odhcp6c logs messages related to its activity when invoked with -v, but
there is no way to configure this from within OpenWrt. This adds a UCI
option to turn on odhcp6c logging, disabled by default. To enable, set,
for example, network.wan6.verbose = 1.
Signed-off-by: Mark Mentovai
---
package
that the kernel's objtool and related configuration have seen a
major overhaul since kernel 5.15, and may need more attention again
after 22922deae13f, in kernel 5.19.
Signed-off-by: Mark Mentovai
---
include/kernel.mk | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/kernel.mk b
Thibaut wrote:
I’m experiencing a strange bug on Yuncore AX820 (mt7621/mt7905/mt7975,
DSA-enabled) when using a bridge-vlan setup. This bug affects at least
OpenWRT 22.03.0-rc6.
I’m not sure whether this bug is related to this particular SoC or only
to DSA as I was unable to test with
It sets a hostname of
"OpenWrt-failsafe" in failsafe mode which is rendered in the shell's
prompt as a reminder of the mode during interactive failsafe use.
Previously, no hostname was set, which resulted in the kernel-default
hostname, "(none)", appearing in failsafe shell prompts.
at
package/kernel/mac80211/patches/build/001-fix_build.patch.
Signed-off-by: Mark Mentovai
---
package/libs/libmnl/patches/001-fix_build.patch | 11 +++
1 file changed, 11 insertions(+)
create mode 100644 package/libs/libmnl/patches/001-fix_build.patch
diff --git a/package/libs/libmnl
Hauke Mehrtens wrote:
On 6/21/22 16:38, Mark Mentovai wrote:
From: Mark Mentovai
This fixes the libmnl build on macOS, which ships with an outdated bash
at /bin/bash. During the OpenWrt build, a modern host bash is built and
made available at staging_dir/host/bin/bash, which is present before
nsoles (not just the last-defined one). It sets a hostname of
"OpenWrt-failsafe" in failsafe mode which is rendered in the shell's
prompt as a reminder of the mode during interactive failsafe use.
Previously, no hostname was set, which resulted in the kernel-default
hostname, "(n
From: Mark Mentovai
This fixes the libmnl build on macOS, which ships with an outdated bash
at /bin/bash. During the OpenWrt build, a modern host bash is built and
made available at staging_dir/host/bin/bash, which is present before
/bin/bash in the build's PATH.
This is similar to 8f7ce3aa6dda
ing interactive failsafe use.
Previously, no hostname was set, which resulted in the kernel-default
hostname, "(none)", appearing in failsafe shell prompts.
Signed-off-by: Mark Mentovai
---
.../files/lib/preinit/10_indicate_failsafe | 7 ++-
.../files/lib/preinit/30_failsafe_wa
util-linux getopt (such as one provided by an
OS) shadows the util-linux getopt in the PATH.
Signed-off-by: Mark Mentovai
---
include/prereq-build.mk | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/include/prereq-build.mk b/include/prereq-build.mk
index e1918f002787..1b
h multiple console= arguments,
including x86 and bcm27xx (Raspberry Pi).
Signed-off-by: Mark Mentovai
---
.../files/lib/preinit/30_failsafe_wait| 60 ++-
.../files/lib/preinit/99_10_failsafe_login| 15 ++---
2 files changed, 40 insertions(+), 35 deletions(-)
diff --git
ithout any need for reconfiguration can be shared by
devices with the more traditional (for x86) VGA console.
Signed-off-by: Mark Mentovai
---
.../files/lib/preinit/99_10_failsafe_login| 15 ---
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/package/base-files
r and group are statically allocated
uid 101 and gid 101, respectively, per
package/base-files/files/etc/passwd and USERID in
package/network/services/hostapd/Makefile, this patch also uses a
constant 101 for the uid and gid.
Cc: Daniel Golle
Signed-off-by: Mark Mentovai
---
.../610-hostapd_cli_ujai
There's no such thing as ucidef_set_interfaces_lan. It's
ucidef_set_interface_lan.
Cc: David Bauer
Signed-off-by: Mark Mentovai
---
target/linux/mediatek/mt7622/base-files/etc/board.d/02_network | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/target/linux/mediatek/mt7622
2825dd77b3 ipq806x: dwmac: set forced speed when using fixed-link
Signed-off-by: Mark Mentovai
Run-tested: ipq806x/ubnt,unifi-ac-hd
Cc: Ansuel Smith
---
.../082-ipq8064-dtsi-tweaks.patch | 36
...-dwmac-ipq806x-qsgmii-pcs-all-ch-ctl.patch | 83 +++
2 files ch
Signed-off-by: Mark Mentovai
Run-tested: ipq806x/ubnt,unifi-ac-hd
Cc: Ansuel Smith
---
.../ipq806x/patches-5.10/0069-arm-boot-add-dts-files.patch | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git
a/target/linux/ipq806x/patches-5.10/0069-arm-boot-add-dts-files.patch
b
Ansuel Smith wrote:
Tested-by: Ansuel Smith
I also tested this with the qca8k driver on 5.10. Also can you include
this with kernel 5.10
Yes, thanks. The broken fa731838c524 hasn't made it into patches-5.10 yet,
so I'm going to merge this patch, that one, and 75ca641f1b84 into an
update
d, setting the speed as directed by
fixed-link if present, and otherwise clearing it as was done previously.
Fixes: fa731838c524 ("ipq806x: dwmac: clear forced speed during probe")
Signed-off-by: Mark Mentovai
Tested-by: Hannu Nyman
Run-tested: ipq806x/ubnt,unifi-ac-hd, ipq806x/netgear,r7800
/ethernet/stmicro/stmmac/stmmac_platform.c
stmmac_probe_config_dt.
Signed-off-by: Mark Mentovai
Build-tested: ipq806x/ubnt,unifi-ac-hd
Run-tested: ipq806x/ubnt,unifi-ac-hd
---
.../ipq806x/patches-5.4/083-ipq8064-dtsi-additions.patch | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff
during dwmac initialization when
SGMII is in use, this port becomes usable.
This patch is upstreamable, and will be sent upstream after successful
testing in OpenWrt.
Signed-off-by: Mark Mentovai
Build-tested: ipq806x/ubnt,unifi-ac-hd
Run-tested: ipq806x/ubnt,unifi-ac-hd
---
...-dwmac-ipq806x
the at803x driver to load.
Signed-off-by: Mark Mentovai
Build-tested: ipq806x/ubnt,unifi-ac-hd
Run-tested: ipq806x/ubnt,unifi-ac-hd
---
.../arm/boot/dts/qcom-ipq8064-unifi-ac-hd.dts | 38 +--
1 file changed, 19 insertions(+), 19 deletions(-)
diff --git
a/target/linux/ipq806x/files
a switch on board, so there's no possibility
to remap ports via switch configuration. "ip link set $interface name"
is used instead, during preinit before networking is configured.
Signed-off-by: Mark Mentovai
Build-tested: ipq806x/ubnt,unifi-ac-hd
Run-tested: ipq806x/ubnt,unifi-ac-hd
---
.
:
- correctly configures that register,
- properly configures the on-board PHYs for both interfaces, and
- reorders eth0 and eth1 so that gmac2/MAIN is eth0 and gmac1/SECONDARY
is eth1.
Mark Mentovai (3):
ipq806x: dwmac: clear forced speed during probe
ipq806x: ubnt,unifi-ac-hd: use on-board PHYs
User-space firmware loading is handled by hotplug in procd. It’s directed
by /etc/hotplug.json. Paraphrasing:
[
[ "case", "ACTION", {
"add": [
[ "if",
[ "has", "FIRMWARE" ],
[
[ "exec", "/sbin/hotplug-call", "%SUBSYSTEM%" ],
[ "load-firmware", "/lib/firmware" ],
[ "return" ]
]
]
],
} ],
]
Felix, is this patch acceptable?
The v2 that made it to Patchwork applies cleanly to the trunk without any
whitespace problems.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
Felix Fietkau wrote:
> On 2015-11-21 22:43, Mark Mentovai wrote:
> > r47288 updated to Busybox 1.24.1 but did not update the configuration.
> >
> > The configuration is updated by running
> >
> > cd config
> > ../convert_menuconfig.pl .../b
Signed-off-by: Mark Mentovai <m...@moxienet.com>
---
package/utils/busybox/Config-defaults.in | 159 +++---
package/utils/busybox/config/archival/Config.in | 10 ++
package/utils/busybox/config/coreutils/Config.in | 116 +++--
package/utils/busybox/conf
ikely, all of OpenWrt’s busybox Config.in files need
to be updated in this way after a Busybox update.
Signed-off-by: Mark Mentovai <m...@moxienet.com>
diff --git a/package/utils/busybox/config/coreutils/Config.in
b/package/utils/busybox/config/coreutils/Config.in
index f50823f012de..e25
John Crispin wrote:
patch does not apply to current trunk
Sorry, the tabs got swallowed. I re-sent it.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Signed-off-by: Mark Mentovai <m...@moxienet.com>
---
package/utils/busybox/Config-defaults.in | 159 +++---
package/utils/busybox/config/archival/Config.in | 10 ++
package/utils/busybox/config/coreutils/Config.in | 116 +++--
package/utils/busybox/conf
rc. Likely, all of OpenWrt’s busybox Config.in files
need to be updated in this way after a Busybox update.
Signed-off-by: Mark Mentovai <m...@moxienet.com>
diff --git a/package/utils/busybox/config/coreutils/Config.in
b/package/utils/busybox/config/coreutils/Config.in
index f50823f012de..e25
I wrote:
OpenWrt’s busybox/config/coreutils/Config.in needs to mirror Busybox’ own
coreutils/Config.src. Likely, all of OpenWrt’s busybox Config.in files need
to be updated in this way after a Busybox update.
I posted this as an alternative, more complete patch, under the title
[PATCH]
Yousong Zhou wrote:
I just sent a patch for this with you in the cc list. Could you give it a
try and tell if it can work for you?
Thanks, Yousong. Your patch fixes this for me.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
In the latest OpenWrt trunk, I found that config_get has stopped loading
uncommitted uci changes from /tmp/.uci. I rely on this behavior, which had
worked well for years.
I found a change[1] in uci that’s responsible.
The uci change makes uci_add_delta_path() reject any attempt to add
deca1064 rps...@arcor.de wrote:
This patch adds platform machine support for the Netgear WNDR3700v4.
Signed-off-by: Ralph Perlich rps...@arcor.de
The WNDR3700v4 and WNDR4300 are almost identical, so much so that
nearly everything you’ve added here duplicates the definitions that
already exist
etienne.champet...@free.fr wrote:
zabbix-agentd select BUSYBOX_CONFIG_UNAME and BUSYBOX_CONFIG_HOSTNAME,
It's working ok in r39105 (419db8ef231eae6c0a514f32ff6c423c384900ca)(just
before busybox config changes),
but it's not in r39185 (13b222d757237eb7772eb7cf8433306ffd6b8ccd)(latest
I wrote:
Are you suggesting that dnsmasq should be started out of a hotplug
script?
I suppose the answer is “yes:” I see r39152.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
It seems that at some point in the last three months (between r37989 and
r39097), something about the way network interfaces are brought up changed.
Interfaces that were formerly available at a certain point during startup
are no longer necessarily available at that point.
I first noticed this
the problem. This seems to be how compound conditions are generally
handled in OpenWrt Makefiles.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: feeds/packages/net/nginx/Makefile
===
--- feeds/packages/net/nginx/Makefile
This works perfectly. Thanks, jow.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
reflection_src_dip? That matches src_dip as used for SNAT rules, but makes
it clear that it’s for reflection. (src_dip has a matching function instead
of a rewriting function for DNAT rules.)
I’ve got a strong preference to allow an interface name argument (“lan”)
instead of requiring an IP
This would work just fine for me, although configuration’s meaning wouldn’t
be nearly as evident without consulting some reference documentation.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
the up/down state
of its interface.
[1] https://lists.openwrt.org/pipermail/openwrt-devel/2012-August/016540.html
[2] https://lists.openwrt.org/pipermail/openwrt-devel/2012-October/016996.html
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index:
target/linux/ar71xx/patches-3.8/910-phy-add
Has the source address used for NAT reflection changed with firewall3?
At r35938, I’m seeing that when I attempt to connect from a host on my LAN
to a redirected port on my main router’s WAN address, the router reflects
the request back in to my LAN using its own WAN address as the source
Since the recent enhancements to NAT reflection were made, I found
that reflection stopped working on my networks. I was unable to
connect from hosts on a LAN to its router’s WAN-side address and have
the port mappings from “config redirect” sections be effective.
Instead, I got “connection
Florian Fainelli wrote:
On Tuesday 21 August 2012 12:32:17 Mark Mentovai wrote:
Changelog: http://nginx.org/en/CHANGES-1.2
Signed-off-by: Mark Mentovai m...@moxienet.com
Applied in r34495, thanks Mark!
Thanks, Florian. They’re up to 1.2.5 now, so I’ve moved on to using this.
Signed-off
Weedy weedy2...@gmail.com wrote:
On 27/08/12 07:34 AM, Gabor Juhos wrote:
The problem is that PHYLIB in Linux does not fully stops the PHY device
when a
driver issues a phy_stop call.
Copy the attached patches into 'target/linux/ar71xx/patches-3.3', then
do a
'make target/linux/clean
)
@@ -0,0 +1,54 @@
+include $(TOPDIR)/rules.mk
+
+PKG_NAME:=ifplugd_support
+PKG_VERSION:=0.1
+PKG_RELEASE:=1
+
+include $(INCLUDE_DIR)/package.mk
+
+define Package/ifplugd_support
+ SECTION:=base
+ CATEGORY:=Custom
+ TITLE:=ifplugd support
+ MAINTAINER:=Mark Mentovai m...@moxienet.com
+ DEPENDS
Adam Gensler wrote:
I recently acquired a Netgear WNDR3800. To date I've been playing
exclusively on an alix 2D13 platform so I have a few new things to learn.
In any case, I've been tinkering with the LEDs and I'm having some trouble
getting my head around a few things. Specifically:
1.
Gabor Juhos wrote:
The problem is that PHYLIB in Linux does not fully stops the PHY device when a
driver issues a phy_stop call.
Copy the attached patches into 'target/linux/ar71xx/patches-3.3', then do a
'make target/linux/clean world'. With these patches, the PHY of the WAN port
will be
Is there a way to make the link state drop on an ag71xx interface, so that
it appears to whatever equipment is connected to the port that the cable
has been disconnected?
I've got a WNDR3700-family device and I need to make it appear to the
equipment connected to its eth1 interface that the
John Crispin wrote:
not merging any new stuff and/or updates atm ...
--
https://lists.openwrt.org/pipermail/openwrt-devel/2012-August/016411.html
is there a specific reason, such as a security fix or is it just a
feature update ?
No security fixes here, just bug fixes. It's OK to hold
Jo-Philipp Wich wrote:
It looks like conffiles can only save files, not entire directories.
/etc/nginx is a directory. Should conffiles be fixed to work with
directories too?
It works fine with directories, see include/package-ipkg.mk, the code
below ifneq ($$(KEEP_$(1)),)
I see. The
Jo-Philipp Wich wrote:
But they will get recorded by opkg, which is another source sysupgrade
uses to assemble the list of files to get backed up.
This is the part I was missing. Thanks for the explanation. I've submitted
a v2 patch in case it's thought to be generally useful, otherwise I'll
This allows additional files in /etc/nginx/, such as SSL certificates,
tobe saved on sysupgrade.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index:
packages/net/nginx/Makefile
===
--- packages/net/nginx/Makefile (revision
John Crispin wrote:
On 24/08/12 15:14, Mark Mentovai wrote:
This allows additional files in /etc/nginx/, such as SSL certificates,
tobe saved on sysupgrade.
Signed-off-by: Mark Mentovai m...@moxienet.com
after not reading the 100 mails leading up to this patch i expected
something
Jo-Philipp Wich wrote:
Hi.
You do not need to ship a keep file, it is enough to add /etc/nginx to
the conffiles define.
It looks like conffiles can only save files, not entire directories.
/etc/nginx is a directory. Should conffiles be fixed to work with
directories too?
Changelog: http://nginx.org/en/CHANGES-1.2
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: packages/net/nginx/Makefile
===
--- packages/net/nginx/Makefile (revision 33213)
+++ packages/net/nginx/Makefile (working copy
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: packages/net/nginx/Makefile
===
--- packages/net/nginx/Makefile (revision 33213)
+++ packages/net/nginx/Makefile (working copy)
@@ -103,6 +103,8 @@
$(INSTALL_DATA
Felix Fietkau wrote:
I did add code to netifd to take care of this a while ago, but there was
a bug that prevented it from working. This bug is fixed in the latest
version, committed in r32506
Great, I confirmed that this is working as it used to now. Thanks!
I suspect that the netifd changes are related, since that looks like the
only relevant area of major activity in the past month when this began
happening. Then again, the timing-sensitive nature means that the
underlying problem may have been present for a while, and only exposed
with the
Philip Prindeville wrote:
On 6/26/12 11:02 AM, Jo-Philipp Wich wrote:
That is nothing netifd or OpenWrt does, it is the behaviour of Linux
bridges - they'll assume the lowest MAC address of all their member ifaces.
Imo that should be changed in the Kernel, it shoud fix the bridge MAC to
the sticky bit set.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: include/image.mk
===
--- include/image.mk (revision 31782)
+++ include/image.mk (working copy)
@@ -146,7 +146,7 @@
- $(FIND) $(TARGET_DIR
Are you sure this is universally correct for all WNDRMAC units? It's
possible that the N there is actually variable, and part of something
else that shows up at that location in flash. `hexdump -C /dev/mtd5`
(assuming mtd5 is art in /proc/mtd) should help identify what that N is
actually
||
My guess is that this should work on all WNDRMACs, although I only have one
device to test this.
15.06.2012 20:19, Mark Mentovai пишет:
Are you sure this is universally correct for all WNDRMAC units? It's
possible that the N there is actually variable, and part of something
else
.
Local hostnames will remain available for lookup using fully-qualified names.
Signed-off-by: Mark Mentovai m...@moxienet.com
Index: package/dnsmasq/files/dnsmasq.init
===
--- package/dnsmasq/files/dnsmasq.init (revision 31782
.
The default, on, produces the current behavior. Setting it to off avoids
adding a search line to resolv.conf. option domain still retains its other
functions such as providing the value for the --domain option to dnsmasq.
Signed-off-by: Mark Mentovai m...@moxienet.com
Index: package/dnsmasq/files
, it is obsolete and should be removed from the dnsmasq package.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: package/dnsmasq/Makefile
===
--- package/dnsmasq/Makefile(revision 31782)
+++ package/dnsmasq/Makefile
Very well, I'll drop this. I missed the conf-file=/etc/dnsmasq.conf at
the top of /var/etc/dnsmasq.conf.
I'm still proposing the other four dnsmasq patches I sent to the list
in a series earlier today.
___
openwrt-devel mailing list
.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: include/image.mk
===
--- include/image.mk(revision 31782)
+++ include/image.mk(working copy)
@@ -142,7 +142,7 @@
define Image/mkfs/prepare/default
# Use
On the off chance that the root filesystem's /tmp is used directly as a
temporary directory instead of having a tmpfs mounted over it, it should have
the sticky bit set.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: include/image.mk
Changelog: http://nginx.org/en/CHANGES-1.0
This upgrade includes a fix for a security vulnerability, CVE-2012-2089.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: packages/net/nginx/Makefile
===
--- packages/net/nginx
Upgrade nginx to 1.0.14.
Changelog: http://nginx.org/en/CHANGES-1.0
This upgrade includes a fix for a major security vulnerability,
CVE-2012-1180.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: packages/net/nginx/Makefile
Dave Taht wrote:
[PATCH 4/4] Add local mac support to wndr3700 so as to be able to
route not bridge
The wndr3700 at least has no eth0 mac address and usually leverages
the first wireless device's mac when in a bridged scenario. If,
however, you want to route, and not bridge the interfaces,
. It also will not mis-detect
units on which people install more memory.
I have tested this patch on WNDR3700 (v1), WNDR3700v2, and WNDR3800.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: target/linux/ar71xx/base-files/lib/ar71xx.sh
to set the device: field. In r29434, this was
erroneously changed to be WNDR3700 for all models. The tools to flash
factory images (U-Boot's TFTP server and the factory software's upgrade
utility) may refuse to honor images with incorrect device: fields in their
DNI tags.
Signed-off-by: Mark Mentovai
I concur that there's something wrong with mtd -j jffs write, as
used by sysupgrade when preserving configuration. The CRC in the TRX
header is computed before all writes that the CRC protects have
completed. I've worked around this problem in my brcm47xx builds by
running mtd fixtrx as part of
Jonathan McCrohan wrote:
Is there even a need to produce an oversize image warning? I'm not aware
of any other architecture that refuses to build over a certain size. x86
being the obvious exception, but that is different because it doesn't
target any particular device, but a whole
I just had a chance to flash a WNDR3800 with an image that contains
this “machine name” patch (post-r29326). It's broken in two ways that
I can see so far. These are regressions. The WNDR3800 images that used
board=WNDR3700v2 worked properly in these cases. It remains my
recommendation that this
Jonathan Bennett wrote:
This patch seems like a really good idea for the ar71xx family. What
sorts of problems are preventing this from being implemented? As small
as it sounds, an extra 192k would be really useful on quite a few
routers.
I don't know of any actual problems with the patch.
Gabor Juhos wrote:
2011.11.16. 21:20 keltezéssel, Petri Rosenström írta:
This fixes the machine name in /proc/cpuinfo and luci status page machine
name.
Signed-off-by: Petri Rosenström petri.rosenst...@gmail.com
Applied, thanks!
I don't think this patch is a great idea. There are no
Petri Rosenström wrote:
I think it is useful because it adds the specific board information
about WNDR3800 (e.g. the name WNDR3800 on status page or /proc/cpuinfo
show correct board name). I think that this information is important
and it should be available.
Since you can flash a WNDR3700v2
Dave Taht wrote:
Awesome. However I'd like to somehow make for fully field-upgradable
kernels for this device (how to do that?), and reserving 64k strikes
me as too small to account for future growth.
It doesn't reserve 64kB. It doesn't really reserve anything at all.
There can be as little as
/8781.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: target/linux/ar71xx/image/Makefile
===
--- target/linux/ar71xx/image/Makefile (revision 27926)
+++ target/linux/ar71xx/image/Makefile (working copy)
@@ -42,13 +42,13
Signed-off-by: Mark Mentovai m...@moxienet.com
--
Index: toolchain/gcc/patches/linaro/860-fix_extension_elimination.patch
===
--- toolchain/gcc/patches/linaro/860-fix_extension_elimination.patch
(revision 27096)
+++ toolchain/gcc
toolchain/gcc: upgrade Linaro GCC to 4.5-2011.05-0
Signed-off-by: Mark Mentovai m...@moxienet.com
linaro_4.5-2011.05-0.patch
Description: Binary data
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman
-NA and worldwide images, neither of
which are tagged with hd_id.
Signed-off-by: Mark Mentovai m...@moxienet.com
wndr3700v2_images.patch
Description: Binary data
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org
Gábor-
I’d like your thoughts (and others’) on my proposed patch at
https://dev.openwrt.org/ticket/8781
Currently, the WNDR3700 images reserve 1MB for the kernel, whether or
not the kernel is actually anywhere near that size. It can be smaller,
in which case space is wasted, or it can in theory
the discrepancy between the two places that
ieee80211d could be set.
Signed-off-by: Mark Mentovai m...@moxienet.com
ieee80211d.patch
Description: Binary data
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo
Jo-Philipp Wich wrote:
Here's my proposal:
I like the outcome but I find the logic a bit confusing. ${country_ie}
starts out as the value of the country parameter on the wifi-device
(empty or a country string), but then becomes either 0 or 1, a boolean
that controls enabling the feature in
This applies Richard Sandiford's patch for Linaro GCC as an
alternative to disabling the Linaro-specific extension elimination
optimization altogether.
Original patch: https://bugs.launchpad.net/gcc-linaro/+bug/728315
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: toolchain/gcc
-by: Mark Mentovai m...@moxienet.com
---
Index: toolchain/gcc/patches/linaro/860-fix_extension_elimination.patch
===
--- toolchain/gcc/patches/linaro/860-fix_extension_elimination.patch
(revision 0)
+++ toolchain/gcc/patches/linaro/860
(Third time's the charm? Sorry for the spam.)
This applies Richard Sandiford's patch for Linaro GCC as an alternative to
disabling the Linaro-specific extension elimination optimization altogether.
Original patch: https://bugs.launchpad.net/gcc-linaro/+bug/728315
Signed-off-by: Mark Mentovai m
Felix Fietkau wrote:
Scratch that, got confused by the delay slot for a second there :)
I see the code that adds the wrong bits now, I wonder if there's
something wrong with signed vs unsigned type propagation...
I'll let you know when I find out more about this.
I agree that signed/unsigned
Felix Fietkau wrote:
Actually, it emits -16384 in both the working and non-working case,
I think what makes the difference of working vs non-working is this:
Broken: (0xc000 is in v0)
- e8:00021a02srl v1,v0,0x8
- ec:00021200sll v0,v0,0x8
documented in the UCI
wireless configuration page on the wiki. When this change is
committed, I'll update the documentation as needed.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: package/mac80211/files/lib/wifi/mac80211.sh
I received a request for more information, so in case anyone else is
interested, here’s what was happening.
The problem occurs in hostapd src/ap/ieee802_11.c’s send_assoc_resp.
The C statement that triggers the compiler bug is
reply-u.assoc_resp.aid = host_to_le16((sta ? sta-aid : 0)
bjorn.smed...@venatech.se:
Wouldn't it make more sense to configure beacon interval per
wifi-iface? This is how mac80211 does it and there is at least talk
about supporting different beacon intervals for different ap vifs with
ath9k.
It’s configured via hostapd, which only supports the beacon
with this compiler, I'm not sending this patch
upstream to hostapd at this point.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: package/hostapd/patches/001-mips_linaro_gcc45_codegen.patch
===
--- package/hostapd/patches/001
base-files: librt should depend on libpthread, not the other way around
In uClibc 0.9.32 as well as recent versions of glibc and eglibc, librt
depends on libpthread.
Signed-off-by: Mark Mentovai m...@moxienet.com
---
Index: package/base-files/Makefile
1 - 100 of 118 matches
Mail list logo