Hello,
> The problems we've observed with the backup *restore* functionality are:
>
> * It is applied with a simple reboot, and _apparently_, it has no hooks
> or other way to ensure system sanity before (or after) the backup restore.
>
> Which would be fine if the following point was never true:
Remove GLOB_ONLYDIR patch. Only fstools relies on it. fstools can be fixed
separately.
Remove woresize.h file. It seems to be for an old version of GCC.
Remove features.h and glibc-types files. Same as above.
Replace sys/cdefs.h with what is in alpine, openembedded, and yocto. The
actual header
This release introduces 64-bit time_t, which is needed to avoid the
year 2038 problem.
Remove PKG_SOURCE and PKG_SOURCE_SUBDIR since the defaults are used.
Signed-off-by: Rosen Penev
---
v3: rebase and split commit between version update and header
removal.
v2: removed more glibc headers.
to
On Mon, Jul 20, 2020 at 3:35 PM Rosen Penev wrote:
>
> The proper format is %zu .
>
> Signed-off-by: Rosen Penev
https://downloads.openwrt.org/snapshots/faillogs/powerpc_464fp/base/procd/compile.txt
> ---
> jail/jail.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/jail
The proper format is %zu .
Signed-off-by: Rosen Penev
---
jail/jail.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/jail/jail.c b/jail/jail.c
index 5a6802f..90de411 100644
--- a/jail/jail.c
+++ b/jail/jail.c
@@ -1366,7 +1366,7 @@ static int parseOCIprocessuser(struct blob_a
musl 1.2.0 turns time_t into a 64-bit value, even on 32-bit. This makes it
compatible.
Signed-off-by: Rosen Penev
---
rcS.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/rcS.c b/rcS.c
index c2e1abb..2851fae 100644
--- a/rcS.c
+++ b/rcS.c
@@ -18,6 +18,7 @@
#include
#i
This define is not only useless, it breaks the build as the little
endian header always gets included, which subsequently errors.
Signed-off-by: Rosen Penev
---
package/kernel/mac80211/Makefile | 2 +-
.../mac80211/patches/build/070-no-rtw-byteorder.patch | 10 ++
SYS_settimeofday is no longer present. That is, it's replaced with the
time32 variant. There is no time64 variant.
Note that 5a7c064bdbb71bfbcded073c7c0a8723be306009 switched the patch to
use the syscall instead of the function as musl's settimeofday does not
use the tz argument for anything.
Sig
They fix compilation with musl 1.2.0, which uses 64-bit time_t.
Signed-off-by: Rosen Penev
---
.../utils/busybox/patches/002-time64.patch| 46 ++
.../utils/busybox/patches/003-gettime.patch | 48 +++
2 files changed, 94 insertions(+)
create mode 100644 pack
Update to musl 1.2.0 removes sgidefs.h. This causes various compilation
issues. Add patches from openembedded to fix.
Signed-off-by: Rosen Penev
---
package/devel/gdb/Makefile | 2 +-
package/devel/gdb/patches/140-sgidefs.patch | 32 +++
package/devel/gdb/patche
The ssh symlink was still being created even when dbclient was disabled in the
build configuration. Fix this annoyance.
Signed-off-by: Rui Salvaterra
---
package/network/services/dropbear/Makefile | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/package/network/services/
> > Tested on the EAP245 v1 running the latest firmware (v1.4.0). The
> > binary patch might not apply to uclited from other firmware versions.
> >
> > Signed-off-by: Sander Vanheule
>
> Seems like I was overdue on a proper read of the kernel patch submission
> guidelines. My understanding from t
Hi Adrian,
On Sun, 2020-07-19 at 23:24 +0200, Sander Vanheule wrote:
> TP-Link EAP245 v1 is an AC1750 (802.11ac Wave-1) ceiling mount access
> point.
>
> Device specifications:
> * SoC: QCA9563 @ 775MHz
> * RAM: 128MiB DDR2
> * Flash: 16MiB SPI-NOR
> * Wireless 2.4GHz (SoC): b/g/n, 3x3
> * Wirele
Buildinfo files were added to allow reproducible builds testing. The
combination of the three buildinfo files
{feeds,config,version}.buildinfo allows to recreate the buildsystem and
check if compiled binaries are bit for bit the same as distributed on
official servers.
To support the verification
Signed-off-by: Todor Colov
---
tools/firmware-utils/Makefile | 4 +-
tools/firmware-utils/src/linksys/addfwhdr.c | 195
tools/firmware-utils/src/linksys/bcmdefs.h| 318 +
.../firmware-utils/src/linksys/code_pattern.h | 396
tools/firmw
The source files of addfwhdr are stuctured under separate manufacturer specific
'linksys' directory
Todor Colov (1):
tools firmware-utils add linksys addfwhdr tool for native WEB upgrade
of openwrt-factory over the stock WEB interface.
tools/firmware-utils/Makefile | 4 +
Todor Colov (1):
update factory reset for generic handling of all type of
devices(jffs2,ubi,etc)
package/base-files/files/etc/rc.button/reset | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--
2.27.0
___
openwrt-devel mailing list
openwr
Signed-off-by: Todor Colov
---
package/base-files/files/etc/rc.button/reset | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/package/base-files/files/etc/rc.button/reset
b/package/base-files/files/etc/rc.button/reset
index 2403122ad2..0c11bfe462 100755
--- a/package/base-files
Todor Colov (1):
sysupgrade need fwtool over the next stages to remove the metadata
info for UBI volumes. This patch creates copy of /usr/bin/fwtool to
/tmp directory to preserves it after the sysupgrade ramdisk switch.
Later fwtool is being referenced from platform.sh script for devi
Signed-off-by: Todor Colov
---
package/base-files/files/lib/upgrade/fwtool.sh | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/package/base-files/files/lib/upgrade/fwtool.sh
b/package/base-files/files/lib/upgrade/fwtool.sh
index a0b3fb0a04..e4736c06f3 100644
--- a/package/b
Hello,
I am bringing the topic here because I am pretty sure it is something
that affects any downstream firmware that focus on regular users...
We'd like to know if anyone has specific thoughts on the issue, or has
implemented/is testing/considered solutions to the issues I describe in
this
> -Original Message-
> From: openwrt-devel [mailto:openwrt-devel-boun...@lists.openwrt.org]
> On Behalf Of Sander Vanheule
> Sent: Montag, 20. Juli 2020 18:28
> To: m...@adrianschmutzler.de
> Cc: yn...@true.cz; openwrt-devel@lists.openwrt.org; Julien Dusser
>
> Subject: Re: [RFC PATCH v2 3
Hi Adrian,
On Mon, 2020-07-20 at 00:25 +0200, m...@adrianschmutzler.de wrote:
> Hi,
>
> > Signed-off-by: Julien Dusser
> > Signed-off-by: Sander Vanheule
>
> technically, if Julien is first SoB, you should also put him in the
> author (From:) field.
The DTSI is derived from Julien's DTS for t
When retrieving the PID for hostapd and wpa_supplicant via ubus the
wrong service name is currently used. This leads to the following error
in the log:
netifd: radio0 (1409): WARNING (wireless_add_process):
executable path /usr/sbin/wpad does not match process path (/proc/exe)
Fixing the service
Don't kill the wireless daemon on teardown. hostapd as well as
wpa_supplicant are managed by procd which would detect the shutdown of
either process as a crash loop.
Signed-off-by: David Bauer
---
package/kernel/mac80211/files/lib/netifd/wireless/mac80211.sh | 2 --
1 file changed, 2 deletions(-
Matt Johnston [2020-07-19 11:32:56]:
Hi,
> It looks like gpg-agent added support in 2.2.5. OpenSSH client 7.7 or later
> will print a warning about it.
> https://dev.gnupg.org/T3880
Debian 9 (EOL in June 30, 2022) has following: gpg-agent (GnuPG) 2.1.18,
OpenSSH_7.4p1
Ubuntu 18.04 (EOL
From: Nathaniel Wesley Filardo
This allows us to attach a hwrng, for example.
At most one sample will be taken every time we also add entropy via the
jitter mechanism. We won't try reading if the stream hasn't indicated
its readiness, and we'll go back to waiting if ever the stream produces
0 b
From: Nathaniel Wesley Filardo
There's no need to hold one of these structures across invocations of
low_entropy_cb. Just put one on the stack and use that.
Signed-off-by: Nathaniel Wesley Filardo
---
urngd.c | 62 +
1 file changed, 18 i
From: Nathaniel Wesley Filardo
These look like there had been plans to stay asleep if there was
"sufficient" entropy in the system, but the code does not do so.
Signed-off-by: Nathaniel Wesley Filardo
---
urngd.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/urngd.c b/urngd.c
index f782
From: Nathaniel Wesley Filardo
Now that entropy can come from two places, indicate which
Signed-off-by: Nathaniel Wesley Filardo
---
urngd.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/urngd.c b/urngd.c
index 1168674..9eedb6b 100644
--- a/urngd.c
+++ b/urngd.c
> > --- /dev/null
> > +++ b/target/linux/ipq806x/base-files/etc/rc.button/reset
>
> What's that for? Implementation of reset button is available with KEY_RESTART
> OOTB.
> -> NOTE: Linksys E8350 uses ubi volumes not jffs2 . Until the "factory reset"
> script properly handle such volumes(which is
> - My opinion is that the current approach for specific firmware utils is
> lacking of scalability. All custom fw tools are in the same directory, what
> if they become 1000 or more?
I don't see how scalability will be better if every piece of non-specific code
is copied in _again_ for every n
Hi,
> Signed-off-by: Kevin Darbyshire-Bryant
Acked-by: Jo-Philipp Wich
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
> > + model = "Linksys EA8300 V1 WiFi Router";
>
> So, is it V1? Then add that to the compatible etc.: linksys,e8350-v1
> -> I've not sen V2 yet, so I can remove the versioning of this device in the
> updated patch.
Well, the question is not whether a v2 is there _yet_, but whether you are
absol
> -Original Message-
> From: openwrt-devel [mailto:openwrt-devel-boun...@lists.openwrt.org]
> On Behalf Of Johannes Kimmel
> Sent: Montag, 20. Juli 2020 08:05
> To: openwrt-devel@lists.openwrt.org
> Cc: Johannes Kimmel ; mschiffer@universe-
> factory.net
> Subject: [PATCH v2 1/2] vxlan: rem
35 matches
Mail list logo
