Re: [OpenWrt-Devel] [PATCH] package/madwifi: work-around broken frequency-list and txpower-list
On 2011-07-18 3:54 PM, Daniel Golle wrote: Hi! When using madwifi, the txpowerlist and freqlist seem to be uninitialized unless a vif is created and ifup'ed. Example for txpowerlist: root@OpenWrt:~# iwinfo wifi0 txpowerlist * 0 dBm ( 1 mW) 10 dBm ( 10 mW) 15 dBm ( 31 mW) 20 dBm ( 100 mW) 25 dBm ( 316 mW) 30 dBm (1000 mW) 35 dBm (3162 mW) 40 dBm (1 mW) root@OpenWrt:~# wlanconfig ath0 create wlandev wifi0 mode sta root@OpenWrt:~# ifconfig ath0 up root@OpenWrt:~# wlanconfig ath0 destroy root@OpenWrt:~# iwinfo wifi0 txpowerlist * 0 dBm ( 1 mW) 8 dBm ( 6 mW) 10 dBm ( 10 mW) 12 dBm ( 15 mW) 14 dBm ( 25 mW) 16 dBm ( 39 mW) 18 dBm ( 63 mW) 20 dBm ( 100 mW) To work-around this bug, I created a simple way to initialize the wifi driver even if the device is disabled and/or there are no vifs, so they there are sane values in the webinterface from the first moment on and the users won't get too confused. I guess the right way[tm] to do this would be fixing madwifi, but we all know that won't happen... How about putting this into a separate init script instead of adding hackery to a generic script? - Felix ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 1/1] ledtrig-ide-disk: make ledtrig-ide-disk modular
On 2011-05-15 7:51 AM, Philip Prindeville wrote: Normally ledtrig-ide-disk requires ide-disk.c to be compiled knowing that ledtrig-ide-disk is enabled. This is less than optimal, since you should be able to add ledtrig targets later and/or separately. So we work around this by having a hook that can be set, in the same way that netfilters and ip routing do. Please forward port to newer kernel versions and resubmit. - Felix ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] linux 3.0 is out - patch for generic and brcm47xx
On 07/24/2011 01:37 AM, Philip Prindeville wrote: On 7/22/11 7:07 AM, Peter Wagner wrote: Hi, this patch updates generec and brcm47xx to linux 3.0. runs on brcm47xx like a charm (for a few minutes by now ;) ) /Peter Wagner Signed-off-by: Peter Wagner tripo...@gmx.at I couldn't actually get it to build: Applying /home/philipp/openwrt2/target/linux/generic/patches-3.0/020-ssb_update.patch using plaintext: patching file drivers/ssb/driver_chipcommon_pmu.c patching file drivers/ssb/driver_gige.c patching file drivers/ssb/driver_pcicore.c Hunk #4 FAILED at 514. 1 out of 4 hunks FAILED -- saving rejects to file drivers/ssb/driver_pcicore.c.rej patching file drivers/ssb/main.c patching file drivers/ssb/pcihost_wrapper.c patching file drivers/ssb/scan.c Patch failed! Please fix /home/philipp/openwrt2/target/linux/generic/patches-3.0/020-ssb_update.patch! make[3]: *** [/home/philipp/openwrt2/build_dir/toolchain-i386_gcc-linaro_eglibc-2.12/linux-3.0/.prepared] Error 1 make[3]: Leaving directory `/home/philipp/openwrt2/toolchain/kernel-headers' make[2]: *** [toolchain/kernel-headers/prepare] Error 2 I'm trying to vet the OCF version bump for linux-3.0 but I can't do that since 3.0 itself isn't in a known/good state. Hi Philip, This patch is already in trunk (see r27722) and for me all the patches are applying. Are you sure you are using linux-3.0 and not some rc version? Hauke ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] package/madwifi: work-around broken frequency-list and txpower-list
On Sun, Jul 24, 2011 at 10:53:42AM +0200, Felix Fietkau wrote: On 2011-07-18 3:54 PM, Daniel Golle wrote: To work-around this bug, I created a simple way to initialize the wifi driver even if the device is disabled and/or there are no vifs, so they there are sane values in the webinterface from the first moment on and the users won't get too confused. I guess the right way[tm] to do this would be fixing madwifi, but we all know that won't happen... How about putting this into a separate init script instead of adding hackery to a generic script? A seperate init script needs additional hackery in LuCI/ucitrigger to get called once the user changes the affected config entries. That would be even more overkill than my previous solution ;) I figured the most simple and non-intrusive way to achieve the desired result is moving things inside /lib/madwifi.sh: As scan_atheros gets called before enable_atheros in every case, the parts in enable_atheros to set countrycode, regdomain and outdoor can simply be moved over to scan_atheros. This results in the settings being applied even if the interface is disabled or without any vifs. Obviously, no changes to /sbin/wifi are needed to do that. Signed-off-by: Daniel Golle dgo...@allnet.de Index: package/madwifi/files/lib/wifi/madwifi.sh === --- package/madwifi/files/lib/wifi/madwifi.sh (revision 27690) +++ package/madwifi/files/lib/wifi/madwifi.sh (working copy) @@ -6,6 +6,24 @@ local wds local adhoc ahdemo sta ap monitor + config_get regdomain $device regdomain + [ -n $regdomain ] echo $regdomain /proc/sys/dev/$device/regdomain + + config_get country $device country + [ -z $country ] country=0 + echo $country /proc/sys/dev/$device/countrycode + + config_get_bool outdoor $device outdoor 0 + echo $outdoor /proc/sys/dev/$device/outdoor + + ifname=$(wlanconfig ${device}-temp create wlandev $device wlanmode ap) + [ $? -ne 0 ] { + echo scan_atheros($device): Failed to set up temporary vif 2 + continue + } + ifconfig $ifname up + wlanconfig $ifname destroy + config_get vifs $device vifs for vif in $vifs; do @@ -68,16 +86,6 @@ enable_atheros() { local device=$1 - config_get regdomain $device regdomain - [ -n $regdomain ] echo $regdomain /proc/sys/dev/$device/regdomain - - config_get country $device country - [ -z $country ] country=0 - echo $country /proc/sys/dev/$device/countrycode - - config_get_bool outdoor $device outdoor 0 - echo $outdoor /proc/sys/dev/$device/outdoor - config_get channel $device channel config_get vifs $device vifs config_get txpower $device txpower ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] linux 3.0 is out - patch for generic and brcm47xx
Am Sonntag, 24. Juli 2011, 01:37:46 schrieb Philip Prindeville: On 7/22/11 7:07 AM, Peter Wagner wrote: Hi, this patch updates generec and brcm47xx to linux 3.0. runs on brcm47xx like a charm (for a few minutes by now ;) ) /Peter Wagner Signed-off-by: Peter Wagner tripo...@gmx.at I couldn't actually get it to build: Applying /home/philipp/openwrt2/target/linux/generic/patches-3.0/006-arm_kernel_xz_ support.patch using plaintext: patching file arch/arm/Kconfig patching file arch/arm/boot/compressed/Makefile patching file arch/arm/boot/compressed/decompress.c patching file arch/arm/boot/compressed/piggy.xzkern.S patching file lib/xz/xz_dec_stream.c Applying /home/philipp/openwrt2/target/linux/generic/patches-3.0/010-mips_wire_up_n ew_sendmmsg_syscall.patch using plaintext: patching file arch/mips/include/asm/unistd.h patching file arch/mips/kernel/scall32-o32.S patching file arch/mips/kernel/scall64-64.S patching file arch/mips/kernel/scall64-n32.S patching file arch/mips/kernel/scall64-o32.S Applying /home/philipp/openwrt2/target/linux/generic/patches-3.0/020-ssb_update.pat ch using plaintext: patching file drivers/ssb/driver_chipcommon_pmu.c patching file drivers/ssb/driver_gige.c patching file drivers/ssb/driver_pcicore.c Hunk #4 FAILED at 514. 1 out of 4 hunks FAILED -- saving rejects to file drivers/ssb/driver_pcicore.c.rej patching file drivers/ssb/main.c patching file drivers/ssb/pcihost_wrapper.c patching file drivers/ssb/scan.c Patch failed! Please fix /home/philipp/openwrt2/target/linux/generic/patches-3.0/020-ssb_update.pat ch! make[3]: *** [/home/philipp/openwrt2/build_dir/toolchain-i386_gcc-linaro_eglibc-2.12/li nux-3.0/.prepared] Error 1 make[3]: Leaving directory `/home/philipp/openwrt2/toolchain/kernel-headers' make[2]: *** [toolchain/kernel-headers/prepare] Error 2 I'm trying to vet the OCF version bump for linux-3.0 but I can't do that since 3.0 itself isn't in a known/good state. ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel 3.0 is allready in trunk -just update your svn co/git clone /Peter ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v4 1/1] ocf-linux: version bump to 20110720
On 07/24/2011 12:55 AM, Philip Prindeville wrote: Fixes problem with TFM allocation in cryptosoft.c When checking in, please do: svn rename patches-2.6.39/941-ocf_20105030.patch patches-2.6.39/941-ocf_20110720.patch for each patches directory before committing (but after applying patch, obviously). Redux to include 2.6.30 - 32, as per Hauke's comment. Signed-off-by: Philip Prindeville phil...@redfish-solutions.com Hi Philip, I applied the patch with some changes in r27753. * remove ubsec_ssb package and take it from ocf-linux * use patches from ocf-linux package * refresh all patches * readd some build fixes for OpenWrt. * readd CRYPTO_MANAGER dependency Hauke ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] Fix detection of standalone ldd
When standalone ldd is selected in config the binary is installed in /usr/bin/ldd. Signed-off-by: Michael Heimpold m...@heimpold.de --- Index: package/base-files/files/etc/profile === --- package/base-files/files/etc/profile(revision 27363) +++ package/base-files/files/etc/profile(working copy) @@ -12,4 +12,4 @@ [ -z $KSH_VERSION -o \! -s /etc/mkshrc ] || . /etc/mkshrc [ -x /sbin/arp ] || arp() { cat /proc/net/arp; } -[ -z /bin/ldd ] || ldd() { LD_TRACE_LOADED_OBJECTS=1 $*; } +[ -x /usr/bin/ldd ] || ldd() { LD_TRACE_LOADED_OBJECTS=1 $*; } ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 1/1] geos: include pppstats and pppdump in default packaging
On 7/24/11 1:42 AM, Felix Fietkau wrote: On 2011-07-23 4:15 AM, Philip Prindeville wrote: As the Geos box includes an ADSL modem that's targeted towards PPPoE, PPPoA, and bridged utilization, including these useful tools allows for easier configuration and troubleshooting. Signed-off-by: Philip Prindevillephil...@redfish-solutions.com Applied in r27743 Thanks! ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 1/1] ppp: version bump to 2.4.5
On 7/24/11 2:34 AM, Felix Fietkau wrote: On 2011-07-22 7:58 AM, Philip Prindeville wrote: On 7/6/11 1:41 AM, Philip Prindeville wrote: Bump to the current version of PPP. A few of the patches are either in 2.4.5 or something functionally equivalent is, so those patches were removed. 1 patch was added to accommodate pppol2pv3_addr which is referenced in a linux kernel header file, but isn't defined in the kernel header files packaged with PPP. Signed-off-by: Philip Prindevillephil...@redfish-solutions.com Can I please get a code review of this? Here's a few things spotted during review: - You forgot to refresh the patches, there's lots of fuzz/offsets when applying them - The change that removes the exponential backoff during pppoe discovery was dropped, please add it back Looking through the code, the patch was failing to comment that snippet out because it's no longer present. - Did you test whether pppd still properly survives errors during pppoe initialization (e.g. when receiving broken data from the AC)? - Felix I only have access to PPPoA and Bridged DSL functionality. I put out a request for people to test the patch but didn't get any responses. -Philip ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] linux 3.0 is out - patch for generic and brcm47xx
On 7/24/11 3:02 AM, Hauke Mehrtens wrote: On 07/24/2011 01:37 AM, Philip Prindeville wrote: On 7/22/11 7:07 AM, Peter Wagner wrote: Hi, this patch updates generec and brcm47xx to linux 3.0. runs on brcm47xx like a charm (for a few minutes by now ;) ) /Peter Wagner Signed-off-by: Peter Wagner tripo...@gmx.at I couldn't actually get it to build: Applying /home/philipp/openwrt2/target/linux/generic/patches-3.0/020-ssb_update.patch using plaintext: patching file drivers/ssb/driver_chipcommon_pmu.c patching file drivers/ssb/driver_gige.c patching file drivers/ssb/driver_pcicore.c Hunk #4 FAILED at 514. 1 out of 4 hunks FAILED -- saving rejects to file drivers/ssb/driver_pcicore.c.rej patching file drivers/ssb/main.c patching file drivers/ssb/pcihost_wrapper.c patching file drivers/ssb/scan.c Patch failed! Please fix /home/philipp/openwrt2/target/linux/generic/patches-3.0/020-ssb_update.patch! make[3]: *** [/home/philipp/openwrt2/build_dir/toolchain-i386_gcc-linaro_eglibc-2.12/linux-3.0/.prepared] Error 1 make[3]: Leaving directory `/home/philipp/openwrt2/toolchain/kernel-headers' make[2]: *** [toolchain/kernel-headers/prepare] Error 2 I'm trying to vet the OCF version bump for linux-3.0 but I can't do that since 3.0 itself isn't in a known/good state. Hi Philip, This patch is already in trunk (see r27722) and for me all the patches are applying. Are you sure you are using linux-3.0 and not some rc version? Hauke Actually, I had reverted because some patches I had submitted that were still in my tree were conflicting with an update. I'll update and resolve the conflicts. -Philip ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v4 1/1] ocf-linux: version bump to 20110720
On 7/24/11 7:54 AM, Hauke Mehrtens wrote: On 07/24/2011 12:55 AM, Philip Prindeville wrote: Fixes problem with TFM allocation in cryptosoft.c When checking in, please do: svn rename patches-2.6.39/941-ocf_20105030.patch patches-2.6.39/941-ocf_20110720.patch for each patches directory before committing (but after applying patch, obviously). Redux to include 2.6.30 - 32, as per Hauke's comment. Signed-off-by: Philip Prindeville phil...@redfish-solutions.com Hi Philip, I applied the patch with some changes in r27753. * remove ubsec_ssb package and take it from ocf-linux What was the problem with ubsec_ssb? * use patches from ocf-linux package I had used the patches from ocf-linux... the change I made was to replace linux-2.6.??.orig/ with a/ and linux-2.6.??/ with b/ in the --- and +++ lines... but other than that they should have been identical. * refresh all patches * readd some build fixes for OpenWrt. What build fixes were these? * readd CRYPTO_MANAGER dependency For cryptosoft? According to the Kconfig that comes with the latest ocf-linux, it no longer has a dependency on crypto-manager, that's why I took it out. Hauke ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 1/1] ppp: version bump to 2.4.5
On 2011-07-24 9:58 PM, Philip Prindeville wrote: On 7/24/11 2:34 AM, Felix Fietkau wrote: On 2011-07-22 7:58 AM, Philip Prindeville wrote: On 7/6/11 1:41 AM, Philip Prindeville wrote: Bump to the current version of PPP. A few of the patches are either in 2.4.5 or something functionally equivalent is, so those patches were removed. 1 patch was added to accommodate pppol2pv3_addr which is referenced in a linux kernel header file, but isn't defined in the kernel header files packaged with PPP. Signed-off-by: Philip Prindevillephil...@redfish-solutions.com Can I please get a code review of this? Here's a few things spotted during review: - You forgot to refresh the patches, there's lots of fuzz/offsets when applying them - The change that removes the exponential backoff during pppoe discovery was dropped, please add it back Looking through the code, the patch was failing to comment that snippet out because it's no longer present. I checked. The 'timeout *= 2' part is still there. - Did you test whether pppd still properly survives errors during pppoe initialization (e.g. when receiving broken data from the AC)? - Felix I only have access to PPPoA and Bridged DSL functionality. I put out a request for people to test the patch but didn't get any responses. You could put in a test change that simulates an occasional connection failure. I think this should be tested because an old patch that was removed touched parts of that and I want to make sure that the update does not cause a regression there. - Felix ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 1/1] ppp: version bump to 2.4.5
On 7/24/11 1:07 PM, Felix Fietkau wrote: On 2011-07-24 9:58 PM, Philip Prindeville wrote: On 7/24/11 2:34 AM, Felix Fietkau wrote: On 2011-07-22 7:58 AM, Philip Prindeville wrote: On 7/6/11 1:41 AM, Philip Prindeville wrote: Bump to the current version of PPP. A few of the patches are either in 2.4.5 or something functionally equivalent is, so those patches were removed. 1 patch was added to accommodate pppol2pv3_addr which is referenced in a linux kernel header file, but isn't defined in the kernel header files packaged with PPP. Signed-off-by: Philip Prindevillephil...@redfish-solutions.com Can I please get a code review of this? Here's a few things spotted during review: - You forgot to refresh the patches, there's lots of fuzz/offsets when applying them - The change that removes the exponential backoff during pppoe discovery was dropped, please add it back Looking through the code, the patch was failing to comment that snippet out because it's no longer present. I checked. The 'timeout *= 2' part is still there. - Did you test whether pppd still properly survives errors during pppoe initialization (e.g. when receiving broken data from the AC)? - Felix I only have access to PPPoA and Bridged DSL functionality. I put out a request for people to test the patch but didn't get any responses. You could put in a test change that simulates an occasional connection failure. I think this should be tested because an old patch that was removed touched parts of that and I want to make sure that the update does not cause a regression there. - Felix Can you point it out to me? It must have been buried in a mix with something else... -Philip ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 1/1] ppp: version bump to 2.4.5
On 2011-07-24 10:21 PM, Philip Prindeville wrote: On 7/24/11 1:07 PM, Felix Fietkau wrote: On 2011-07-24 9:58 PM, Philip Prindeville wrote: On 7/24/11 2:34 AM, Felix Fietkau wrote: On 2011-07-22 7:58 AM, Philip Prindeville wrote: On 7/6/11 1:41 AM, Philip Prindeville wrote: Bump to the current version of PPP. A few of the patches are either in 2.4.5 or something functionally equivalent is, so those patches were removed. 1 patch was added to accommodate pppol2pv3_addr which is referenced in a linux kernel header file, but isn't defined in the kernel header files packaged with PPP. Signed-off-by: Philip Prindevillephil...@redfish-solutions.com Can I please get a code review of this? Here's a few things spotted during review: - You forgot to refresh the patches, there's lots of fuzz/offsets when applying them - The change that removes the exponential backoff during pppoe discovery was dropped, please add it back Looking through the code, the patch was failing to comment that snippet out because it's no longer present. I checked. The 'timeout *= 2' part is still there. - Did you test whether pppd still properly survives errors during pppoe initialization (e.g. when receiving broken data from the AC)? - Felix I only have access to PPPoA and Bridged DSL functionality. I put out a request for people to test the patch but didn't get any responses. You could put in a test change that simulates an occasional connection failure. I think this should be tested because an old patch that was removed touched parts of that and I want to make sure that the update does not cause a regression there. - Felix Can you point it out to me? It must have been buried in a mix with something else... I was talking about 350-survive_bad_pads_packets.patch. I believe dropping the patch was the right thing to do, but I still want to make sure that we don't introduce any regressions here. - Felix ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH 1/1] linux-3.0: missing symbols
When building x86, the following symbols turn up missing. Signed-off-by: Philip Prindeville phil...@redfish-solutions.com Index: target/linux/generic/config-3.0 === --- target/linux/generic/config-3.0 (revision 27738) +++ target/linux/generic/config-3.0 (working copy) @@ -174,6 +174,7 @@ # CONFIG_BATTERY_DS2782 is not set # CONFIG_BATTERY_MAX17040 is not set # CONFIG_BATTERY_MAX17042 is not set +# CONFIG_BATTERY_DS2780 is not set # CONFIG_BAYCOM_EPP is not set # CONFIG_BAYCOM_PAR is not set # CONFIG_BAYCOM_SER_FDX is not set @@ -355,6 +356,7 @@ # CONFIG_CFG80211 is not set # CONFIG_CGROUPS is not set # CONFIG_CHARGER_GPIO is not set +# CONFIG_CHARGER_MAX8903 is not set # CONFIG_CHELSIO_T1 is not set # CONFIG_CHELSIO_T3 is not set # CONFIG_CHELSIO_T4 is not set @@ -1280,6 +1282,7 @@ # CONFIG_LIBIPW_DEBUG is not set # CONFIG_LINE6_USB is not set # CONFIG_LIRC_STAGING is not set +# CONFIG_INTEL_MEI is not set # CONFIG_LKDTM is not set CONFIG_LLC=y # CONFIG_LLC2 is not set @@ -1797,6 +1800,7 @@ # CONFIG_OCF_OCFNULL is not set # CONFIG_OCF_SAFE is not set # CONFIG_OCF_TALITOS is not set +# CONFIG_OCF_UBSEC_SSB is not set # CONFIG_OC_ETM is not set # CONFIG_OMFS_FS is not set # CONFIG_ORION_WATCHDOG is not set @@ -2162,6 +2166,7 @@ # CONFIG_SCSI_FC_ATTRS is not set # CONFIG_SCSI_FUTURE_DOMAIN is not set # CONFIG_SCSI_GDTH is not set +# CONFIG_SCSI_ISCI is not set # CONFIG_SCSI_GENERIC_NCR5380 is not set # CONFIG_SCSI_GENERIC_NCR5380_MMIO is not set # CONFIG_SCSI_HPSA is not set @@ -2249,6 +2254,7 @@ # CONFIG_SENSORS_F71805F is not set # CONFIG_SENSORS_F71882FG is not set # CONFIG_SENSORS_F75375S is not set +# CONFIG_SENSORS_FAM15H_POWER is not set # CONFIG_SENSORS_FSCHMD is not set # CONFIG_SENSORS_G760A is not set # CONFIG_SENSORS_GL518SM is not set @@ -3167,3 +3173,4 @@ CONFIG_ZONE_DMA_FLAG=1 # CONFIG_ZRAM is not set # CONFIG_SIGMA is not set +# CONFIG_GOOGLE_FIRMWARE is not set ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [packages] Fix mysql dependencies
Trying to fix some issues with php5 I discoverd that php5's mysql extension doesn't load: snip root@OpenWrt:/tmp# php-cli -m PHP Warning: PHP Startup: Unable to load dynamic library '/usr/lib/php/mysql.so' - File not found in Unknown on line 0 [PHP Modules] Core ctype curl date ereg gettext json openssl pcntl pcre posix Reflection session sockets SPL sqlite3 standard sysvmsg sysvsem sysvshm zlib [Zend Modules] /snap After installing uclibcxx, the extension loads fine: snip root@OpenWrt:/tmp# opkg install uclibcxx Installing uclibcxx (0.2.2-3) to root... Downloading http://192.168.1.2:8081/brcm47xx/packages/uclibcxx_0.2.2-3_brcm47xx.ipk. Configuring uclibcxx. root@OpenWrt:/tmp# php-cli -m [PHP Modules] ... mysql ... [Zend Modules] /snap So the mysqlclient library needs uclibcxx, not only mysql-server. But since mysql-server is depended of the mysqlclient library we can shorten its dependencies. I also cross-checked that e.g. a mysql-enabled lighttpd shows the same behaviour when uclibcxx is not installed. snip root@OpenWrt:/tmp/log/lighttpd# lighttpd -f /etc/lighttpd/lighttpd.conf 1970-01-01 05:07:30: (plugin.c.169) dlopen() failed for: /usr/lib/lighttpd/mod_mysql_vhost.so File not found 1970-01-01 05:07:30: (server.c.650) loading plugins finally failed /snap Signed-off-by: Michael Heimpold m...@heimpold.de --- Index: libs/mysql/Makefile === --- libs/mysql/Makefile (revision 27363) +++ libs/mysql/Makefile (working copy) @@ -9,7 +9,7 @@ PKG_NAME:=mysql PKG_VERSION:=5.1.53 -PKG_RELEASE:=2 +PKG_RELEASE:=3 PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.gz PKG_SOURCE_URL:=\ @@ -29,7 +29,7 @@ define Package/libmysqlclient/Default SECTION:=libs CATEGORY:=Libraries - DEPENDS:=+zlib + DEPENDS:=+uclibcxx +zlib TITLE:=MySQL client library URL:=http://dev.mysql.com/ endef @@ -37,7 +37,7 @@ define Package/mysql-server/Default SECTION:=utils CATEGORY:=Utilities - DEPENDS:=+zlib +libmysqlclient +uclibcxx + DEPENDS:=+libmysqlclient TITLE:=MySQL Server URL:=http://dev.mysql.com/ SUBMENU:=database ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v4 1/1] ocf-linux: version bump to 20110720
On 07/24/2011 10:04 PM, Philip Prindeville wrote: On 7/24/11 7:54 AM, Hauke Mehrtens wrote: On 07/24/2011 12:55 AM, Philip Prindeville wrote: Fixes problem with TFM allocation in cryptosoft.c When checking in, please do: svn rename patches-2.6.39/941-ocf_20105030.patch patches-2.6.39/941-ocf_20110720.patch for each patches directory before committing (but after applying patch, obviously). Redux to include 2.6.30 - 32, as per Hauke's comment. Signed-off-by: Philip Prindeville phil...@redfish-solutions.com Hi Philip, I applied the patch with some changes in r27753. * remove ubsec_ssb package and take it from ocf-linux What was the problem with ubsec_ssb? That is not a problem. ubsec_ssb is some driver for the ipsec accelerator on some bcm47xx SoCs and it was in some package in OpenWrt. Now ocf linux contains this (copied the OpenWrt version) and it is cleaner to just use that version. * use patches from ocf-linux package I had used the patches from ocf-linux... the change I made was to replace linux-2.6.??.orig/ with a/ and linux-2.6.??/ with b/ in the --- and +++ lines... but other than that they should have been identical. * refresh all patches * readd some build fixes for OpenWrt. What build fixes were these? I do not exactly any more I think the patches did not apply for kernel 2.6.37. * readd CRYPTO_MANAGER dependency For cryptosoft? According to the Kconfig that comes with the latest ocf-linux, it no longer has a dependency on crypto-manager, that's why I took it out. This dependency was added in OpenWrt (see: https://dev.openwrt.org/changeset/26402 ). Without it I get the following error when loading the module: cryptosoft: Unknown symbol crypto_alloc_ahash (err 0) cryptosoft: Unknown symbol crypto_ahash_setkey (err 0) cryptosoft: Unknown symbol crypto_alloc_ablkcipher (err 0) cryptosoft: Unknown symbol crypto_ahash_digest (err 0) I also added these 3 things: https://dev.openwrt.org/changeset/21373/ https://dev.openwrt.org/changeset/21849/ https://dev.openwrt.org/changeset/21854/ Hauke ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v4 1/1] ocf-linux: version bump to 20110720
On 7/24/11 2:43 PM, Hauke Mehrtens wrote: [snip] * readd CRYPTO_MANAGER dependency For cryptosoft? According to the Kconfig that comes with the latest ocf-linux, it no longer has a dependency on crypto-manager, that's why I took it out. This dependency was added in OpenWrt (see: https://dev.openwrt.org/changeset/26402 ). Without it I get the following error when loading the module: cryptosoft: Unknown symbol crypto_alloc_ahash (err 0) cryptosoft: Unknown symbol crypto_ahash_setkey (err 0) cryptosoft: Unknown symbol crypto_alloc_ablkcipher (err 0) cryptosoft: Unknown symbol crypto_ahash_digest (err 0) I also added these 3 things: https://dev.openwrt.org/changeset/21373/ https://dev.openwrt.org/changeset/21849/ https://dev.openwrt.org/changeset/21854/ Hauke And these are upstreamed now? -Philip ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 1/1] ppp: version bump to 2.4.5
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi, I took a look at it and it seems that upstream now gracefully handles errors in the PADS processing, on a first glance I see no condition anymore that may lead into a fatal() so dropping this patch is ok. ~ Jow -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk4sncgACgkQdputYINPTPM5DgCfTEX09MmFhDjJIzriSgmisran Da4An1aMGr9bc9DJ7BQ/WTOrQ5FEKlom =IHFO -END PGP SIGNATURE- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH 1/1] alix2: include WAN interface by default
Include WAN interface by default, instead of putting both ports in the LAN bridge. Signed-off-by: Philip Prindeville phil...@redfish-solutions.com Index: target/linux/x86/alix2/base-files/etc/config/network === --- target/linux/x86/alix2/base-files/etc/config/network(revision 27763) +++ target/linux/x86/alix2/base-files/etc/config/network(working copy) @@ -6,9 +6,15 @@ option ipaddr 127.0.0.1 option netmask 255.0.0.0 +# N.B. the labels on the Netgate boxes actually reverse the LAN and WAN +# port markings. config interface lan - option ifname eth0 eth1 + option ifname eth1 option type bridge option protostatic option ipaddr 192.168.1.1 option netmask 255.255.255.0 + +config interface wan + option ifname eth0 + option protodhcp ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH 1/1] linux: bump 2.6.39.x to .3
2.6.39.3 is now latest/stable. Signed-off-by: Philip Prindeville phil...@redfish-solutions.com Index: include/kernel-version.mk === --- include/kernel-version.mk (revision 27761) +++ include/kernel-version.mk (working copy) @@ -26,8 +26,8 @@ ifeq ($(LINUX_VERSION),2.6.38.8) LINUX_KERNEL_MD5SUM:=d27b85795c6bc56b5a38d7d31bf1d724 endif -ifeq ($(LINUX_VERSION),2.6.39.2) - LINUX_KERNEL_MD5SUM:=51be93d92028658ec93f68b79a378b17 +ifeq ($(LINUX_VERSION),2.6.39.3) + LINUX_KERNEL_MD5SUM:=5afede829846587e798f2631c2ece84f endif ifeq ($(LINUX_VERSION),3.0) LINUX_KERNEL_MD5SUM:=398e95866794def22b12dfbc15ce89c0 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel