So far fixtrx was calculating checksum over amount of data matching
partition erase size. It was mostly a workaround of checksum problem
after changing anything in initial TRX content (e.g. formatting JFFS2).
Its main purpose was to make bootloader accept modified TRX. This didn't
provide much protection of flash data against corruption.
This new option lets caller request calculating checksum over a bigger
amount of data. It may be used e.g. to include whole kernel data for
checksum and hopefully make bootloader go info failsafe mode if
something goes wrong.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
We plan to adjust usage of the main buffer to allow reading custom
amount of data for CRC32. This means we need another buffer that will be
always block aligned.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
1) Put sanity checks in one place
2) Respect provided offset
3) Read only as much data as needed for MD5 calculation
Thanks to the last change this is a great speedup and memory saver. On
devices with NAND flash we were allocating & reading about 128 MiB while
something about 8 MiB is enough.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
This avoid long (and unneeded) process of reading all data in case of
running on MTD not containig Seama entity.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
On platforms supporting both: TRX and Seama calling "fixtrx" was
resulting in trying to fix Seama as well.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Remove the public unatteded buildkey from the opkg package to avoid
having hardcoded keys in tree. Use the external keyring package instead
which can be easily updated by users.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
When the same package file is specified multiple times on the opkg install
command line, the name pointer on the argv array becomes stale after the
package structures have been merged, leading to invalid memory accesses
upon install.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Respect user configured mount flags such as "ro" or "sync" when processing
external overlay mount points.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
SVN-Revision: 49223
procd from revision b12bb150ed38a4409bef5127c77b060ee616b860 supports
writing a pidfile. This adds support for setting that parameter with
standard init script hooks:
procd_set_param pidfile /var/run/someprocess.pid
Signed-off-by: Karl Palsson <karlp@etactica.com>
SVN-Revision: 48984
MD5s are shorter than SHA256, so reading buffer has to be bigger to read the
whole hash.
Signed-off-by: Michal Hrusecky <michal.hrusecky@nic.cz>
SVN-Revision: 48917
Extend /etc/config/system with a parameter to set the
default respawn retry for procd launched services that
have respawn enabled.
config service
option respawn_retry -1
All services that don't specify specific respawn parameters
will get their defaults added by procd.sh. If respawn_retry
is specified in /etc/config/system the default retry limit
will be set to this value by procd.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
SVN-Revision: 48915
The image_check currently fails when it cannot read all magic bytes in a
single chunk. But this can happen when the data are read from a pipe. This
currently breaks the openmesh.sh upgrade script with musl because it uses
dd with a blocksize of 1 to copy the image file to the mtd process.
The read can simply be repeated until enough bytes are read for the magic
byte check. It only stops when either an error was returned or 0 bytes were
read.
Signed-off-by: Sven Eckelmann <sven.eckelmann@open-mesh.com>
SVN-Revision: 48891
Opkg now uses sha256 by default and expects them. Making it optionally
understand md5s also and detect md5 sum so we can migrate from configuration
that used md5.
Signed-off-by: Michal Hrusecky <Michal.Hrusecky@nic.cz>
SVN-Revision: 48867
There are time that programs need to be notified of events from
subsystems that are not enumerated in the .json definition, e.g. QEMU
guest agent by default requires /dev/virtio-ports/org.qemu.guest_agent.0
which is a symlink to /dev/vportMpN from virtio-ports subsystem.
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
SVN-Revision: 48799
Seama header has MD5 similarly to TRX and its CRC32. We need to update
it after replacing anything in Seama entity content to make bootloader
accept it.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 48688
Recently TRX checking code was changed to detect Seama format and don't
abort whole writing operation because of it. This isn't a good long-term
solution. It's a poor idea to teach every format handler recognizing all
possible formats. Instead it should be handled in a generic code which
should run check depending on the detected format.
This will also allow further improvements like fixing formats other than
TRX after replacing JFFS2.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 48639
- update to latest version 20160104
- remove cpu dependency (PKGARCH:=all)
- set myself as package maintainer
Signed-off-by: Christian Schoenebeck <christian.schoenebeck@gmail.com>
SVN-Revision: 48271
This fixes diverging executable search paths observed in programs
launched throughn etifd which in turn inherited the search path from
procd early on boot.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 48235
Add 'platform_nand_pre_upgrade' callback to allow platform specific
preparation right before flashinng, when already in ramdisk.
Example uses might be setting correct values for CI_{KERNPART,UBIPART}.
Signed-off-by: Nikolay Martynov <mar.kolya@gmail.com>
SVN-Revision: 47879
Bump ubus to current Git HEAD in order to:
* Fix a null pointer access when user ACLs are loaded into memory
* Fix wrong permission handling for uid=0, gid!=0 peers
* Fix return code for permission defined cases
* Fix socket error handling when the server connection dies
* Make ACL file path configurable
* Remove invalid usages of poll() when using a zero timeout
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 47816
This version fixes a potential uncontrolled format string problem. This
makes it possible to activate the format security check.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 47600
This version fixes a potential uncontrolled format string problem. This
makes it possible to activate the format security check.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 47599
This version fixes a potential uncontrolled format string problem. This
makes it possible to activate the format security check.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 47598
I accidentally deactivated it. This should be set to active it in any case
independently what was set in the user config.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 47587
Now we have all targets using platform_pre_upgrade (and calling
nand_do_upgrade directly) we don't need nand_upgrade_stage1 to be in
sysupgrade_pre_upgrade hooks anymore.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 47489
This is done with existing code from the WRT1900AC port.
It makes sure the "auto_recovery" bootloader option is set,
and resets the s_env boot counter after a successful boot.
This gives users without a serial console connection some
measure of safety.
Signed-off-by: Claudio Leite <leitec@staticky.com>
SVN-Revision: 47433
Recently all targets have been updated to call nand_do_upgrade directly.
It's not needed anymore to leave a magic /tmp/sysupgrade-nand-path mark
during image check.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Acked-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 47079
'opkg list' command only displays the available packages' name, version and
description. It would be useful to also see the approximate size of the
available package.
This patch extends "opkg list" command with "--size" to optionally show also
the *.ipk size.
* Default behaviour is to list the available packages as earlier:
"name - version - description"
* with "--size" the output of is "name - version - size - description".
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
SVN-Revision: 46980
this need to be applied after my work on ujail (procd git)
ujail doesn't depend on seccomp and some archs dont support seccomp
Signed-off-by: Etienne CHAMPETIER <champetier.etienne@gmail.com>
SVN-Revision: 46936
Expose the "lists_dir" configuration file option as command line
switch so that we can override it for the Image Builder environment.
Also add a more standard PKG_MAINTAINER variable while touching
the Makefile.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 46911
As the OpenWrt build system only resolves build dependencies per directory,
all opkg variants were causing libopenssl to be downloaded and built,
not only opkg-smime. Fix this by applying the same workaround as in
ustream-ssl.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 46850
The u-boot boot counter was never reset after a successful boot,
which sometimes could make some variables become out of sync.
This patch adds support for the boot counter and enables
auto_recovery unconditionally.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
Signed-off-by: Rob Mosher <nyt-openwrt@countercultured.net>
Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
SVN-Revision: 46690
fix generating syscall-names.h
Sometimes the syscall number is not defined with a number but with an
offset to an other syscall and then make_syscall_h.sh created some
broken header file.
For example the bit/syscall.h from musl for i386 has this:
#define __NR_timer_create 259
#define __NR_timer_settime (__NR_timer_create+1)
With this patch the resulting array looks like this:
[259] = "timer_create",
[(__NR_timer_create+1)] = "timer_settime",
This closes#20195.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 46612
Split the opkg.conf into three files, to make it easier to support custom
feeds and configs:
* /etc/opkg.conf -> base opkg configuration
* /etc/opkg/distfeeds.conf -> default Openwrt package feeds
* /etc/opkg/customfeeds.conf -> custom package feeds
Of these three, only the base opkg.conf and the customfeeds.conf is marked
as to be kept, so that the distfeeds.conf from the image is always used.
To ease migration, a script is added that moves any feeds from /etc/opkg.conf
to /etc/opkg/customfeeds.conf on first boot.
Also ensure that any keys used for verification are also kept in upgrade.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 46491
This update adds a couple of file plugin enhancements, such as base64 support
and setting mode when files are written.
Signed-off-by: Luka Perkov <luka@openwrt.org>
SVN-Revision: 46384
trx.c in mtd would not include endian.h, so on systems that do not have
this header implicitly included from the other headers (like musl), both
__BYTE_ORDER and __BIG_ENDIAN would be undefined and thus 0, leading to
it always presuming a big-endian system. this would lead to issues when
running mtd fixtrx on little-endian systems, as it would never recognize
the TRX magic as result of the broken STORE32_LE() macro.
Signed-off-by: Shiz <hi@shiz.me>
Tested-by: Shiz <hi@shiz.me>
SVN-Revision: 45896
there is a conceptual design flaw in our interface events. workaround this by
disabling duplicate message supression in procd. we need to fix this properly
for the next release
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 45883
this is an ugly hack that will be removed when the netifd maintainers
have time to look at the problem.
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 45876
This introduces a common macro to assemble the correct url templates to
avoid code duplication and have the feed config handling in a central place.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 45799
The last line of procd.sh has a reference to procd_add_interface_reload. procd_add_interface_reload
doesn't seem to exist. I've removed the reference of it to minimize confusion.
Signed-off-by: Eric Schultz <eschultz@prplfoundation.org>
SVN-Revision: 45487
To prevent future confusion around '/overlay' vs. 'overlay' simply reject
relative path specifications as mount points since such paths result in
undefined behaviour anyway.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 45404
this allow us to use syslog tcp with \0 trailer
instead of \n trailer (logread -0 option)
Signed-off-by: Etienne CHAMPETIER <champetier.etienne@gmail.com>
SVN-Revision: 45329
Use xattr to store the filesystem initialization state of the overlay.
As long as the filesystem is not marked as initialized yet (happens in
/etc/init.d/done), all overlay data (except for sysupgrade.tgz) will be
discarded before the system is allowed to boot
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
SVN-Revision: 44942
In mtd_verify(), the return value of md5sum() has been
interpreted as error if nonzero, while the function
returns number of processed bytes, which caused
mtd_verify() to always fail.
This patch fixes error checking to interpret only
negative values as errors.
Signed-off-by: Zefir Kurtisi <zefir.kurtisi@neratec.com>
SVN-Revision: 44605
Remove the need for the header file to be exported - we don't need most
of it anyway; all we care about are the offset of the rootfs length and
header crc fields.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 44557
Procd enabled init scripts can now specify:
procd_set_param stdout 1
procd_set_param stderr 1
... to relay their respective standard IO streams to the system log.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 44547
Update to git head in order to switch the block, mount_root etc. logging to
the common ulog() api for putting early boot messages into dmesg.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 44545
Update to git head in order to switch the kmodloader logging to the common
ulog() api for putting early boot messages into dmesg.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 44544
Update to git head in order to switch the procd logging to the common ulog()
api for putting early boot messages into dmesg.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 44543
- Support external overlays on non-MTD systems again
- Account for new directory structure of overlayfs partitions (additional upper/ component)
- Support executing block from either overlay or rom
- Support loading fstab from either overlay or rom
- Log extroot failures to dmesg
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 44535
This is required during early boot in the extroot setup phase to support
loading kmods from the overlay that depend on kmods on the rom partition.
The subsequent mount_root updates will use the LD_LIBRARY_PATH env
variable to pass kmod directories to kmodloader.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 44533
Update to latest git head in order to support disabling the build
of upgraded which causes linker errors on avr32.
This also adds some fixes to the ubus system.info and system.board
methods.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 44134
fixes a bug where wdt write happened with no valid fd available
--> procd: WDT failed to write: Bad file descriptor
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 43711
if the initramfs kernel and payload were flashed in one block, the payload might not be at the start of the ubi partition due to bad blocks inside the kernel partition.
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 43513
it is now possible to flash a initramfs kernel and a sysupgrade tar file inside the ubi partition. on first boot, the takeover script will find the tar file, extract and finally sysupgrade it. this allows us to flash owrt/ubi images in a 2 phase setup using ODM webuis that are not ubi aware. this is needed by some mediatek and brokencom devices.
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 43505
this can be used on nand flashes and will skip bad blocks and run ecc on the
read data before dumping it.
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 43500
mtd.c:544:7: warning: format ‘%x’ expects argument of type ‘unsigned int’, but argument 3 has type ‘ssize_t’ [-Wformat]
mtd.c:602:1: warning: label ‘done’ defined but not used [-Wunused-label]
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 43383
1) nand_upgrade_tar: use a simpler condition that matches if () { }
logic and fix check of $kernel_length variable
2) nand_do_upgrade_stage2: use case, otherwise one could believe we
always call nand_upgrade_tar
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 43373
setsid is called fixing the pgrp issue
trigger the wdt while modules are being inserted
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 43193
Note, that licensing stuff is a nightmare: many packages does not clearly
state their licenses, and often multiple source files are simply copied
together - each with different licensing information in the file headers.
I tried hard to ensure, that the license information extracted into the OpenWRT's
makefiles fit the "spirit" of the packages, e.g. such small packages which
come without a dedicated source archive "inherites" the OpenWRT's own license
in my opinion.
However, I can not garantee that I always picked the correct information
and/or did not miss license information.
Signed-off-by: Michael Heimpold <mhei@heimpold.de>
SVN-Revision: 43155
- Consider not installed feeds as well
- Add option to decide whether to comment disabled feeds
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 42931
[base-files] shell-scripting: fix wrong usage of '==' operator
normally the '==' is used for invoking a regex parser and is a bashism.
all of the fixes just want to compare a string. the used busybox-ash
will silently "ignore" this mistake, but make it portable/clean at least.
this patch does not change the behavior/logic of the scripts.
Signed-off-by: Bastian Bittorf <bittorf@bluebottle.com>
SVN-Revision: 42911
Implementing "add-cert.sh" functionality described at
http://wiki.openwrt.org/doc/howto/wget-ssl-certs into Makefile
otherwise you need to create symbolic links for certificate hashes
yourself.
Signed-off-by: Christian Schoenebeck <christian.schoenebeck@gmail.com>
SVN-Revision: 42660
Without running fixtrx the image will not boot at the second time,
because the CRC the boot loader check is invalid at that time.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
SVN-Revision: 42639
when running the postinst trigger we set PKG_UPGRADE=1 if this is an upgrade
and not an install.
Signed-off-by: John Crispin <blogic@openwrt.org>
SVN-Revision: 42468
On brcm47xx (Huawei E970), I noticed that sysupgrade now tries to flash trx images which don't fit into the flash, resulting in non-booting system.
I found that mtd does trx size checking only when flashing to a partition called "linux". This patch changes this to be "firmware".
This matches the following patch which changed partition name in bcm47xxpart.c driver to "firmware":
commit 86b4d5ef68ca77ff6724ebb9bddd6b93239c87fc
Author: hauke <hauke@3c298f89-4303-0410-b956-a3cf2f4a3e73>
Date: Sun Dec 1 16:32:32 2013 +0000
brcm47xx: use "firmware" partition name
...
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 38973
Signed-off-by: Mathias Adam <m.adam--openwrt@adamis.de>
SVN-Revision: 42403
somebody started to set a function returncode in the validation
stuff and everybody copies it, e.g.
myfunction()
{
fire_command
return $?
}
a function automatically returns with the last returncode,
so we can safely remove the command 'return $?'. reference:
http://tldp.org/LDP/abs/html/exit-status.html
"The last command executed in the function or script determines the exit status."
Signed-off-by: Bastian Bittorf <bittorf@bluebottle.com>
SVN-Revision: 42278
This changeset implements a new menuconfig option to generate separate
repositories for each enabled package feed instead of one monolithic one.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
SVN-Revision: 42002
A missing path prevents the rootfs type contained in a SysupgradeNAND
tar file from being determined correctly.
This fixes it, and also corrects a minor spelling mistake.
Signed-off-by: Ben Mulvihill <ben.mulvihill@gmail.com>
SVN-Revision: 41564
