Felix Fietkau [Sun, 10 May 2015 11:46:56 +0000 (11:46 +0000)]
ipq806x: fix boot freeze on zImage kernel
ARCH_QCOM is using the ARCH_MULTIPLATFORM option, as now recommended
on most ARM architectures. This automatically calculate ZRELADDR by
masking PHYS_OFFSET with 0xf8000000.
On IPQ806x though, the first ~20MB of RAM is reserved for the hardware.
In newer bootloader, when DT is used, this is not a problem, we just
reserve this memory in the device tree. But if the bootloader doesn't
have DT support, then ATAGS have to be used. In this case, the ARM
decompressor will position the kernel in this low mem, which will not be
in the RAM section mapped by the bootloader, which means the kernel will
freeze in the middle of the boot process trying to map the memory.
As a work around, this patch allows disabling AUTO_ZRELADDR when
ARCH_QCOM is selected. It makes the zImage usage possible on bootloaders
which don't support device-tree, which is the case on certain early
IPQ806x based designs.
Felix Fietkau [Sun, 10 May 2015 11:46:41 +0000 (11:46 +0000)]
generic: add CRYPTO_DEV_QCE option to default config
This option has been added in kernel 3.17. It shows-up only when both
ARCH_QCOM and CRYPTO are enabled. So we'll disable these two by default
to avoid stalling the build when these conditions are met.
Felix Fietkau [Sun, 10 May 2015 11:17:29 +0000 (11:17 +0000)]
build: don't call prereq for any package/symlinks rules
Most of the time, we want to make sure OpenWrt has been configured and
setup before start running make. However, in case of package/symlinks,
forcing prereq as a dependency creates multiple issues:
*when executed on a clean workspace, it will prompt for user input
and open a menuconfig window before executing the feeds command
*the only way around that is to provide a .config. However, the "prereq"
target would then run a "make defconfig", which will remove all the
packages in the .config but from external feeds, as feeds have not been
installed yet.
The only way to currently work around this, is to generate a fake config
by running "make defconfig", then "make package/symlinks", copy the real
config (which at this point disregards the previously generated config),
and run make defconfig again. Something like this:
make defconfig
make package/symlinks
cp real.config .config
make defconfig
This change is removing the need for the first defconfig, making the
process more logical for OpenWrt users using the package/symlinks target.
Felix Fietkau [Fri, 8 May 2015 14:02:23 +0000 (14:02 +0000)]
ar71xx/image: move TPLINK-LZMA image to new build code
There are 2 images missing: TLWR2543 TLWR1043V2 which have special properties.
v2: set correct DEVICE_PROFILES for all images.
v2: migrate TP-LINK TL-WR710N v2 which was committed after v1.
v2: split very very long line `TARGET_DEVICES +=` into smaller parts
Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
SVN-Revision: 45643
Felix Fietkau [Fri, 8 May 2015 12:23:45 +0000 (12:23 +0000)]
ar71xx: fix ethernet on wnr2000-v4
Most people report broken ethernet with upstream. Last year, user "franz.flasch"
authored a working mach-file. His patch is outdated so I modernized it. Original
patch and user commentary on page 1:
https://forum.openwrt.org/viewtopic.php?pid=260861#p260861
I have figured out what the critical differences are between the two that caused
upstream ethernet to break.
1) Both ath79_init_mac() functions calls must be invocated before any GMAC init
2) must init GMAC0 before GMAC1
That was enough to get upstream to function, but I wanted to enjoy my confidence
having tested franz's patch for a week sucessfully, so I put his whole
function in, which only features more differences in order of function calls.
An expert should consider these changes, which could pose potential bugs/issues:
1) No longer using the flag AR934X_ETH_CFG_SW_PHY_SWAP in the
ath79_setup_ar934x_eth_cfg() call.
2) Possible consequence of no longer explicitly setting ethernet duplex/speed.
Review: With this patch, my ethernet and wireless works.
Signed-off-by: Michael J. Bazzinotti <mbazzinotti@gmail.com>
SVN-Revision: 45633
Felix Fietkau [Fri, 8 May 2015 10:44:08 +0000 (10:44 +0000)]
trace-cmd: update to 2.4.2
fixes compile error:
(observed w. target arm64, all kernel build opts enabled)
error: call to '__open_missing_mode' declared with attribute
error: open with O_CREAT in second argument needs 3 arguments
Felix Fietkau [Fri, 8 May 2015 10:43:48 +0000 (10:43 +0000)]
ustream-ssl: correct year in PKG_VERSION string
ustream-ssl: correct the year in the PKG_VERSION string, as both r45157 and
r45441 left the old year 2014 there. For a casual user it may seem that the
current code is from April 2014, although a4ca61527236e89eb9efb782fd9bfd04796144e3 is from April 2015.
Steven Barth [Tue, 5 May 2015 13:22:33 +0000 (13:22 +0000)]
netsupport: l2tp-ip: only depend on IPv6 if IPv6 support is enabled
Before r45593 kmod-l2tp-ip did not depend on kmod-ipv6.
With r45593 support for L2TP IPv6 encapsulation was added and
included in the kmod-l2tp-ip package. This change also
added the dependency to kmod-ipv6 to kmod-l2tp-ip, regardless
of whether the user chose to generally include IPv6 support
or not.
Change this so L2TP over IPv6 and the resulting dependency
to kmod-ipv6 is only included in kmod-l2tp-ip if IPv6 support
is enabled.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
SVN-Revision: 45612
Jo-Philipp Wich [Mon, 4 May 2015 08:49:21 +0000 (08:49 +0000)]
openvpn: disable CBC record splitting in PolarSSL/mbedTLS (#19101)
OpenVPN assumes that its control channel messages are sent and received
unfragmented, this assumption is broken when CBC record splitting is
enabled in mbedTLS.
The record splitting is intended as countermeasure against BEAST attacks
which do not apply to OpenVPN, therefore we simply disable it until
upstream OpenVPN gains the ability to process fragmented control
messages.
Disabling the splitting also works around a (not remotely triggerable)
segmentation fault in mbedTLS.
Jonas Gorski [Tue, 28 Apr 2015 13:31:08 +0000 (13:31 +0000)]
kernel: disable INET_LRO by default
INET_LRO deprecated and there are exactly two drivers using it, neither
being included in any of the targets enabling INET_LRO. At the same time
both drivers needing it select it.
So just disable it for everyone.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 45584
This reverts commit ff84c27a281bc19df19bc62ee8688cca5586f6e3.
This tool has really broken size handling (many values hardcoded), it
crashes right now in case of NVRAM not filling whole MTD partition.
For years Broadcom devices use 64 KiB NVRAM partition size and some of
them indeed have it filled in more than 50%. This change allows handling
whole NVRAM e.g. on Netgear WNDR4500 and Netgear R8000.
The same fix was applied to kernel in upstream commit 6ab7c29.
projects / thirdparty / openwrt.git / log
