Re: [OpenWrt-Devel] Packages were found, but none compatible with the architectures configured
> # opkg install kmod-sched > Collected errors: > * Packages were found, but none compatible with the architectures configured I had the same problem with weekend upgrading some packages on same devices running 8.09.1 tag, and packages were compiled using the trunk toolchain. I don't know exactly if this is your case, but for example the Atheros target changed architecture name from "mips" to "atheros_uClibc-0.9.30.1". So if your firmware was compiled before the target name change, and you are now compiling the packages with a recent toolchain, the opkg will fail its check because it thinks you are installing a package for a wrong target. I have no idea if this only applies for the Atheros target or also for other targets. Saverio ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Packages were found, but none compatible with the architectures configured
I recall having similar issues when using a recent trunk compiled image (after the kernel upgrade) and trying to use a kmod package from the 8.09.1 repository ... I guess it's the kernel version. ede On 20.10.2009 12:33, ZioPRoTo (Saverio Proto) wrote: # opkg install kmod-sched Collected errors: * Packages were found, but none compatible with the architectures configured I had the same problem with weekend upgrading some packages on same devices running 8.09.1 tag, and packages were compiled using the trunk toolchain. I don't know exactly if this is your case, but for example the Atheros target changed architecture name from "mips" to "atheros_uClibc-0.9.30.1". So if your firmware was compiled before the target name change, and you are now compiling the packages with a recent toolchain, the opkg will fail its check because it thinks you are installing a package for a wrong target. I have no idea if this only applies for the Atheros target or also for other targets. Saverio ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] bcm63xx_enet: update upstream fixes
Hi all, This patch integrates http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ec1652af18ef02c7c6ceeabb64f56f16eaf40ae9 to 060-bcm63xx_enet_upstream_fixes.patch . Signed-off-by: Guillaume LECERF -- Guillaume LECERF GeeXboX developer - www.geexbox.org ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] bcm63xx_enet: update upstream fixes (with the patch)
With the patch this time. -- Guillaume LECERF GeeXboX developer - www.geexbox.org Index: target/linux/brcm63xx/patches-2.6.30/060-bcm63xx_enet_upstream_fixes.patch === --- target/linux/brcm63xx/patches-2.6.30/060-bcm63xx_enet_upstream_fixes.patch (revision 18097) +++ target/linux/brcm63xx/patches-2.6.30/060-bcm63xx_enet_upstream_fixes.patch (working copy) @@ -8,6 +8,15 @@ #include #include "bcm63xx_enet.h" +@@ -91,7 +90,7 @@ + if (enet_readl(priv, ENET_IR_REG) & ENET_IR_MII) + break; + udelay(1); +- } while (limit-- >= 0); ++ } while (limit-- > 0); + + return (limit < 0) ? 1 : 0; + } @@ -321,7 +320,7 @@ if (len < copybreak) { struct sk_buff *nskb; ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] firewall/iptables
This patch is for firewall/iptables. Most of it is to fix some hush script problems. Adds also some SPI rule to firewall script. diff -ruN package/firewall/files/20-firewall package-after/firewall/files/20-firewall --- package/firewall/files/20-firewall2009-10-06 15:41:25.0 +0300 +++ package-after/firewall/files/20-firewall2009-10-20 13:51:40.0 +0300 @@ -4,8 +4,8 @@ [ "$ifname" == "lo" ] && exit 0 load_zones() { -local name -local network + name= + network= config_get name $1 name config_get network $1 network [ -z "$network" ] && network=$name @@ -20,17 +20,18 @@ [ ifup = "$ACTION" ] && { for z in $ZONE; do -local loaded + loaded= config_get loaded core loaded -[ -n "$loaded" ] && addif "$INTERFACE" "$ifname" "$z" +[ -n "$loaded" ] && [ -n "$z" ] && addif "$INTERFACE" "$ifname" "$z" done } [ ifdown = "$ACTION" ] && { -local up -config_get up "$INTERFACE" up +up= for z in $ZONE; do -[ "$up" == "1" ] && delif "$INTERFACE" "$ifname" "$z" + up= +config_get up $z up +[ "$up" == "1" ] && [ -n "$z" ] && delif "$INTERFACE" "$ifname" "$z" done } diff -ruN package/firewall/files/uci_firewall.sh package-after/firewall/files/uci_firewall.sh --- package/firewall/files/uci_firewall.sh2009-10-06 15:41:25.0 +0300 +++ package-after/firewall/files/uci_firewall.sh2009-10-20 14:37:52.0 +0300 @@ -24,7 +24,7 @@ NOTRACK_DISABLED= find_item() { -local item="$1"; shift +item="$1"; shift for i in "$@"; do [ "$i" = "$item" ] && return 0 done @@ -42,13 +42,14 @@ } create_zone() { -local exists - +exists= + [ "$1" == "loopback" ] && return -config_get exists $ZONE_LIST $1 +eval_ZONE_LIST=$(eval "echo $ZONE_LIST") +config_get exists $eval_ZONE_LIST $1 [ -n "$exists" ] && return -config_set $ZONE_LIST $1 1 +config_set $eval_ZONE_LIST $1 1 $IPTABLES -N zone_$1 $IPTABLES -N zone_$1_MSSFIX @@ -67,11 +68,12 @@ } addif() { -local network="$1" -local ifname="$2" -local zone="$3" +network="$1" +ifname="$2" +zone="$3" -local n_if n_zone +n_if= +n_zone= config_get n_if core "${network}_ifname" config_get n_zone core "${network}_zone" [ -n "$n_zone" ] && { @@ -101,9 +103,9 @@ } delif() { -local network="$1" -local ifname="$2" -local zone="$3" +network="$1" +ifname="$2" +zone="$3" logger "removing $network ($ifname) from firewall zone $zone" $IPTABLES -D input -i "$ifname" -j zone_$zone @@ -123,8 +125,8 @@ } load_synflood() { -local rate=${1:-25} -local burst=${2:-50} +rate=${1:-25} +burst=${2:-50} echo "Loading synflood protection" $IPTABLES -N syn_flood $IPTABLES -A syn_flood -p tcp --syn -m limit --limit $rate/second --limit-burst $burst -j RETURN @@ -133,8 +135,8 @@ } fw_set_chain_policy() { -local chain=$1 -local target=$2 +chain=$1 +target=$2 [ "$target" == "REJECT" ] && { $IPTABLES -A $chain -j reject target=DROP @@ -220,9 +222,9 @@ } fw_zone() { -local name -local network -local masq +name= +network= +masq= config_get name $1 name config_get network $1 network @@ -238,18 +240,18 @@ } fw_rule() { -local src -local src_ip -local src_mac -local src_port -local src_mac -local dest -local dest_ip -local dest_port -local proto -local icmp_type -local target -local ruleset +src= +src_ip= +src_mac= +src_port= +src_mac= +dest= +dest_ip= +dest_port= +proto= +icmp_type= +target= +ruleset= config_get src $1 src config_get src_ip $1 src_ip @@ -265,14 +267,18 @@ src_port_first=${src_port%-*} src_port_last=${src_port#*-} -[ "$src_port_first" -ne "$src_port_last" ] && { \ -src_port="$src_port_first:$src_port_last"; } +if [ -n "$src_port_last" ] ; then +[ "$src_port_first" -ne "$src_port_last" ] && { +src_port="$src_port_first:$src_port_last"; } +fi dest_port_first=${dest_port%-*} dest_port_last=${dest_port#*-} -[ "$dest_port_first" -ne "$dest_port_last" ] && { \ -dest_port="$dest_port_first:$dest_port_last"; } - +if [ -n "$dest_port_last" ] ; then +[ "$dest_port_first" -ne "$dest_port_last" ] && { +dest_port="$dest_port_first:$dest_port_last"; } +fi + ZONE=input TARGET=$target [ -z "$target" ] && target=DROP @@ -280,15 +286,14 @@ [ -n "$src" -a -n "$dest" ] && ZONE=zone_${src}_forward [ -n "$dest" ] && TARGET=zone_${dest}_$target add_rule() { -$IPTABLES -I $ZONE 1 \ -${proto:+-p $proto} \ -${icmp_type:+--icmp-type $icmp_type} \ -${src_ip:+-s $src_ip} \ -${src_port:+--sport $src_port} \ -${src_mac:+-m mac --ma
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
On Tuesday 20 October 2009 14:58:49 Ugur DOGRU wrote: > This patch is for firewall/iptables. Most of it is to fix some hush > script problems. >[...] As most other init/hotplug scripts don't work with hush as well, I wonder why you went for the firewall only :) What you "fixed" are actually not problems in the scripts, but workarounds for hush-limitations and un-localizing variables is prone to breaking some stuff (I fixed one such issue earlier this year). The eval-code you used for assigning variables looks fishy as well but I never worked with hush so maybe PITAresque stuff like this is needed. Maybe you should just add the 50k and use ash instead? Cheers, Malte -- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi Malte. > [...] > Maybe you should just add the 50k and use ash instead? As far as I know is "hush" a requirement of the platform (no mmu) and I think "ash" does not work on non-mmu + non-fork platforms. ~ JoW -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkrd2scACgkQdputYINPTPN//gCeJk490y55LpSJKQP++wghzZ6K Tv4AniggWzPOQT1RSAdy314bwXyP1TqJ =W339 -END PGP SIGNATURE- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi. What about "emulating" local with something like the following? local() { eval "$1=''"; } This way you can keep the current scripts and handle the not-implemented local gracefully. ~ JoW -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkrd29IACgkQdputYINPTPND4ACaArek5LSOYNkoNV1FFpmK81lb GzEAn1f9SPvnwPucGRp/Y2PT58WItq7C =rKsF -END PGP SIGNATURE- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
Hi, On Tuesday 20 October 2009 17:48:34 Jo-Philipp Wich wrote: > Hi. > > What about "emulating" local with something like the following? > > local() { eval "$1=''"; } > > This way you can keep the current scripts and handle the not-implemented > local gracefully. I definitively prefer that solution, anytime hush makes more progress towards having local variables support or ash finally works on no-mmu systems this "hack" can be removed. ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The following function supports multiple args like the original "local". if ! type "local" >/dev/null; then local() { for _v in $*; do eval "$_v=''"; done } fi If you add that to /etc/profile (is that supported by hush?) then it should be available system wide, in any script. ~ JoW -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkrd5AoACgkQdputYINPTPP2RwCgjlMDgHn/slP2BbQ+488LpzwN GOIAnik/wOoQCFG7p9m7NXbKd3SZdIIs =eQjM -END PGP SIGNATURE- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
Jo-Philipp Wich wrote: > The following function supports multiple args like the original "local". > > if ! type "local" >/dev/null; then > local() { > for _v in $*; do eval "$_v=''"; done > } > fi > > > If you add that to /etc/profile (is that supported by hush?) then it > should be available system wide, in any script. /etc/functions.sh would be a better place, imho. I think /etc/profile is not automatically sourced by shell scripts. This function doesn't currently cover all scripts, it needs to handle things like local var="value" as well without inserting extra = characters. - Felix ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
hi Malte, I forgot to say that, this is for ubicom32 cpu that has no mmu, and other no-mmu cpu's, there is no memory concern. this will not be the end of hush scripts changes, others are on the way. "eval" may be fishy for ash, i don't know. May be making a global "switch" mechanism between ash/hush is the solution to "isolate" different behaviour of shells. regards ugur -Original Message- From: openwrt-devel-boun...@lists.openwrt.org on behalf of Malte S. Stretz Sent: Tue 20.10.2009 16:51 To: OpenWrt Development List Subject: Re: [OpenWrt-Devel] [PATCH] firewall/iptables On Tuesday 20 October 2009 14:58:49 Ugur DOGRU wrote: > This patch is for firewall/iptables. Most of it is to fix some hush > script problems. >[...] As most other init/hotplug scripts don't work with hush as well, I wonder why you went for the firewall only :) What you "fixed" are actually not problems in the scripts, but workarounds for hush-limitations and un-localizing variables is prone to breaking some stuff (I fixed one such issue earlier this year). The eval-code you used for assigning variables looks fishy as well but I never worked with hush so maybe PITAresque stuff like this is needed. Maybe you should just add the 50k and use ash instead? Cheers, Malte -- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
Good idea. but, not sure that "type" supported by hush. ugur Jo-Philipp Wich wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The following function supports multiple args like the original "local". if ! type "local" >/dev/null; then local() { for _v in $*; do eval "$_v=''"; done } fi If you add that to /etc/profile (is that supported by hush?) then it should be available system wide, in any script. ~ JoW -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkrd5AoACgkQdputYINPTPP2RwCgjlMDgHn/slP2BbQ+488LpzwN GOIAnik/wOoQCFG7p9m7NXbKd3SZdIIs =eQjM -END PGP SIGNATURE- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] firewall/iptables
Latest busybox version is 1.15.2 and includes local builtin implemented. You can check the following link: http://busybox.net/. As noted, in the present time, hush is a must for nommu processors like we have: ubicom32. Bayram -Original Message- From: openwrt-devel-boun...@lists.openwrt.org on behalf of Florian Fainelli Sent: Tue 10/20/2009 09:03 To: openwrt-devel@lists.openwrt.org Subject: Re: [OpenWrt-Devel] [PATCH] firewall/iptables Hi, On Tuesday 20 October 2009 17:48:34 Jo-Philipp Wich wrote: > Hi. > > What about "emulating" local with something like the following? > > local() { eval "$1=''"; } > > This way you can keep the current scripts and handle the not-implemented > local gracefully. I definitively prefer that solution, anytime hush makes more progress towards having local variables support or ash finally works on no-mmu systems this "hack" can be removed. ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel