Hi,
procd causes trouble when running in a LXD container, however would be
nice to use OpenWrt as small VMs.
There are patches [0] "fixing" the issue, but they're not really
upstream ready (for obvious reasons).
Would it be possible to add a detection to procd if it's running in an
Hi all,
thanks to this[0] bug report I found that the ar71xx/generic
ImageBuilder doesn't create any firmware images for profile OM2P when
used with BIN_DIR.
My first guess is that it's somewhat related to a hard coded destination
which doesn't handle BIN_DIR. At the very end of an "successful"
Hi,
I couldn't find a mapping between the profiles and the
"supported_devices" in the resulting firmware metadata. With a script I
extracted the metadata and created a mapping, maybe it is useful for
someone.
https://chef.libremesh.org/download/openwrt/18.06.1/mapping.txt
It only contains
Hi,
tl;dr: please run zsyncmake[0] on created snapshot SDK and IB archives
I'm offering a web service to download custom firmware images via an
API[1], this includes snapshot images. To stay up to date, all snapshot
ImageBuilders are daily re-downloaded. While being convenient for some
Signed-off-by: Alexander Couzens
[fixed waiting for unlock and commit message]
Signed-off-by: Paul Spooren
---
iptables.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/iptables.c b/iptables.c
index a095621..559fe7d 100644
--- a/iptables.c
+++ b/iptables.c
@@ -55,6 +55,8 @@
#include &
From: Alexander Couzens
To lock a second lock file at the same time, introduce fw3_{un,}lock_path.
fw3_lock_path support the path as parameter in difference to fw3_lock which
only locks the fw3 lock file (/var/run/fw3.lock)
Signed-off-by: Alexander Couzens
---
utils.c | 34
show license in `opkg info `.
CC: Jo-Philipp Wich
Signed-off-by: Paul Spooren
---
libopkg/opkg_cmd.c | 3 +++
libopkg/pkg.c | 10 ++
libopkg/pkg.h | 1 +
libopkg/pkg_parse.c | 5 +
libopkg/pkg_parse.h | 1 +
5 files changed, 20 insertions(+)
diff --git a/libopkg
Double `!!` results in false positives, making the snapshot unusable as
procd won't mount essential things on *non virtial machines*.
Signed-off-by: Paul Spooren
---
container.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/container.h b/container.h
index dd2e432..3c388ce
Hi all,
currently OpenWrt only offers a very basic Docker image[0] for testing
of the packages.git repo. The image is not directly usable as the CI
does most of the work, like setting up the SDK and adding feeds[1]. I'd
like to propose two additional images and added CI test examples for
detects if running in a docker container, which then requires special
treatment of mounts. OpenWrt within Docker is useful for CI testing.
Signed-off-by: Paul Spooren
---
container.h | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/container.h b/container.h
index d33fa32
- 1.8.2-3 - GPL-2.0
iptables - 1.8.2-3 - GPL-2.0
jshn - 2019-02-27-eeef7b50-1 - ISC
jsonfilter - 2018-02-04-c7e938d6-1 - ISC
...
CC: Jo-Philipp Wich
Signed-off-by: Paul Spooren
---
v2 <- v1 ignore license via PFM_LICENSE in opkg_cmd.c when not actually needed
libopkg/opkg_cmd.c |
- 1.8.2-3 - GPL-2.0
iptables - 1.8.2-3 - GPL-2.0
jshn - 2019-02-27-eeef7b50-1 - ISC
jsonfilter - 2018-02-04-c7e938d6-1 - ISC
...
CC: Jo-Philipp Wich
Signed-off-by: Paul Spooren
---
libopkg/opkg_cmd.c | 3 +++
libopkg/opkg_conf.c | 1 +
libopkg/opkg_conf.h | 1 +
libopkg/pkg.c | 11
Hi Ted,
> This sounds like a good idea - would giving you commit access to our
> docker hub [0] repo be of any help? Is there anything else I could
> help with?
I'd be happy with commit access to the docker hub (not to mistake with
git commit access).
Please find here a reworked version with a
show license in `opkg info `.
Signed-off-by: Paul Spooren
---
This patch isn't complete yet. It show the license info as long as the package
isn't installed. Once it's installed it seems to forget the information. I can't
figure out why. Also printing `opkg list` doesn't add the License, neither
On 5/14/19 6:24 PM, Daniel Golle wrote:
> I don't think adding formatter-information into the code makes sense
> at this stage, as this is meant for exeptions from an otherwise applied
> style. Here, obviously clang-format defintions don't reflect the
> current style.
They don't, they reflect
-openwrt/blob/master/patches/procd-master/0003-docker-fix-problem-stopping-container.patch
Signed-off-by: Paul Spooren
---
container.h | 4 +++-
state.c | 4
2 files changed, 7 insertions(+), 1 deletion(-)
diff --git a/container.h b/container.h
index d33fa32..04d63ed 100644
On 5/23/19 10:42 AM, Petr Štetiar wrote:
> Paul Spooren [2019-05-22 19:24:19]:
>
>> Signed-off-by: Paul Spooren
>> ---
>> state.c | 3 +++
>> 1 file changed, 3 insertions(+)
>>
>> diff --git a/state.c b/state.c
>> index ff1734f..9f196ee 10064
Hi,
On 5/23/19 12:30 PM, Petr Štetiar wrote:
> I'm wondering if this Docker related changes shouldn't be included as a part
> of a patch series which would add complete Docker support to OpenWrt (do you
> plan to do this?), so it could be tested together.
>
> Or is it now possible to just
On 5/23/19 2:16 PM, Petr Štetiar wrote:
> Paul Spooren [2019-05-23 13:36:37]:
>
>> No, thanks for the link. Seems that Docker now uses LXC anyway which
> No more[1] :-)
>
> Remove LXC support.
>
> The LXC driver was deprecated in Docker 1.8. Following the deprecat
copied from the official kernel style:
https://github.com/torvalds/linux/blob/master/.clang-format
Signed-off-by: Paul Spooren
---
.clang-format | 118 ++
1 file changed, 118 insertions(+)
create mode 100644 .clang-format
diff --git a/.clang
copied from the official kernel style:
https://github.com/torvalds/linux/blob/master/.clang-format
also modify gitignore to unignore the style file
Signed-off-by: Paul Spooren
---
.clang-format | 117 ++
.gitignore| 1 +
2 files changed
Currently some files use a mixture of spaces and tabs within a single
file, instead of fixing style manually, a tool could do the job.
Signed-off-by: Paul Spooren
---
container.h | 3 +-
inittab.c | 109 +++-
libc-compat.h | 5 +-
log.h | 12
rto to the Cc loop]
>
> Paul Spooren [2019-05-04 17:48:03]:
>
> Hi,
>
>> from my point of view the current OpenWrt device documentation lacks
>> some unity, meaning the table of hardware is partly incomplete or
>> inconsistent. Also documentations are often rewritten wit
On 5/14/19 4:08 PM, Petr Štetiar wrote:
> Paul Spooren [2019-05-10 17:36:21]:
>
> Hi,
>
>> Currently some files use a mixture of spaces and tabs within a single
>> file, instead of fixing style manually, a tool could do the job.
> I find consistent coding style impo
using clang-format-9
Signed-off-by: Paul Spooren
---
container.h | 5 +-
inittab.c | 108 +
libc-compat.h | 5 +-
log.h | 12 ++--
preload.h | 49 ++-
procd.c | 16 ++---
rcS.c | 18 +++---
signal.c | 22
copied from the official kernel style:
https://github.com/torvalds/linux/blob/master/.clang-format
also modify gitignore to unignore the style file
Signed-off-by: Paul Spooren
---
Added a single OpenWrt specific ForEachMacro rule
.clang-format | 496
github.com/containercraft/openwrt-lxd/blob/master/patches/procd-openwrt-18.06/001_lxd_no_mounts.patch
Signed-off-by: Paul Spooren
---
initd/early.c | 19 +++
initd/zram.c| 10 ++
plug/coldplug.c | 13 -
procd.c | 7 ++-
4 files changed, 31 inserti
Hi all,
from my point of view the current OpenWrt device documentation lacks
some unity, meaning the table of hardware is partly incomplete or
inconsistent. Also documentations are often rewritten with different
precision and "tongue".
To unify things I created a *PoC*[0], mostly based on the
he cleanup.
[0]:
https://github.com/containercraft/openwrt-lxd/blob/master/patches/procd-openwrt-18.06/001_lxd_no_mounts.patch
Signed-off-by: Paul Spooren
---
container.h | 22 ++
initd/early.c | 20
initd/zram.c| 11 +++
plug/coldp
he cleanup.
[0]:
https://github.com/containercraft/openwrt-lxd/blob/master/patches/procd-openwrt-18.06/001_lxd_no_mounts.patch
Signed-off-by: Paul Spooren
---
container.h | 22 ++
initd/early.c | 20
initd/zram.c| 11 +++
plug/coldp
Hi all,
just as a notification as I don't want to spread the discussion to all
kinds of platforms:
I proposed (actually jow in 2017) to merge openwrt-routing/packages into
openwrt/packages subfolder net.
Please find the current discussion at GitHub[0].
This is intended to be a transfer, not a
The PR was accepted, I'd be happy if newly ported devices could follow
this style!
I'll do some work on reorganizing existing targets.
Best,
Paul
> [0]: https://github.com/openwrt/openwrt/pull/2124
> [1]: https://sudhanshu16.github.io/openwrt-firmware-selector/
>
>
>
>
our thoughts of this transfer!
Thanks for your time,
Paul
> or we fall back in the
> past problematic situation.
>
> Cheers!
> Gio
>
> On Friday, 5 July 2019 14:26:31 CEST Paul Spooren wrote:
>> Hi all,
>>
>> just as a notification as I don't want to spread t
Hi all,
to improve security of the router sysupgrade process, it's sane to check
firmware images for signatures of trusted parties. While this should
always be optional (aka no vendor locking), it helps *basic* users to
easily verify that they are installing the image they intended.
It is
to use procd in LXC containers they have to support SIGPWR to shutdown.
Signed-off-by: Paul Spooren
---
initd/init.c | 1 +
signal.c | 2 ++
2 files changed, 3 insertions(+)
diff --git a/initd/init.c b/initd/init.c
index 0349e6e..29eee50 100644
--- a/initd/init.c
+++ b/initd/init.c
Hi all,
some days ago I wrote this proposal[0] to split DEVICE_TITLE in
DEVICE_VENDOR, DEVICE_MODEL and DEVICE_VARIANT (holding revisions and
storage sizes if applicable).
I extended the image.mk to to use the newly introduced variables if
DEVICE_TITLE is empty, meaning it can be used as a drop
Hey,
on commit 5e928acf22cdc956eabe6e4b2327b34eb0ee66da applying fails:
error: patch failed: package/libs/ustream-ssl/Makefile:10
error: package/libs/ustream-ssl/Makefile: patch does not apply
However removing the ustream-ssl part of the patch results in a working
version for (at least) all
when started via terminal. Before it would only try
to connect to an existing ubus instance.
The -D arg handling was kindly created (with < 60 seconds RTT) by John,
I just created the patch and removed pid checking.
CC: John Crispin
Signed-off-by: Paul Spooren
---
v2: added usage/help mess
2.manifest
Signed-off-by: Paul Spooren
---
v2: removed senseless newline - sorry
include/image.mk | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/image.mk b/include/image.mk
index e55ac22d0d..c6a6ab7993 100644
--- a/include/image.mk
+++ b/include/image.mk
@@ -10
2.manifest
Signed-off-by: Paul Spooren
---
include/image.mk | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/include/image.mk b/include/image.mk
index e55ac22d0d..cc333052a2 100644
--- a/include/image.mk
+++ b/include/image.mk
@@ -106,7 +106,7 @@ define add_jffs2_mark
-by: Paul Spooren
---
target/linux/ipq40xx/Makefile | 1 +
target/linux/ipq40xx/generic/target.mk | 1 +
target/linux/ipq806x/Makefile | 1 +
target/linux/ipq806x/generic/target.mk | 1 +
4 files changed, 4 insertions(+)
create mode 100644 target/linux/ipq40xx/generic/target.mk
Hi John,
This commit adds the Generic subtarget resulting in consistent naming.
and
already uses `x/generic/` as subfolder as if the subtarget would exist.
I'm very much in favor of consistent names[0][1][2] as it reduces the
hassle when trying automate things, like building images via an
On 25.08.19 22:08, Bjørn Mork wrote:
Paul Spooren writes:
as 19.07 is *just around the corner* I was wondering if there's a
better way of distinguishing between versions.
Right now, I see 4 different *channels* which somewhat match the
Debian style, therefore a possible mapping:
18.06.N
This will not only create config.buildinfo as before, but also
version.buildinfo and feeds.buidinfo.
CC: Jo-Philipp Wich
Signed-off-by: Paul Spooren
---
phase1/master.cfg | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/phase1/master.cfg b/phase1/master.cfg
index
Allow to run procd from a bash script and still let it fire up system
services. This is handy when running from within a container that does
not start any init command on pid 1, like on GitLab CI.
Signed-off-by: Paul Spooren
---
procd.c | 5 +
1 file changed, 1 insertion(+), 4 deletions
Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.
CC: Hauke Mehrtens
Signed-off-by: Paul Spooren
---
target/linux/bcm53xx/Makefile | 1 +
target/linux/bcm53xx/generic/target.mk | 1 +
2 files changed, 2 insertions(+)
create mode 100644 target
Perkov
CC: Roman Yeryomin
CC: Sergey Ryazanov
CC: Tim Harvey
CC: Tomasz Maciej Nowak
Signed-off-by: Paul Spooren
---
target/linux/ath25/Makefile | 1 +
target/linux/ath25/generic/target.mk | 1 +
target/linux/cns3xxx/Makefile| 1 +
target/linux/cns3xxx/generic
Hi,
various targets at OpenWrt/snapshots[0] are no longer build and cause
(at least for me) from time to time some trouble. Would it be possible
to add a cron job removing all files from snapshots/ which are older
than, a week? The folder is not intented as an archive, right?
Sunshine,
Hi all,
as 19.07 is *just around the corner* I was wondering if there's a better
way of distinguishing between versions.
Right now, I see 4 different *channels* which somewhat match the Debian
style, therefore a possible mapping:
18.06.N -> stable
19.07-rcN -> testing
19.07-SNAPSHOT ->
On 22.08.19 00:11, John Crispin wrote:
On 22/08/2019 08:47, Paul Spooren wrote:
Hi John,
This commit adds the Generic subtarget resulting in consistent naming.
and
already uses `x/generic/` as subfolder as if the subtarget would exist.
I'm very much in favor of consistent names[0][1
I posted this a while ago could somebody please apply this?
On 22.08.19 23:02, Paul Spooren wrote:
As in 853e4dd OpenWrt should follow a unified structure, where every
device has a target/subtarget combination, if there is only one
subtarget, call it "Generic". This introduces p
What you suggest is about what we have right now. This kind of creates a misleading situation where
for some targets subtargets are present, while for others paths and image names are
"fixed" in several places to include a "generic". The reason for Paul's patch
was to get rid of the fixes at
when started via terminal. Before it would only try
to connect to an existing ubus instance.
The -D arg handling was kindly created (with < 60 seconds RTT) by John,
I just created the patch and removed pid checking.
CC: John Crispin
Signed-off-by: Paul Spooren
---
procd.c | 10 +-
1 f
-fix-problem-stopping-container.patch
Signed-off-by: Paul Spooren
---
state.c | 25 +++--
1 file changed, 15 insertions(+), 10 deletions(-)
diff --git a/state.c b/state.c
index ccf4104..b7cba88 100644
--- a/state.c
+++ b/state.c
@@ -20,6 +20,7 @@
#include
#include
Add option BUILD_LOG_DIR to menuconfig to change log destination.
The mix-up of *DIR* and *FOLDER* is confusing however.
Signed-off-by: Paul Spooren
---
config/Config-devel.in | 7 +++
rules.mk | 2 +-
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/config
Main motivation for this commit is the introduction of
`-ffile-prefix-map=` which alows reproducible build path.
Compiling tested without errors on the following targets:
* ath79
* brcm2708
* brcm63xx
* ixp4xx
* ramips
* sunxi
* x86
Signed-off-by: Paul Spooren
---
Please let me know if I
Hi Scott,
On 23.09.19 16:15, Scott via openwrt-devel wrote:
First and foremost, as a user of OpenWrt and LuCI by way of TurrisOS thanks for
everyone who has contributed to OpenWrt. As someone who's seeking to give back
to the project via localization I noticed that while OpenWrt has
Signed-off-by: Paul Spooren
---
config/Config-build.in | 2 --
.../patches/120-strip-cflags-from-binary.patch | 2 +-
rules.mk | 8
toolchain/gcc/Config.in| 8 ++--
toolchain/gcc
This is a follow up of 8cb13f4e6d which sets the right timestamp for the
/init file in initramfs. The previous patch doesn't cover it as the
files appear to come from a later step during compilation.
CC: Alexander Couzens
Signed-off-by: Paul Spooren
---
include/kernel-defaults.mk | 1 +
1
On 17.09.19 00:25, Jonas Gorski wrote:
On Sun, 15 Sep 2019 at 12:49, Paul Spooren wrote:
What you suggest is about what we have right now. This kind of creates a misleading situation where
for some targets subtargets are present, while for others paths and image names are
"fixed"
On 28.09.19 11:04, Daniel Golle wrote:
Hi Paul,
On Sat, Sep 28, 2019 at 10:44:48AM -1000, Paul Spooren wrote:
Main motivation for this commit is the introduction of
`-ffile-prefix-map=` which alows reproducible build path.
Imho definitely a good reason to move forward with the switch to GCC 8
SIGNATURE valuess over multiple build with the same commit.
This is another step in the direction of reproducible OpenWrt images.
Signed-off-by: Paul Spooren
---
I'm currently trying to step up my commit message game, please comment
if the message is unclear or confusing!
include/image.mk
This commit changes the behaviour of UUID generation. Instead of using
the rand module of perl, generating a random UUID per compile, it now
hashes the SOURCE_DATE_EPOCH, resulting in a reproducible UUID.
Signed-off-by: Paul Spooren
---
target/linux/x86/image/Makefile | 2 +-
1 file changed, 1
the same as for the original
script as the latter does not sort the targets before printing, as the
Python implementation does.
CC: Jo-Philipp Wich
[0]:
https://git.openwrt.org/?p=buildbot.git;a=blob;f=scripts/dumpinfo.pl;h=aa97f8d60379076a41b968402e9337cea824ece5;hb=HEAD
Signed-off-by: Paul
) root (0) 33 2019-11-05 23:43:08.00
v_0.0.0.md5
Signed-off-by: Paul Spooren
---
I could not runtime test this, I guessed the changed owner shouldn't be
a problem for the device.
target/linux/ramips/image/mt7620.mk | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
Thanks for debugging this!
On 11/5/19 1:23 PM, Petr Štetiar wrote:
Commit 881ed09ee6e2 ("build: create JSON files containing image info")
has removed the crucial empty new line from the image copy step
resulting in the following errors during make function expansion:
GZ_SUFFIX :=
bash:
Thank you for reworking this! Daniel already told me that a fallback is
required if SOURCE_DATE_EPOCH is unset, this looks like a clean solution.
On 11/7/19 4:09 AM, Petr Štetiar wrote:
From: Paul Spooren
x86, mvebu and tegra targets are currently using more or less same
SIGNATURE variable
) root (0) 33 2019-11-05 23:43:08.00
v_0.0.0.md5
Signed-off-by: Paul Spooren
---
v2: rename mt7520 to mt7620
target/linux/ramips/image/mt7620.mk | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/target/linux/ramips/image/mt7620.mk
b/target/linux/ramips/image
Oh thanks, updated...
On 11/5/19 10:01 PM, Bjørn Mork wrote:
subject typo. I was curious about that new chip :-)
Bjørn
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
LGTM
On 11/13/19 10:09 AM, Thomas Weißschuh wrote:
Signed-off-by: Thomas Weißschuh
---
include/prereq-build.mk | 2 ++
1 file changed, 2 insertions(+)
diff --git a/include/prereq-build.mk b/include/prereq-build.mk
index 4c59910056..65baf1f5e5 100644
--- a/include/prereq-build.mk
+++
/763e1148f68f03cb2fa85d022500acf8c66af222
Signed-off-by: Paul Spooren
---
Tested and working on Debian 10
https://buildmaster.aparcar.org/#/builders/79/builds/4/steps/64/logs/stdio
scripts/signall.sh | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/scripts/signall.sh b/scripts/signall.sh
index 269375f
the buildbot.git contains some hacks to create images that
have signature verification capabilities while not storing private keys
on buildbot slaves. This commit allows to disable these steps for the
buildbots and only perform signing on the master.
Signed-off-by: Paul Spooren
---
v3: set "y if !BUI
the buildbot.git contains some hacks to create images that
have signature verification capabilities while not storing private keys
on buildbot slaves. This commit allows to disable these steps for the
buildbots and only perform signing on the master.
Signed-off-by: Paul Spooren
---
v2: fix missing endif
Looks good to me, I ran into this issue before and would appreciate
moving forward to 6.
On 11/11/19 10:16 PM, Petr Štetiar wrote:
Currently minimal GNU supported GCC version is 7 (from May 2, 2017), buildbots
are using default GCC version 6 on Debian 9 (old stable), current Debian
stable 10
verification capabilities while not storing private keys
on buildbot slaves. This commit allows to disable these steps for the
buildbots and only perform signing on the master.
Signed-off-by: Paul Spooren
---
v4: replace ifdef with ifneq - Makefile magic
-ifdef CONFIG_SIGNED_PACKAGES
+ifneq
+1
11 Dec 2019 15:22:44 Daniel Golle :
> Hi Tomislav,
>
> On Wed, Dec 11, 2019 at 11:24:21AM +0100, Tom Psyborg wrote:
>
> > suck it
> >
>
> As a community, we decided to give our self a set of minimal rules[1].
> And even though it is in the last position, rule #12 "Be nice to each
>
On 9/25/19 3:24 PM, Scott wrote:
‐‐‐ Original Message ‐‐‐
On Tuesday, September 24, 2019 1:13 AM, Paul Spooren wrote:
I think that's a good idea! As it doesn't need to be directly integrated
in any existing workflow nor requires privileges on any Git, it's
basically just a fancy
Bump
On 11/12/19 12:04 AM, Paul Spooren wrote:
This separates the options for signature creation and verification
* SIGNED_PACKAGES create Packages.sig
* SIGNED_IMAGES add ucert signature to created images
* CHECK_SIGNATURE add verification capabilities to images
* INSTALL_LOCAL_KEY add local
Hi Yousong,
A few months ago, I started working on bringing up k3s on OpenWrt. I
managed to have a small k8s cluster with nodes running centos and
openwrt. Likely other linux distros will also do, but I didn't test
combination of different architectures yet.
Cool! That looks like an
the buildbot.git contains some hacks to create images that
have signature verification capabilities while not storing private keys
on buildbot slaves. This commit allows to disable these steps for the
buildbots and only perform signing on the master.
Signed-off-by: Paul Spooren
---
config/Config-build.in
env vars:
* TOP_DIR where to search for sysupgrade.bin images
* BUILD_KEY place of key-build{,.pub,.ucert}
* REMOVE_OTHER_SIGNATURES removes signatures added by e.g. buildbots
Only sysupgrade.bin files are touched as factory.bin signatures wouldn't
be evaluated on stock from.
Signed-off-by: Paul
All other keys are stored as 0644, the only reason this key is 0600 is
as the key-build should be hidden from others. Changing the public key
it to 0644 it is in line with all other keys in /etc/opkg/keys.
Signed-off-by: Paul Spooren
---
phase1/master.cfg | 2 +-
1 file changed, 1 insertion
Aloha,
I'm currently working on the reproducibility of OpenWrt snapshot images
and packages and came across the following problem:
Some of the ath79 images have a signature attached to the factory images
which is pointless as stock roms wont evaluate those anyway. However, as
these factory
and the signature is (obviously) gone, perfect!
Can't do any runtime tests, but I guess it's good for merging!
Thanks for the quick response!
Paul
Ref: http://lists.infradead.org/pipermail/openwrt-devel/2019-October/019205.html
Reported-by: Paul Spooren
Signed-off-by: Petr Štetiar
images which
contains metadata which are causing image reproducibility issues, so
let's build factory images from the scratch. While at it, refactor the
shared vars into common base as well.
Ref: http://lists.infradead.org/pipermail/openwrt-devel/2019-October/019205.html
Reported-by: Paul Spooren
Hi,
On 10/7/19 6:38 PM, Petr Štetiar wrote:
as discussed on IRC, this issue is caused by your custom build step[1] and
doesn't exist in the tree, so the proposed patch can be seen just a workaround
and not proper fix so I've rejected it and not going to apply.
Fine for me if the enhancements
.
Compiling tested without errors on the following targets:
* ath79
* brcm2708
* brcm63xx
* ixp4xx
* ramips
* sunxi
* x86
CC: Andre Heider
CC: Petr Štetiar
Thanks to Andre for the iremap fixup
Thanks to Petr for his patience on helping which screws to tweak
Signed-off-by: Paul Spooren
---
v3
Hi team,
there are two recent patches introducing reproducibility of most
packages and firmware images. Applying them to 19.07 would only change
timestamps and shorten debug information (of buildpath), therefore
fairly secure to merge.
Please consider backporting:
4ed356fa71 kernel.mk: add
g in build path independent
reproducible builds.
Signed-off-by: Paul Spooren
---
include/kernel.mk | 1 +
1 file changed, 1 insertion(+)
diff --git a/include/kernel.mk b/include/kernel.mk
index 73645330fe..439e910ebf 100644
--- a/include/kernel.mk
+++ b/include/kernel.mk
@@ -104,6 +104,
-by: Paul Spooren
---
phase1/master.cfg | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/phase1/master.cfg b/phase1/master.cfg
index f0151f6..0aaa079 100644
--- a/phase1/master.cfg
+++ b/phase1/master.cfg
@@ -920,7 +920,7 @@ for target in targets:
factory.addStep(ShellCommand
dition to use `generic` if the subtarget
is empty.
Tested for the kirkwood target.
Signed-off-by: Paul Spooren
---
include/image.mk | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/image.mk b/include/image.mk
index 46d592e8dc..fd04d4020b 100644
--- a/include/image.mk
Hi,
This PR refactors the JSON creation to store individual files in
$(KDIR)/tmp and create an single overview file called `profiles.json` in
the target dir.
As before, this creation is enabled by default only for the BUILDBOT.
To archive the previous behaviour the option
Hi,
for the last few days I get the following error for multiple targets,
would it be possible to rerun the snapshot servers to rebuild the
package `iw`?
$ make manifest PROFILE="devolo_dvl1750e"
[...]
Collected errors:
* satisfy_dependencies_for: Cannot satisfy the following dependencies
implementation used the functions `json.dumps()` which seem
to have caused broken files. Now the `pathlib` library is used to deal
with files and the `json` library only reads/writes into variables.
Tested via buildroot & ImageBuilder on ath79/generic.
Signed-off-by: Paul Spooren
---
v2:
* One ins
On 01.03.20 02:34, Petr Štetiar wrote:
Paul Spooren [2020-02-29 16:48:50]:
FYI:
$ grep JSON .config
CONFIG_JSON_OVERVIEW_IMAGE_INFO=y
$ cat bin/targets/imx6/generic/profiles.json
{}
This problem occurs also fox x86, the problem is that the image function
is not properly called
.
Thanks, included in v3
On 3/1/20 3:48 AM, Paul Spooren wrote:
JSON info files contain machine readable information of built profiles
and resulting images. These files where added via 881ed09ee6e2. They are
useful for firmware wizards and script checking for reproducibility.
Currently all JSON
Hi,
A first step could be to establish a *versions.json* file at the root of
downloads.openwrt.org! The file would allow to check if a device still runs
the latest release. JSON seems common enough and is well supported by LuCIs
JavaScript implementation and also via jshn.sh on a CLI/script
implementation used the functions `json.dumps()` which seem
to have caused broken files. Now the `pathlib` library is used to deal
with files and the `json` library only reads/writes into variables.
Tested via buildroot & ImageBuilder on ath79/generic.
Signed-off-by: Paul Spooren
---
v2:
* One ins
Hi Petr,
On 02.03.20 23:12, Petr Štetiar wrote:
Paul Spooren [2020-03-02 16:19:05]:
- .IGNORE: $(BIN_DIR)/$(call IMAGE_NAME,$(1),$(2))
[...]
$(BIN_DIR)/$(call IMAGE_NAME,$(1),$(2)): $(KDIR)/tmp/$(call
IMAGE_NAME,$(1),$(2))
- cp $$^ $$@
+ -cp $$^ $$@
The prefixed dash
son` files is created as it would be empty anyway.
As before, this creation is enabled by default only if `BUILDBOT` is set.
Tested via buildroot & ImageBuilder on ath79/generic, imx6 and x86/64.
Signed-off-by: Paul Spooren
---
v2:
* One instead of three CONFIG options
* Only created `profiles
1 - 100 of 590 matches
Mail list logo
