Bug related to this issue: https://bugzilla.redhat.com/show_bug.cgi?id=1360131
Now dracut only attempts to copy GlobalKnownHostsFile while generating kdump's
initramfs. This method will cause kdump's failure if users set customized
UserKnownHostsFile in /etc/ssh/ssh_config. This patch simply attempts to copy
those files too while going through /etc/ssh/ssh_config. Note that we need to
make sure ~/foo will be copied as /root/foo in kdump's initramfs.
Extend "rd.memdebug" to "4", and "make_trace_mem" to "4+:komem".
Add new "cleanup_trace_mem" to cleanup the trace if active.
Signed-off-by: Xunlei Pang <xlpang@redhat.com>
The current method for memory debug is to use "rd.memdebug=[0-3]",
it is not enough for debugging kernel modules. For example, when we
want to find out which kernel module consumes a large amount of memory,
"rd.memdebug=[0-3]" won't help too much.
A better way is needed to achieve this requirement, this is useful for
kdump OOM debugging.
The principle of this patch is to use kernel trace to track slab and
buddy allocation calls during kernel module loading(module_init), thus
we can analyze all the trace data and get the total memory consumption.
As for large slab allocation, it will probably fall into buddy allocation,
thus tracing "mm_page_alloc" alone should be enough for the purpose(this
saves quite some trace buffer memory, also large free is quite unlikey
during module loading, we neglect those memory free events).
The trace events include memory calls under "tracing/events/":
kmem/mm_page_alloc
We also inpect the following events to detect the module loading:
module/module_load
module/module_put
Since we use filters to trace events, the final trace data size won't
be too big. Users can adjust the trace buffer size via "trace_buf_size"
kernel boot command line as needed.
We can get the module name and task pid from "module_load" event which
also mark the beginning of the loading, and module_put called by the
same task pid implies the end of the loading. So the memory events
recorded in between by the same task pid are consumed by this module
during loading(i.e. modprobe or module_init()).
With these information, we can record the rough total memory(the larger,
the more precise the result will be) consumption involved by each kernel
module loading.
Thus we introduce this shell script to find out which kernel module
consumes a large amount of memory during loading. Use "rd.memdebug=4"
as the tigger.
After applying this patch and specifying "rd.memdebug=4", during booting
it will print out something extra like below:
0 pages consumed by "pata_acpi"
0 pages consumed by "ata_generic"
1 pages consumed by "drm"
0 pages consumed by "ttm"
0 pages consumed by "drm_kms_helper"
835 pages consumed by "qxl"
0 pages consumed by "mii"
6 pages consumed by "8139cp"
0 pages consumed by "virtio"
0 pages consumed by "virtio_ring"
9 pages consumed by "virtio_pci"
1 pages consumed by "8139too"
0 pages consumed by "serio_raw"
0 pages consumed by "crc32c_intel"
199 pages consumed by "virtio_console"
0 pages consumed by "libcrc32c"
9 pages consumed by "xfs"
From the print, we see clearly that "qxl" consumed the most memory.
This file will be installed as a separate executable named "tracekomem"
in the following patch.
Signed-off-by: Xunlei Pang <xlpang@redhat.com>
Some crashkernel targets still use legacy NTLM auth, which
require those (bsc#869496). This patch enumerates all dependent
hash algorithems, because even though most of them are probably
compiled in, older ones (e.g. md4 and arc4) usually aren't.
Contrary to previous intel pinctrl drivers, the cherryview driver can be
and usually is built as a module. However, it sets up the SDIO pinout
so sdhci can make use of the SD card reader, which may subsequently
hold a root file system on a card (bsc#998440).
IMA validates file signatures based on the security.ima xattr. As of
Linux-4.7, instead of cat'ing the IMA policy into the securityfs policy,
the IMA policy pathname can be written, allowing the IMA policy file
signature to be validated.
This patch first attempts to write the pathname, but on failure falls
back to cat'ing the IMA policy contents .
Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Signed-off-by: Mimi Zohar <zohar@linux.vnet.ibm.com>
bnx2x can take no longer than 3 seconds to initialize the link in some setups
which can cause fipvlan to fail and thus the fcoe interface(s) won't be
created.
Add another 10 seconds to give the link enough time to initialize.
Signed-off-by: Chad Dupuis <chad.dupuis@cavium.com>
This is a further improvement on PR #105 which re-adds support for nfs:// urls to root=live:nfs://... Symptoms prior to applying this patch are that sysroot.mount times out when following the nfs:// syntax for the path to the live image. An additional case is added to livenet-generator to support the nfs protocol.
ip=2620:0052:0000:2220:0226:b9ff:fe81:cde4::[2620:0052:0000:2220:0000:0000:0000:03fe]:64::ibft0:none
should be
ip=[2620:0052:0000:2220:0226:b9ff:fe81:cde4]::[2620:0052:0000:2220:0000:0000:0000:03fe]:64::ibft0:none
https://bugzilla.redhat.com/show_bug.cgi?id=1322592#c19
(cherry picked from commit b8e6c051c6)
use inst() instead of inst_simple()
/etc/pki/tls/certs/ca-bundle.crt is a symlink to
../../ca-trust/extracted/pem/tls-ca-bundle.pem
with inst() we install the original file also.
https://bugzilla.redhat.com/show_bug.cgi?id=1341280
(cherry picked from commit 1b23c6c65c)
IPv6 addresses should be specified in brackets so that the
ip= scanning code doesn't get confused.
References: bnc#887542
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.com>
Both 'utmp' and 'root' groups are mentioned in tmpfiles.d/systemd.conf
and as such should be included.
It's probably better to have something equiv to inst_rule_group_owner()
for udev rules which parses out users and groups and adds them to the
passwd/group db respectively.
Could also rely on sysusers but as the initramfs is static in this
sense, it's more efficient to pre-define the users IMO.
This will bundle clock drivers into the initramfs on arm
Tested on ti dm8148-t410 where adpll is needed on 4.6+ kernel
This will avoid to rely on (maybe broken) bootloader clocks.
Theses modules are also usually loaded early. Having them bundled into
the initramfs will avoid lot of deferred probes and others delay.
Signed-off-by: Nicolas Chauvet <kwizart@gmail.com>
If journald.conf already contains Storage=persistent, journald will
write to /var/log/journal/, which ends up at /run/initramfs/log/journal/
after switching root. We want to make sure early boot logs are written
to /run/log/journal/ so they can be flushed to /var/log/journal/ after
switching root.
Users can pass the DNS information throught "nameserver=" cmdline,
there maybe duplicated inputs.
"/etc/resolv.conf" have some restrictions on the number of DNS items
effective, so make sure that this file contains no duplicated items.
We achieve this by simply making the file have no duplicated lines.
Signed-off-by: Xunlei Pang <xlpang@redhat.com>
I met a problem when passing kdump dns to dracut via "nameserver=x.x.x.x",
the dns I provided didn't appear in the "/etc/resolv.conf".
After some debugging, found that when setup dhcp DNS, in setup_interface()
and setup_interface6(), it has:
echo "search $search $domain" > /tmp/net.$netif.resolv.conf
So if "$search $domain" isn't NULL(this is ture in my kdump environment),
the dns contents(that is, dns1, dns2, nameserver) in "ifup" before dhcp
will be discarded.
This patch addresses it by handling dhcp first. In fact this is also the
way the NetworkManager in 1st kernel works.
Signed-off-by: Xunlei Pang <xlpang@redhat.com>
Avoid keymap/font not found error when loadkeys/setfont
are compiled with the default data directory path.
Signed-off-by: Guido Trentalancia <guido@trentalancia.net>
- use local variables with _
- use associative array for the kernel modules
- install emergency hook even in the systemd case
- follow device path until /sys is reached
- set kernel version for modprobe checking
If the initramfs was built with prefix=/run/... /run can't be mounted
with noexec, otherwise no binary can be run.
Guard against it by looking where /bin/sh is really located.
Trigger the acpi subsystem. This will ensure hv_vmbus gets loaded before
plymouth is started, which will make the graphics device become
available before plymouth is started too (and the keyboard ! which might
also be important for plymouth in some setups).
https://bugzilla.redhat.com/show_bug.cgi?id=1218130
(cherry picked from commit d2846fdcce9b8de0edecdf0e06a4b86fc8de542c)
It is expected that a watchdog module will disable an active watchdog when
its probe is called ie, when it is loaded. So an early load of the module
will help to disable it earlier.
This can be helpful in some corner cases where kdump and watchdog daemon
both are active.
Testing:
-- When watchdog kernel modules were added
# dracut --no-hostonly initramfs-test.img -a watchdog
# lsinitrd initramfs-test.img -f etc/cmdline.d/00-watchdog.conf
rd.driver.pre=iTCO_wdt,lpc_ich,
Signed-off-by: Pratyush Anand <panand@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Harald Hoyer <harald@redhat.com>
Recently following patches have been added in upstream Linux kernel, which
(1) fixes parent of watchdog_device so that
/sys/class/watchdog/watchdogn/device is populated. (2) adds some sysfs
device attributes so that different watchdog status can be read.
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=6551881c86c791237a3bebf11eb3bd70b60ea782http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=906d7a5cfeda508e7361f021605579a00cd82815http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=33b711269ade3f6bc9d9d15e4343e6fa922d999b
With the above support, now we can find out whether a watchdog is active or
not. We can also find out the driver/module responsible for that watchdog
device.
Proposed patch uses above support and then adds module of active watchdog
in initramfs generated by dracut for hostonly mode. Kernel module for
inactive watchdog will be added as well for none hostonly mode.
When an user does not want to add kernel module, then one should exclude
complete dracut watchdog module with --omit.
Testing:
-- When watchdog is active watchdog modules were added
# cat /sys/class/watchdog/watchdog0/identity
iTCO_wdt
# cat /sys/class/watchdog/watchdog0/state
active
# dracut --hostonly initramfs-test.img -a watchdog
# lsinitrd initramfs-test.img | grep iTCO
-rw-r--r-- 1 root root 9100 Feb 24 09:19 usr/lib/modules/.../kernel/drivers/watchdog/iTCO_vendor_support.ko
-rw-r--r-- 1 root root 19252 Feb 24 09:19 usr/lib/modules/.../kernel/drivers/watchdog/iTCO_wdt.ko
-- When watchdog is inactive then watchdog modules were not added
# cat /sys/class/watchdog/watchdog0/state
inactive
# dracut --hostonly initramfs-test.img -a watchdog
# lsinitrd initramfs-test.img | grep iTCO
-- When watchdog is inactive, but no hostonly mode, watchdog modules were added
# cat /sys/class/watchdog/watchdog0/state
inactive
# dracut --no-hostonly initramfs-test.img -a watchdog
# lsinitrd initramfs-test.img | grep iTCO
-rw-r--r-- 1 root root 9100 Feb 24 09:19 usr/lib/modules/.../kernel/drivers/watchdog/iTCO_vendor_support.ko
-rw-r--r-- 1 root root 19252 Feb 24 09:19 usr/lib/modules/.../kernel/drivers/watchdog/iTCO_wdt.ko
Signed-off-by: Pratyush Anand <panand@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Harald Hoyer <harald@redhat.com>
When systemd is present, let it manage watchdog feed.
Signed-off-by: Pratyush Anand <panand@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Harald Hoyer <harald@redhat.com>
How to reproduce:
host# ./dracut.sh -o 'dracut-systemd systemd systemd-initrd' --local -f ./initramfs.img
host# qemu-system-x86_64 -initrd ./initramfs.img \
-append 'root=/dev/sda1 rd.cmdline=ask rd.hostonly=0' \
...
Enter additional kernel command line parameter (end with ctrl-d or .)
> rd.break
> .
...
There is no "Break before switch_root"
...
Signed-off-by: Evgeny Vereshchagin <evvers@ya.ru>
* Multipath device names only start with the mpath-prefix if the option
use_userfriendly_names is set true in /etc/multipath.conf and if user
has not set any aliases in the said file. Thus the for-loop should go
through all files in /dev/mapper/, not just ones starting with 'mpath'
* Bash is perfectly capable to extend `/dev/mapper/*` notation without a
need to pass it to an external ls
* Changed the function to use a local variable $_dev instead of the
global $dev, which seemed to be the original intention as the local
_dev was defined but not used
crypt/parse-crypt.sh generate initqueue job which always call
dev_unit_name() with an argument beginning with "-". This results
in the following error:
dracut-initqueue[307]: + systemd-escape -p -cfb4aa43-2f02-4c6b-a313-60ea99288087
dracut-initqueue[307]: systemd-escape: invalid option -- 'c'
Add a systemd generator for root=nbd:.. so that systemd has a correct
sysroot.mount unit.
Use export names instead of port numbers, because port number based
exports are deprecated and were removed.
rename iface_has_link() to iface_has_carrier() to clarify usage
Only assign static "wildcard interface" settings, if the interface has a
carrier.
If the interface name was specified with a name, do not do carrier
checking for static configurations.
8b5ee88ff6 removed the check for SQUASHED,
assuming, that the if clause above was the only place, where SQUASHED is
set.
This patch reverts to the old logic, because SQUASHED can be set
earlier.
Persistent, uncompressed live installations can avoid overlays with a new rd.live.overlay=none flag.
Non-persistent rd.live.ram boots can also take advantage of persistent home.img filesystems.
Resolves issues where systemd attempts to boot a live URL as an NFS mount.
This patch uses systemd's generator arg[2] to generate an early sysroot.mount
that preempts systemd-fstab-generator.
See https://bugzilla.redhat.com/show_bug.cgi?id=1280103
When using rd.zfcp= parameter in generic.prm file, wrong format
parameters will prevent the zfcp driver to add the correct SCSI
disk. dracut should die when a wrong rd.zfcp= parameter supplied.
Signed-off-by: Zhiguo Deng <bjzgdeng@linux.vnet.ibm.com>
znetconf command is part of s390utils-base package. It depends on
awk and getopt.
This patch is used to fix the following error:
dracut:/#
znetconf -c
/usr/sbin/znetconf: line 70: awk: command not found
/usr/sbin/znetconf: line 1138: getopt: command not found
Signed-off-by: Mei Liu <liumbj@linux.vnet.ibm.com>
all_ifaces_up() is true, if all interfaces are up.
all_ifaces_setup() is true, if all interfaces are up and the gateways
and nameserver are setup.
(cherry picked from commit 63e75dc4cd)
By default, dracut only builds in dm-service-time into the initramfs as
that is the default multipath.conf path selector. If the user changes
the path selector to "round robin" on the fly and runs dracut, multipath
does not find any paths on boot and the user will be dropped into a
shell.
Apparently, in RHEL7 dracut defaults to "hostonly" mode, i.e. modules
not currently in use at the time dracut runs do not get built into
initramfs. This is definitely one case where this doesn't work. A change
to reconfigure multipath probably should not render the system
unbootable.
https://bugzilla.redhat.com/show_bug.cgi?id=1195392
(cherry picked from commit f11d7a81e2)
- check if modprobe was successful
- add a timeout for /sys/firmware/edd
- only remove the module, if it was loaded by the script
(cherry picked from commit 34203d03c0)
rd.net.dhcp.retry=<cnt>
If this option is set, dracut will try to connect via dhcp
<cnt> times before failing. Default is 1.
rd.net.timeout.dhcp=<arg>
If this option is set, dhclient is called with "-timeout <arg>".
rd.net.timeout.iflink=<seconds>
Wait <seconds> until link shows up. Default is 60 seconds.
rd.net.timeout.ifup=<seconds>
Wait <seconds> until link has state "UP". Default is 20 seconds.
rd.net.timeout.route=<seconds>
Wait <seconds> until route shows up. Default is 20 seconds.
rd.net.timeout.ipv6dad=<seconds>
Wait <seconds> until IPv6 DAD is finished. Default is 50 seconds.
rd.net.timeout.ipv6auto=<seconds>
Wait <seconds> until IPv6 automatic addresses are assigned.
Default is 40 seconds.
rd.net.timeout.carrier=<seconds>
Wait <seconds> until carrier is recognized. Default is 5 seconds.
(cherry picked from commit d8ad687e1a)
https://bugzilla.redhat.com/show_bug.cgi?id=742564
Once lvm2 starts using /run (bug 742554), it should be no longer
necessary to disable file-based locking in the vgchange call in
fedora-storage-init.
Removing '--sysinit' will make it safe to call LVM operations
concurrently from other units.
The --sysinit is a compound option consisting of:
-> --ignorelockingfailure - not needed anymore, the /run/lock/lvm is
available rw soon in boot process
-> --ignoremonitoring - not needed since /run is available, this would
require the dm-event.service to be run before
fedora-storage-init.service (and new lvm2-activation.service when
deployed). But that's a one line change - I'll have a look whether it
plays well with other services and if yes, I'll commit the change. N.B.:
This has a consequence that all volumes activated on vgchange -ay will
be monitored at the same time they're activated (which is a plus I
think). The lvm2-monitor will just grab all the other volumes not
activated at the time of the boot's vgchange -ay call. But that's not an
issue (for already monitored volumes, calling vgchange --monitor y will
just be a NOOP).
-> --poll n - not needed, we can run the polldaemon as the /run is
available and rw (in case there's unfinished merge or mirror sync from
previous system run, the poll-daemon will be triggered at boot now).
Actually, the polldaemon should be triggered as a service like dmeventd
is, not forked off from the LVM command itself, like from vgchange in
this case - we still need to change this - there's a bug open for this
request already (bug #814857). However, we don't have this feature ready
yet so I need to check whether this is OK with the early boot process
with the current state.
Transient snapshots can take advantage of smaller,
non-persistent metadata structures.
Make the --readonly option explicit rather than inferred
for the readonly_overlay target.
Assure that the live-base target is on the BASE_LOOPDEV.
Temporary snapshots can take advantage of smaller, non-persistent metadata structures.
Make the --readonly option explicit rather than inferred for the readonly_overlay target.
Assure that the live-base target is on the BASE_LOOPDEV.
Some hosting providers need a static route set in order to be
able to reach the default gateway. Be sure to retry adding
the default gateway after setting the static routes.
parse-cmdline sets up an initial initiator-name to let iscsid start.
iscsid is started before doing any iscsistart business.
iscsistart is done with systemd-run asynchrone to do things in
paralllel. Also restarted for every new interface which shows up.
If rd.iscsi.waitnet (default) is set, iscsistart is done only
after all interfaces are up.
If not all interfaces are up and rd.iscsi.testroute (default) is set,
the route to a iscsi target IP is checked and skipped, if there is none.
If all things fail, we issue a "dummy" interface iscsiroot to retry
everything in the initqueue/timeout.
This patch supports loading keys either on the _ima keyring or, as of
Linux 3.17, on the trusted .ima keyring. Only certificates signed by
a key on the system keyring can be loaded onto the trusted .ima keyring.
Changelog:
- Update 98integrity/README
Setting and unsetting the IFS variable is tricky. To be on the
safe side we should always reset the IFS variable to its original
value after parsing.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
For large machine, suppose there's a lot of multipath devices, multipath layer
will use a lot of memory. For kdump kernel memory is very limited thus it causes
oom. To avoid oom, we only add necessary multipath devices in kdump kernel
multipath.conf.
This is done by use mpathconf --allow, a new option which is like whitelist.
Signed-off-by: Dave Young <dyoung@redhat.com>
"ifup -m" was thought to be used by humans in the emergency shell.
Using it programatically shows some other flaw in the execution logic.
Also, "ifup -m" was configuring the interface multiple times on "add"
and "change" uevent, because the "$netif.did-setup" test was not
executed.
This reverts commit cfa365a32d.
The logic in commit cfa365a was added to prevent (odl) lvms from
activating snapshots which should not be activated.
Newer lvms however do this automatically (not enabling an LV if the
the 'k' attribute set), thus we can revert the previous commit.
The fcoe-uefi module should test for EFI firmware when called
in 'hostonly' mode; of no EFI firmware is found then the module
doesn't need to be included.
References: bnc#882412
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
When systemd's crypttab generator parsed crypttab, it tells
systemd about several devices which may not appear until later
in the boot sequence, and which are not needed while dract is running.
This can particularly happen when an md array is encrypted,
and the array is newly degraded so that it doesn't appear until
dracut runs mdraid_start.sh.
This can result in systemd printing warning messages which are
inappropriate.
So tell systemd that the timeout for each of these is zero.
This is involves splitting some functionality out of wait_for_dev()
That function does two things:
- creates 'finished' hooks so that dracut will wait for the device,
and
- sets the systemd timeout for the device to zero, so systemd doesn't
wait.
We only want the second of these for most encrypted devices.
So split that out into a new function set_systemd_timeout_for_dev(),
and call it from parse-crypt.sh
Signed-off-by: NeilBrown <neilb@suse.de>
--
This version fixes the missing redirect from /etc/crypttab
NeilBrown
dracut: Make host only mode more resilient to missing swaps.
This patch set allows swap devices to disappear without cocking up a
host-only initramfs boot.
Although various bits are in place to cancel waiting for the /dev/resume
device (actually a symlink), we don't actually ever wait for it.
Also as the udev rule may create the symlink, silence any errors from
our manual ln -s call from the settled job.
The only reason we add swap devices to host-only mode (added in
dd5875499e) is to allow us to process
resume= arguments passed on the kernel command line when the swap
partition lives on something slightly more complex than a normal
partion (e.g. in an LVM or RAID setup).
By adding the device to host_devs, the necessary LVM and RAID hooks
are added and thus the underlying storage will be initialised OK, and
the 95resume module handles the waiting for the device (via udev rules
creating the /dev/resume symlink).
So ultimately, we do not need to hard-code the waiting for the swap
devices into the initramfs at build time as the waiting part can be
dynamic.
This makes things more resiliant to swap partitions disappearing and
being reformatted etc.
Inspired by a patch by Martin Whitaker on Mageia bug:
https://bugs.mageia.org/show_bug.cgi?id=12305
If crypt-lib.sh is sourced from any #!/bin/sh script, a POSIX shell
quirk is in effect that causes variable assignments to "special
builtins" (such as "shift") to leak to their context. So the buggy
code works even despite the missing semicolons.
But if it is sourced by "bash acting under its own name", i.e. from
any #!/bin/bash script, the quirk is disabled, tty_cmd/tty_prompt are
undefined, and ask_for_password doesn't do anything if plymouth is not
present.
AMD's HSA Linux kernel driver (amdkfd) has been merged into the mainline
kernel since kernel 3.19.
However, for the driver to work, it needs to be included in the default
initramfs image, together with the amd_iommu_v2 driver.
The radeon driver (AMD's kernel graphic driver) calls amdkfd during its
initialization and probing stages. Because radeon is included in the
initramfs image, it tries to initialize amdkfd during the early boot
stages. However, as amdkfd is not present there, it fails.
That doesn't harm radeon operation. However, it disables the HSA
abilities in the machine.
Because of the current design, if you later try to "modprobe amdkfd",
you won't be able to run HSA applications, even though the driver will
be loaded.
https://bugzilla.redhat.com/show_bug.cgi?id=1205222
When 'systemctl daemon-reload' is run, systemd will clean out
/run/systemd/generator and re-run all the generators.
So it is important that the generators always create the required
files.
rootfs-generator.sh currently does *not* create the desired files
if $hookdir/initqueue/finished/devexists-${_name}.sh
exists.
This is not removed by "systectl daemon-reload" so the first time this
generator is run it will do the right thing. Subsequent times it
won't.
This results in incorrect timeouts after "daemon-reload" is run.
So let the existence of each file only guard the creation that file.
Signed-off-by: NeilBrown <neilb@suse.de>
A number of timeout scripts can be registered. If any one of them
makes progress - e.g. assembles a degraded md array - then
the main loop should wait a bit longer rather than pressing forward.
This is particularly important is resume-from-hibernate requires a
degraded md array. Both the script to forcibly assemble the md array
and the script to abort hibernation if the device doesn't appear
are 'timeout' scripts. There needs to be a reasonable delay between
these running.
So: if any script has indicated that progress was made, break of out
the loop and go back to normal waiting.
Signed-off-by: NeilBrown <neilb@suse.de>
mdraid_start is a number of scripts which run after a timeout.
If it makes progress, it should tell the main loop so that it
knows that it is worth waiting a bit longer.
So in that case, create the initqueue/work file which the main loop
checks for.
Signed-off-by: NeilBrown <neilb@suse.de>
When current dracut receives an ip with netmask of 255.255.255.255 via DHCP,
setting the also supplied default gateway fails (because it is obviously not
within the netmask).
The setup with a netmask of /32 is quite common in colocation datacenters
where you don't want the machines of two different customers to directly talk
to each other. At least two of the biggest colocation providers in Germany
(1&1 and Strato) do it that way. NetworkManager supports this kind of setup
and the dhclient-scripts of several distributions too.
In this patch I have implemented a simple approach very similar to what is
found in Debian. The dhclient-script from Fedora uses a more sophisticated
approach, but that relies on the ipcalc utility which would introduce a
dependency on Fedora-initscripts for dracut.
Signed-off-by: Gerd von Egidy <gerd.von.egidy@intra2net.com>
When the 'loop' kernel module isn't loaded in a running system, it gets
excluded from the hostonly initrd. Given that the crypt-loop dracut
module has to be loaded explicitly anyway, it makes sense to always
include the requisite loop kernel module.
When booting with 'rd.info', the 'info' statements in the crypt-loop
module's 'loop_decrypt' function are output to stdout along with the key
that gets piped into the 'cryptsetup' command, which causes the crypt
device unlocking to fail.
There are two possible simple solutions to this problem:
1. Redirect the info messages to stderr (just add '>&2' at the end of
the info statements).
or
2. Remove the info statements altogether.
I have tested both and they both work, but this commit implements #2.
The existing info messages are long (they overflow 80 characters
easily) and redundant (the password prompt clearly indicates what is
happening), and just generally not useful. Given that no one has
reported or fixed this bug in the three years that this module has
existed, no one will miss these info messages.
The commit also changes an error message in the same function to be more
descriptive.
Basic systemd functionality is in 00systemd now.
Switching root and the initrd.target is in 00systemd-initrd.
Dracut additions to the systemd initrd are in 98dracut-systemd.
Previously rd.live.fsimg only supported filesystems residing in
(compressed) archives.
Now rd.live.fsimg can also be used when a squashfs image is used.
This is achieved by extracting the rootfs image from the squashfs and
then continue with the default routines for rd.live.fsimg.
In addition some code duplication got removed and some documentation
got added.
Signed-off-by: Fabian Deutsch <fabiand@fedoraproject.org>
This option changes the underlying mechanism for the overlay in the
dmsquash module.
Instead of a plain dm snapshot a dm thin snapshot is used. The advantage
of the thin snapshot is, that the TRIM command is recognized, which
means that at runtime, only the occupied blocks will be claimed from
memory, and freed blocks will really be freed in ram.
Signed-off-by: Fabian Deutsch <fabiand@fedoraproject.org>
Dracut will generate systemd units for additional devices that should be
brought up during boot, e.g. swap devices. These unit files are broken
symlinks with \ in the filename, e.g.
/etc/systemd/system/initrd.target.wants/dev-disk-by\x2duuid-e6a54f99\x2da4fd\x2d4931\x2da956\x2d1c642bcfee5e.device.
Both the backslash and the broken symlink causes problems for shell
scripts, [ -e "$file" ] isn't enough and read requires the additional -r
argument to not react on the \.
The function 99shutdown/shutdown.sh:_check_shutdown() assumes that
shutdown scripts report success or failure via their return value.
However, "dmsetup remove_all" always reports success, even if some
of the device mappings could not be removed.
I submitted a patch for dmsetup but the lvm2 folks rejected it,
asserting that its behaviour is correct, that "remove_all" should
only be used by developers and that the proper solution would be
to invoke "dmsetup remove" on each device. This does report success
or failure via the return value.
Apart from fixing that issue, this commit also adds the dmsetup
option "--noudevsync". Without it, dmsetup would hang after removal
of a device while trying to communicate with systemd-udevd, which
is no longer running at this shutdown stage.
[harald: replaces backticks with $() ]
The sdhci-pci module is currently not being included in the initramfs,
even though other sdhci modules are. This breaks boot on systems that
rely on this driver to access the root filesystem.
Instead of looking for modules that use sdhci_pltfm_init, look for
sdhci_add_host. I checked 3.18 kernel sources, and this change
does not remove any of the previously-matched SDHCI drivers.
It should result in the addition of sdhci-pci, sdhci-s3c, sdhci-spear
and sdhci-acpi.
For targets with colons in the iSCSI target name:
"iqn.2000-09.com.foo:storage-system.e2000:00000001cm1p1"
the parser was confused with the optional iscsi_iface_name and
iscsi_netdev_name.
This patch reintroduces the old IQN, EUI and NAA parsing and enhances
the fallback parser by checking the LUN for a numerical value.
As reported in https://bugs.mageia.org/show_bug.cgi?id=14799
the xhci module got splitted up in upstream linux merged during
3.18 release cycle:
>From 29e409f0f7613f9fd2235e41f0fa33e48e94544e Mon Sep 17 00:00:00 2001
From: Andrew Bresticker <abrestic@chromium.org>
Date: Fri, 3 Oct 2014 11:35:29 +0300
Subject: xhci: Allow xHCI drivers to be built as separate modules
so we need to adjust 90kernel-modules accordingly.
Signed-off-by: Thomas Backlund <tmb@mageia.org>
rootfs is on nbd
In not hostonly mode, require_binaries will still complain.
If in hostonly mode and the module is explicitly added via -a nbd, then
install() section will still complain later:
dracut-install: ERROR: installing 'nbd-client'
Signed-off-by: Thomas Renninger <trenn@suse.de>
95rootfs-block would not terminate the commandline parameter with
a space or newline, instead it'll rely on the main routine from
dracut.sh to do this.
Which will cause unexpected problems for any modules called
after this.
So terminate the commandline parameters correctly here and remove
the newline from dracut.sh.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
This aligns other places piping cmdline() output to cmdline.d files
with the earlier fix for 95rootfs-block.
Signed-off-by: Thorsten Behrens <tbehrens@suse.com>
Signed-off-by: Thomas Renninger <trenn@suse.de>
If there's a root fallback, at least attempt to have it falling
back to the last root filesystem this system ran off of.
Signed-off-by: Thorsten Behrens <tbehrens@suse.com>
Signed-off-by: Thomas Renninger <trenn@suse.de>
cmdline_journal does not contain linefeeds anymore, so read
silently skipped it altogether.
Signed-off-by: Thorsten Behrens <tbehrens@suse.com>
Signed-off-by: Thomas Renninger <trenn@suse.de>
nvidia driver needs this via modprobe script.
Needs to do change the group after a device node got created.
Add chown instead of chgrp which can also change the owner of a file.
Ask Stefand Dirsch <sndirsch@suse.de> for details.
Signed-off-by: Thomas Renninger <trenn@suse.de>
The module_setup.sh script was missing an '_arch' declaration,
causing network not to be installed on s390.
Signed-off-by: Thomas Renninger <trenn@suse.de>
SUSE is using specific settings for dracut, so add them
to the suse.conf.example file.
Add vi and find to the debug module add some help text to the suse.conf
file when and how to use it.
Signed-off-by: Thomas Renninger <trenn@suse.de>
lldpad is a system-wide process, which must be started only once.
So we should be separate it from fcoe-up, as it might be called
several times.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
DCB & fipvlan can only be called on real devices, not VLAN
ones. So skip any VLAN devices which might been added to the
list of network interfaces.
References: bnc#878583
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
When running with --hostonly-cmdline we should be storing
the current configuration in /etc/cmdline.d so that dracut
will be configure the system automatically.
References: bnc#877288
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
The 'create' sysfs entry has been removed for newer fcoe modules,
so just check if the module directory exists.
References: bnc#877288
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
Due to some obsure reason the IFS parameter is not set correctly
when evaluating get_ibft_mod(). So change the parsing to not rely
on IFS altogether.
References: bnc#886199
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
cmdline
If the rootfs is an nfs mount, also know as nfsroot, add the correct
parameter to the dracut cmdline.
Signed-off-by: Thomas Renninger <trenn@suse.de>
In case of systemd is used the timeout already is set to 180s, compare
with file: modules.d/98systemd/dracut-initqueue.sh
Do the same if systemd is not used, e.g. in kdump case.
Signed-off-by: Thomas Renninger <trenn@suse.de>
When 'initqueue' is called with an invalid command it'll generate
invalid job scripts. This will lead to confusing error messages
later on.
So abort in these cases and print out a warning.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
If kernel cmdline has no root= provided, *and* the initrd carries
a default value, create a systemd mount unit to have it available
in /sysroot
References: bnc#855258
Signed-off-by: Thorsten Behrens <tbehrens@suse.com>
Signed-off-by: Thomas Renninger <trenn@suse.de>
Add more corner cases from systemd's
unit_name_from_path_instance() C function.
Signed-off-by: Thorsten Behrens <tbehrens@suse.com>
Signed-off-by: Thomas Renninger <trenn@suse.de>
When generating units for devices the administrator might
want to use a different timeout than the default.
So implement a new parameter 'rd.timeout' for this.
References: bnc#878770
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
Make sure duplicates of iscsi.initiator vanish.
Only get one rd.iscsi.* paramter value. If getargs is used and several
parameters are parsed, one gets two values separated by whitespace in a
variable which breaks later code and is not suppported.
Signed-off-by: Thomas Renninger <trenn@suse.de>
When installing on an software iscsi root we need to generate
the correct commandline, otherwise the system cannot boot.
References: bnc#880108
Patch: 0093
When using software iscsi we should be using the existing
configuration from the system for the time being.
References: bnc#884768
Patch: 0095
Avoid bad ip route call on empty address
Patch: 0143
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
This fixes up some 95iscsi/module-setup.sh which might print out empty
commandline files.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
bnx2i is using a separate iSCSI offload engine with a separate
MAC address. As a result, the iBFT information is displaying
a MAC address which does not relate to any MAC address from
the network interfaces.
In addition, the iSCSI offload engine works independently on
the NIC, so we do not need to enable the NIC for iSCSI offload
to work.
This patch modifies the automatic iBFT detection to not set
the 'ip=ibft' flag when bnx2i offload is detected.
References: bnc#855747
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Thomas Renninger <trenn@suse.de>
Daniel Molkentin