the /tools dir so I can update chapter 6.
I get this when installing util-linux in chapter 5:
/tools/bin/mkdir -p '/usr/lib/systemd/system'
/tools/bin/install -c -m 644 misc-utils/uuidd.service misc-utils/uuidd.socket
'/usr/lib/systemd/system'
/tools/bin/install: cannot
> Date: Thu, 23 Jan 2014 22:46:45 +0100
> From: kre...@email.com
> To: lfs-dev@linuxfromscratch.org
> Subject: Re: [lfs-dev] util-linux-2.24.1 in chapter05
>
> On 01/23/2014 10:40 PM, John Burrell wrote:
>> I periodically recreate
I just noticed this:
x86_64) ln -sv lib /lib64 &&
ln -sv lib /usr/lib64 &&
As far as I'm aware this is the first instance of && in LFS.
Is this the thin end of a wedge or a slip?
Personally, I would prefer you to remove it to leave LFS consistent in this
regard.
jb.
As 210 is the latest version, I thought I'd try it in the systemd version of
LFS.
Without gudev it doesn't install:
/usr/lib/pkgconfig/libsystemd-daemon.pc
/usr/lib/pkgconfiglibsystemd-journal.pc
/usr/lib/pkgconfig/libsystemd-login.pc
Without these files dbus-1.8.0 won't recognise that systemd
>>
>
> systemd version of lfs has systemd-208. 210 is in terrible shape.
> systemd-209 merged those libraries into one library - libsystemd.so and
> all apps now link against it. --enable-compat-libs enable those
> libraries and their pkg-config files, but those libraries shouldn't be
> linked agai
>> Can I try your patch please?
>>
>
> Archlinux also enables compat-libs, but as I said, they're useless on
> LFS since you don't actually link against them - they're necessary if
> you have programs linked to it and are going through a transition to the
> new lib.
>
> http://www.linuxfromscratch.
>
> I have noticed that we don't seem to cover firmware installation
> anywhere in LFS. The trick is that some modern network cards and most
> wireless ones require firmware to function. Radeon cards also require
> firmware (most of them does iirc) but we cover that in BLFS. It doesn't
> stop there
I get the folowing error doing make in
autoconf-2.59 in lfs-gcc4-20050728
Can't locate Data/Dumper.pm in @INC.( @INC
contains: ---blah blah ...
/test/../lib/Autom4te/C4che.pm line 35
BEGIN failed --compilation aborted at
/sources/autoconf-2.59//test/../lib/Autom4te/C4che.pm line 35.
I as
52 matches
Mail list logo
