[OpenWrt-Devel] Curious about procd
Hi, the FeM e.V. is a non-profit organization that for example operates a student computer network in Germany. We therefore currently have about 20 APs (P1020WLAN) deployed (more to come), running an OpenWRT-based firmware in a centralized manner. So the flashed OpenWRT will just download another OpenWRT-firmware (kernel+rootfs) and run it in memory using kexec. As we need very recent kernel support, we opted to base our firmware on OpenWRT-Trunk and after some freezing and testing use it on the campus. The firmware is build by adding a custom feed, a new kernel target and a specific .config - so it was easy to upgrade and maintain. Using a very recent kernel and compat-wireless doesn't only help us but also enables some testing and fixing effort that results in fixes send to upstream, as well as adding and testing new features to existing or new projects (for example hostapd or dhcpsnoopingd). I recently rebased our devel tree to openwrt trunk and found that the devices would no longer boot up. This is because we use /sbin/init as rdinit binary (the image is started using kexec + initrd with) and the package base-files still includes /etc/inittab, which references /etc/init.d/rcS that is lacking. It was removed in a commit saying that it is not needed with procd. The preinit script also still references /sbin/init. Changing the kernel command line to start /sbin/procd helped in that some init scripts were run, but networking was left completely unconfigured (we only used the stock /etc/init.d/network) and procd didn't bother to start a login prompt on the console. So I started searching for information about procd, why it is there and how to use it. But I found hardly anything. Procd is mentioned in the wiki, but not documented. It looks like there was no public discussion about using procd (on a mailing list) nor did anybody bother to sum up the reasons why procd was chosen or informations about how to use it into the wiki. As procd is OpenWRT specific, there is no documentation elsewhere and its source repo doesn't even contain a readme file. I also fail see any obvious reasons why it should be required - given that cron, inetd and init were already there. Script files are also more easy to patch from files added by a feed. So the OpenWRT trunk currently doesn't work for me with neither procd nor init. As it looks like procd is becoming the future, I go against reverting the procd-related changes. But as it is also completely undocumented, I find it difficult to fix the issue I have. This is really disappointing to me. I'm also asking myself why I don't find a discussion about procd - I suppose that the topic is likely to have been discussed among developers. So I would be glad if somebody could point me to documentation about procd and maybe why it is there. Thanks alot, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Curious about procd
Hi, thanks a lot for the quick reply. It helped me tracking down the issue as learned that I needed to look for inittab code in procd and that the code is expected to work out of the box. First my platform seems to be the only target (mpc85xx) that has no console entry in inittab by default in OpenWRT now. This wasn't an issue in the past, as an init.d-Script was fiddling with inittab anyway (to use /bin/login instead of ash) and made init reload it, but it looks like procd cannot handle the HUP signal. Moving this into preinit now avoids the need of the HUP signal. Next, I was facing issues when starting procd with rdinit=/sbin/procd. The errors are "coldplug failed" and "/proc/kmsg does not exist" (or alike). The console then does not get started either as /dev/ttyS0 does not exist. This can be fixed by running /sbin/procd with basename init. I noticed the need for a different basename as the preinit output from procd was missing, which indicates a different code path in main.c from procd. So I now use rdinit=/bin/feminit. The executable file /bin/feminit contains: #!/bin/sh +e ln -s /sbin/procd /init exec /init It cannot be fixed by doing the mount before procd starts, as procd will unmount it. For documentation: procd inittab needs sysinit and shutdown entries to run the rc scripts. Its script file names are ignored, but their arguments are not. ::sysinit:/etc/init.d/rcS S boot ::shutdown:/etc/init.d/rcS K shutdown This was really confusing to me. Maybe /etc/init.d/rcS should be replaced by something like procd:rcS or alike just to avoid confusion? Regards, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] linux-wireless WARNING in ieee80211_sta_ps_deliver_wakeup
Hi, I'm running openwrt Trunk Rev 38850 on Freescale P1020WLAN (mpc85xx) and repeatingly get the following kernel warning during boot. This is 2x ath9k (AR93xx) on a ppc dualcore with each phy spanning multiple BSS with multiple dynamic VLANs each. Regards, M. Braun [ 163.749372] [ cut here ] [ 163.754000] WARNING: at /opt/openwrt/stage2/trunk/build_dir/target-powerpc_8540_eglibc-2.17/linux-mpc85xx_p1020wlan-fem/compat-wireless-2013-11-05/net/mac80211/util.c:2375 [ 163.769287] Modules linked in: ath9k ath9k_htc ath9k_common iptable_nat carl9170 ath9k_hw ath nf_nat_ipv4 nf_conntrack_ipv4 mac80211 ipt_MASQUERADE ebtable_nat ebtable_filter ebtable_broute cfg80211 xt_time xt_tcpudp xt_state xt_nat xt_multiport xt_mark xt_mac xt_limit xt_conntrack xt_comment xt_TCPMSS xt_REDIRECT xt_LOG xt_CT nf_nat nf_defrag_ipv4 nf_conntrack macvlan iptable_raw iptable_mangle iptable_filter ipt_REJECT ip_tables ebtables ebt_vlan ebt_ulog ebt_stp ebt_snat ebt_redirect ebt_pkttype ebt_nflog ebt_mark_m ebt_mark ebt_log ebt_limit ebt_ip6 ebt_ip ebt_dnat ebt_arpreply ebt_arp ebt_among ebt_802_3 crc8 crc7 crc_itu_t crc_ccitt compat sch_teql em_nbyte cls_basic sch_dsmark em_meta sch_gred sch_htb act_ipt sch_tbf em_text sch_codel sch_sfq act_police sch_prio em_cmp sch_red act_skbedit act_mirred em_u32 cls_u32 cls_tcindex cls_flow cls_route cls_fw sch_hfsc sch_ingress ledtrig_morse ledtrig_heartbeat ledtrig_gpio booke_wdt xt_set x_tables ip_set_list_set ip_set_hash_netport ip_set_hash_netiface ip_set_hash_net ip_set_hash_ipportnet ip_set_hash_ipportip ip_set_hash_ipport ip_set_hash_ip ip_set_bitmap_port ip_set_bitmap_ipmac ip_set_bitmap_ip ip_set msdos e1000 dummy tun vfat fat minix nls_utf8 cryptodev ipv6 bridge stp chainiv eseqiv psnap p8022 llc xts gf128mul crypto_wq algif_skcipher algif_hash af_alg sha1_generic krng rng crypto_null michael_mic md5 md4 talitos hmac ecb des_generic deflate cbc authenc aead arc4 crypto_blkcipher mmc_block gpio_keys_polled usb_storage mmc_core input_polldev leds_gpio sd_mod scsi_mod ext4 jbd2 mbcache btrfs zlib_deflate xor i2c_mpc raid6_pq lzo_decompress lzo_compress libcrc32c button_hotplug crc32c crypto_hash [ 163.916584] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.10.18 #9 [ 163.922585] task: c04162e0 ti: ee7e8000 task.ti: c0444000 [ 163.927977] NIP: f3ba7ad4 LR: f3b8131c CTR: c02f3bd8 [ 163.932935] REGS: ee7e9c20 TRAP: 0700 Not tainted (3.10.18) [ 163.938760] MSR: 00029000 CR: 44048042 XER: [ 163.944872] [ 163.944872] GPR00: f3b81310 ee7e9cd0 c04162e0 0003 f3b81310 [ 163.944872] GPR08: efa74cac 0001 f3bd1d7a c02f3bd8 10034aa0 ee1e6030 ee9c64be [ 163.944872] GPR16: efb69264 0004 efb69000 efae6570 ee9c6480 f3ba eebf4500 efa74c20 [ 163.944872] GPR24: ee7e9cd8 efb6911c efb69000 0004 efb691ec [ 163.974692] NIP [f3ba7ad4] ieee80211_smps_is_restrictive+0x60/0x94 [mac80211] [ 163.981830] LR [f3b8131c] ieee80211_sta_ps_deliver_wakeup+0x1f4/0x54c [mac80211] [ 163.989217] Call Trace: [ 163.991667] [ee7e9cd0] [f3b81310] ieee80211_sta_ps_deliver_wakeup+0x1e8/0x54c [mac80211] (unreliable) [ 164.000899] [ee7e9d30] [f3b9a144] ieee80211_rx_irqsafe+0xc34/0x3958 [mac80211] [ 164.008128] [ee7e9dd0] [f3b9cdb4] ieee80211_rx_irqsafe+0x38a4/0x3958 [mac80211] [ 164.015443] [ee7e9e30] [f3b9d4e8] mac80211_ieee80211_rx+0x680/0x848 [mac80211] [ 164.022669] [ee7e9e90] [f3d8b038] ath_rx_tasklet+0xe98/0xfec [ath9k] [ 164.029024] [ee7e9f40] [f3d88068] ath9k_tasklet+0x134/0x340 [ath9k] [ 164.035300] [ee7e9f70] [c002bbf0] tasklet_action+0xec/0x188 [ 164.040872] [ee7e9fa0] [c002bd80] __do_softirq+0xf4/0x1c8 [ 164.046270] [ee7e9ff0] [c000badc] call_do_softirq+0x14/0x24 [ 164.051845] [c0445e60] [c00045a0] do_softirq+0x88/0xec [ 164.056982] [c0445e80] [c002bfd4] irq_exit+0x74/0x8c [ 164.061945] [c0445e90] [c000436c] do_IRQ+0x12c/0x154 [ 164.066910] [c0445ec0] [c000d6bc] ret_from_except+0x0/0x18 [ 164.072398] --- Exception: 501 at arch_cpu_idle+0x24/0x60 [ 164.072398] LR = arch_cpu_idle+0x24/0x60 [ 164.082057] [c0445f80] [c007e928] rcu_idle_enter+0xbc/0xf8 (unreliable) [ 164.088678] [c0445f90] [c005bc08] cpu_startup_entry+0xf8/0x158 [ 164.094515] [c0445fc0] [c03e57c0] start_kernel+0x300/0x314 [ 164.07] [c0445ff0] [c3f8] skpinv+0x2e4/0x320 [ 164.104955] Instruction dump: [ 164.107920] 68830001 7c630034 5463d97e 68630001 4e800020 0fe0 3860 4e800020 [ 164.115694] 3d40f3bd 394a1d7a 892a000a 69290001 <0f09> 2f89 40beffac 3860 [ 164.123641] ---[ end trace 67775a34f53bbbf9 ]--- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] fix dependency for glib2
Index: libs/glib2/Makefile === --- libs/glib2/Makefile (Revision 32264) +++ libs/glib2/Makefile (Arbeitskopie) @@ -30,7 +30,7 @@ define Package/glib2 SECTION:=libs CATEGORY:=Libraries - DEPENDS:=$(ICONV_DEPENDS) $(INTL_DEPENDS) + DEPENDS:=$(ICONV_DEPENDS) $(INTL_DEPENDS) +libpthread +zlib TITLE:=glib 2.0 URL:=http://www.gtk.org/ endef ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] fix dependency for mtd-utils
Index: utils/mtd-utils/Makefile === --- utils/mtd-utils/Makefile(Revision 32264) +++ utils/mtd-utils/Makefile(Arbeitskopie) @@ -37,8 +37,8 @@ ftl_check \ ftl_format \ jffs2dump \ - mkfs.jffs2 \ - mkfs.ubifs:+libuuid \ + mkfs.jffs2:+zlib \ + mkfs.ubifs:+libuuid:+zlib \ mtd_debug \ mtdinfo \ nanddump \ ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] fix dependency for pyopenssl
Index: lang/pyopenssl/Makefile === --- lang/pyopenssl/Makefile (Revision 32264) +++ lang/pyopenssl/Makefile (Arbeitskopie) @@ -27,7 +27,7 @@ CATEGORY:=Languages TITLE:=pyopenssl URL:=http://pyopenssl.sourceforge.net/ - DEPENDS:=+python + DEPENDS:=+python +libopenssl endef define Package/pyopenssl/description ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] fix dependency for apr
Index: libs/apr/Makefile === --- libs/apr/Makefile (Revision 32264) +++ libs/apr/Makefile (Arbeitskopie) @@ -25,7 +25,7 @@ define Package/libapr SECTION:=libs CATEGORY:=Libraries - DEPENDS:=+libpthread +librt + DEPENDS:=+libpthread +librt +libuuid TITLE:=Apache Portable Runtime Library URL:=http://apr.apache.org/ endef ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] wrong permissions in cpio image
Hi, the tgz rootfs image forces all files to be owned by root:root, but this is not the case with the cpio initramfs image. This needs "-R 0:0" to be added as a parameter to cpio in include / image.mk line around 111. Regards, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] Use DEVICE+VLAN naming scheme in hostapd
Hi,
OpenWRT switched to the Interface+VlanID naming scheme for vlan devices
recently and hostapd only supports the "vlan"+VID scheme. This sometimes
leads to race conditions, when either system expects the other mode of
operation. Although the best solution were to get the kernel return the
name of the created interface (or let the user configure it), that would
require more work. Further I prefer the dev+vid scheme over the
"vlan"+VID scheme, as this avoids conflicts were the same vlan-id is
added to more than one device, so I propose to change hostapd instead of
openwrt. If this cannot go upstream, maybe it would still be a good
thing to have in openwrt, hence the CC to OpenWRT.
I therefore propose to change hostapd to use dev+vid naming scheme and
am attaching a patch.
Hostap-Signed-Off: M. Braun --- a/src/ap/vlan_init.c 2012-06-25 15:20:00.934769046 +0200
+++ b/src/ap/vlan_init.c 2012-06-25 15:31:13.881415772 +0200
@@ -403,8 +403,8 @@
/* Determine if a suitable vlan device already exists. */
- os_snprintf(if_request.device1, sizeof(if_request.device1), "vlan%d",
- vid);
+ os_snprintf(if_request.device1, sizeof(if_request.device1), "%s.%d",
+ if_name, vid);
if_request.cmd = _GET_VLAN_VID_CMD;
@@ -501,7 +501,7 @@
vlan->clean |= DVLAN_CLEAN_VLAN;
os_snprintf(vlan_ifname, sizeof(vlan_ifname),
- "vlan%d", vlan->vlan_id);
+ "%s.%d", tagged_interface, vlan->vlan_id);
if (!br_addif(br_name, vlan_ifname))
vlan->clean |= DVLAN_CLEAN_VLAN_PORT;
@@ -542,7 +542,7 @@
if (tagged_interface) {
os_snprintf(vlan_ifname, sizeof(vlan_ifname),
- "vlan%d", vlan->vlan_id);
+ "%s.%d", tagged_interface, vlan->vlan_id);
if (vlan->clean & DVLAN_CLEAN_VLAN_PORT)
br_delif(br_name, vlan_ifname);
ifconfig_down(vlan_ifname);
@@ -682,7 +682,7 @@
if (priv == NULL)
return NULL;
- vlan_set_name_type(VLAN_NAME_TYPE_PLUS_VID_NO_PAD);
+ vlan_set_name_type(VLAN_NAME_TYPE_RAW_PLUS_VID_NO_PAD);
priv->s = socket(PF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
if (priv->s < 0) {
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] fix wget-ssl dependencies
Index: net/wget/Makefile === --- net/wget/Makefile (Revision 34280) +++ net/wget/Makefile (Arbeitskopie) @@ -22,7 +22,7 @@ define Package/wget/Default SECTION:=net CATEGORY:=Network - DEPENDS:=+libpcre + DEPENDS:=+libpcre +libuuid SUBMENU:=File Transfer TITLE:=Non-interactive network downloader URL:=http://www.gnu.org/software/wget/index.html ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] CrashLog broken on P1020WLAN (PPC)
Hi, CrashLog breaks booting the kernel for me on P1020 WLAN (PPC). Kernel output: Crashlog failed to allocate RAM at address 0x2ff0 [...] __ioremap(): phys addr 0x2fffc000 is RAM lr c0362eac Unable to handle kernel paging request for data at address 0x Faulting instruction address: 0xc0362eb0 Oops: Kernel access of bad area, sig: 11 [#1] SMP NR_CPUS=2 P1020 RDB Modules linked in: NIP: c0362eb0 LR: c0362eac CTR: c000f1fc REGS: ee841ef0 TRAP: 0300 Not tainted (3.3.8) MSR: 00029000 CR: 20004022 XER: DEAR: , ESR: TASK = ee83c000[1] 'swapper/1' THREAD: ee84 CPU: 1 GPR00: a1ee ee841fa0 ee83c000 00021000 ee841ea6 0002 GPR08: c0388f94 c0388fe4 c03ba4f0 0001 40004022 10034ad0 3ff83a00 003f2680 GPR16: c900 0012 0fff 017fd000 0015 10011ba8 c03a6000 GPR24: 0001 16c0 c03c c0362e78 c03e c03e429c NIP [c0362eb0] crashlog_init_fs+0x38/0xbc LR [c0362eac] crashlog_init_fs+0x34/0xbc Call Trace: [ee841fa0] [c0362eac] crashlog_init_fs+0x34/0xbc (unreliable) [ee841fb0] [c0001e2c] do_one_initcall+0xdc/0x1a8 [ee841fe0] [c0355844] kernel_init+0xb4/0x13c [ee841ff0] [c000ce90] kernel_thread+0x4c/0x68 Instruction dump: 3860fff4 bfc10008 3fc0c03e 90010014 3bfe429c 801f0004 2f80 41be0084 7c030378 38804000 4bcb0071 3c00a1ee <8123> 6000dead 907e429c 7f89 ---[ end trace 361d25c4ae36ea70 ]--- Kernel panic - not syncing: Attempted to kill init! Rebooting in 180 seconds.. Looks like https://dev.openwrt.org/changeset/34213/ for SPARC, so maybe disable CRASHLOG for PPC, too? Regards, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] Freescale P1020 WLAN support
Hi, is there currently anybody working on OpenWRT P1020-WLAN (Freescale) support? It is because we at the FeM e.V. are working on it trying to port the freescale patches forward and would like to join forces with anybody already working on it. Regards, M. Braun -- http://www.fem.tu-ilmenau.de/ ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [Projekt-wlan] [PATCH 0/7] mpc85xx: P1010RDB and P1020WLAN support
Hi, thanks for all the patches. If tested them on P1020 WLAN Rev D and found USB failing if, - the usb chipset at 23000 stays in dts or [lockup during boot] - usb is not loaded statically [won't recognize the usb bus] Further, booting fails if the free_space mapping is in dts [lockup during boot], the freescale kernel has a slightly different dtsi file (no nand, smaller nor) so I changed that (works fine). The RTC DS1307 needs RTC_CLASS and I2C to be set, so I added that. Further, I enabled and testet kexec - works fine. Additionally, I copied parts of the freescale linux kernel config over, that looked like they mit benefit without interfering. This enables DMA, for example. Already explained CONFIG_HIGHMEM. I did not test the MMC interface. Please find attached my modifications on top of your patch set. Regards, M. Braun From d11c53d51de838d4a589c847676dc4348b701940 Mon Sep 17 00:00:00 2001 From: M. Braun Date: Sat, 31 Mar 2012 05:48:24 +0200 Subject: [PATCH] Fixups for P1020WLAN Working on P1020WLAN support. * get full memory working using CONFIG_HIGHMEM * get usb host support working by removing second usb entry from dtsi file (freescale takes same approach in the dev kit) Further, the ehci_hcd module must be statically linked into the kernel, otherwise the chipset won't be detected. * enable kexec * add some more freescale default options as they did test them with 2.6.32 This makes the rtc appear, for example. --- package/kexec-tools/Makefile |2 +- target/linux/mpc85xx/config-default| 86 +- .../linux/mpc85xx/patches/230-p1020wlan-dts.patch | 33 ++- .../patches/240-p1020wlan-dts-fix-usb-dts.patch| 93 target/linux/mpc85xx/profiles/201-p1020wlan.mk |2 +- 5 files changed, 205 insertions(+), 11 deletions(-) create mode 100644 target/linux/mpc85xx/patches/240-p1020wlan-dts-fix-usb-dts.patch diff --git a/package/kexec-tools/Makefile b/package/kexec-tools/Makefile index 87ad808..5ae3af1 100644 --- a/package/kexec-tools/Makefile +++ b/package/kexec-tools/Makefile @@ -22,7 +22,7 @@ include $(INCLUDE_DIR)/package.mk define Package/kexec-tools SECTION:=utils CATEGORY:=Utilities - DEPENDS:=@armeb||@arm||@i386||@powerpc64||@mipsel||@mips +zlib + DEPENDS:=@armeb||@arm||@i386||@powerpc64||@mipsel||@mips||@TARGET_mpc85xx +zlib TITLE:=Kernel boots kernel URL:=http://kernel.org/pub/linux/kernel/people/horms/kexec-tools/ MENU:=1 diff --git a/target/linux/mpc85xx/config-default b/target/linux/mpc85xx/config-default index 8f3795c..b0668ac 100644 --- a/target/linux/mpc85xx/config-default +++ b/target/linux/mpc85xx/config-default @@ -117,7 +117,7 @@ CONFIG_LXT_PHY=y CONFIG_MAX_ACTIVE_REGIONS=32 CONFIG_MDIO_BOARDINFO=y CONFIG_MIGRATION=y -# CONFIG_MII is not set +CONFIG_MII=y # CONFIG_MLX4_CORE is not set # CONFIG_MMIO_NVRAM is not set # CONFIG_MPC8536_DS is not set @@ -133,11 +133,15 @@ CONFIG_MPIC=y CONFIG_MTD_NAND=y CONFIG_MTD_NAND_ECC=y CONFIG_MTD_NAND_FSL_ELBC=y -# CONFIG_MTD_OF_PARTS is not set +CONFIG_MTD_OF_PARTS=y CONFIG_MTD_PHYSMAP=y CONFIG_MTD_PHYSMAP_OF=y # CONFIG_MTD_SM_COMMON is not set # CONFIG_MTD_WRSBC8260 is not set +CONFIG_MTD_CMDLINE_PARTS=y +CONFIG_MTD_NAND_ECC_SMC=y +CONFIG_MTD_NAND_PLATFORM=y +CONFIG_MTD_NAND_FSL_UPM=y CONFIG_MUTEX_SPIN_ON_OWNER=y # CONFIG_NEED_DMA_MAP_STATE is not set # CONFIG_NEED_PER_CPU_EMBED_FIRST_CHUNK is not set @@ -230,11 +234,13 @@ CONFIG_RWSEM_XCHGADD_ALGORITHM=y CONFIG_SBC8560=y CONFIG_SCHED_HRTICK=y # CONFIG_SCSI_DMA is not set -# CONFIG_SERIAL_8250_DETECT_IRQ is not set +CONFIG_SERIAL_8250=y +CONFIG_SERIAL_8250_PCI=y +CONFIG_SERIAL_8250_DETECT_IRQ=y CONFIG_SERIAL_8250_EXTENDED=y CONFIG_SERIAL_8250_FSL=y -# CONFIG_SERIAL_8250_MANY_PORTS is not set -# CONFIG_SERIAL_8250_RSA is not set +CONFIG_SERIAL_8250_MANY_PORTS=y +CONFIG_SERIAL_8250_RSA=y CONFIG_SERIAL_8250_SHARE_IRQ=y CONFIG_SERIAL_OF_PLATFORM=y # CONFIG_SERIAL_QE is not set @@ -260,6 +266,7 @@ CONFIG_UCC=y CONFIG_UCC_FAST=y CONFIG_UCC_GETH=y # CONFIG_UGETH_TX_ON_DEMAND is not set +CONFIG_USB=y CONFIG_USB_ARCH_HAS_XHCI=y CONFIG_USB_SUPPORT=y CONFIG_USE_GENERIC_SMP_HELPERS=y @@ -268,3 +275,72 @@ CONFIG_VM_EVENT_COUNTERS=y CONFIG_WORD_SIZE=32 # CONFIG_XES_MPC85xx is not set CONFIG_XPS=y + +CONFIG_POSIX_MQUEUE=y +CONFIG_POSIX_MQUEUE_SYSCTL=y +CONFIG_SYSCTL_SYSCALL=y +CONFIG_MODULE_FORCE_LOAD=y +CONFIG_MODULE_FORCE_UNLOAD=y +CONFIG_HIGHMEM=y +CONFIG_KEXEC=y +CONFIG_SECCOMP=y +CONFIG_WIRELESS_EXT_SYSFS=y +CONFIG_FTL=y +CONFIG_OF_SPI=y +CONFIG_NET_VENDOR_MICROCHIP=y +CONFIG_SERIO=y +CONFIG_SERIO_I8042=y +CONFIG_SERIO_SERPORT=y +CONFIG_SERIO_LIBPS2=y +CONFIG_VT=y +CONFIG_CONSOLE_TRANSLATIONS=y +CONFIG_VT_CONSOLE=y +CONFIG_HW_CONSOLE=y +CONFIG_HW_RANDOM=y +CONFIG_NVRAM=y +CONFIG_I2C=y +CONFIG_I2C_COMPAT=y +CONFIG_I2C_HELPER_AUTO=y +CONFIG_SPI_FSL_LIB=y +CONFIG_SPI_FSL_SPI=y +CONFIG_SPI_FSL_ESPI=y +CONFIG_DUMMY_CONSOLE=y +CONFIG_USB_ANNO
[OpenWrt-Devel] initramfs support
Hi, enabling the cpio target (e.g. initramfs not embedded within the elf image, required for example by p1020wlan tftp boot due to u-boot memory constraints), the archive is lacking /init . This results in the kernel panicking because rootfs cannot be mounted. However, I see /init in the staging dir. As a workaround, I now repack the cpio.tgz but this is not intended, or is it? Regards, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] linux 3.3 hostapd issue between 30952 - 31118
Hi, I've also found this on P1020wlan. It looks like it is caused by libnl-tiny (maybe in combination with linux 3.3), as it issues syscalls the let the kernel reject with EALREADYINUSE and debug prints "cannot setup genlink support (mgmt)" (once debug verbosity is increased in .config). Can be fixed by upgrading libnl to 3.2.7 (most recent) and editing hostapd to use libnl instead of libnl-tiny. http://git.fem.tu-ilmenau.de/?p=fem-wlan.git;a=commit;h=57747d0656e73f67adb8e0d501e381e33850be89 (you only need the changes to the package directory). Regards, M. Braun Am 30.03.2012 06:24, schrieb Dave Taht: On Thu, Mar 29, 2012 at 8:48 PM, Dave Taht wrote: On Thu, Mar 29, 2012 at 8:27 PM, Dave Taht wrote: On Tue, Mar 27, 2012 at 9:52 PM, Otto Solares Cabrera wrote: On Tue, Mar 27, 2012 at 05:03:24PM -0700, Dave Taht wrote: > however my issue is that hostapd isn't coming up for ap mode. Yeah in 3.3 the mon.wlan0 i/f is not being setup but I didn't investigate further. Your confirmation rules out cerowrt's setup (with device renaming, etc), so that reduces it to hostapd, nl, wireless-next, or the 3.3 kernel, sometime in the last 150 commits. sigh. I confirm there's no mon device being created. Go looking for a newer hostapd? bisect? root@OpenWrt:/lib/wifi# wifi start Configuration file: /var/run/hostapd-phy0.conf nl80211: Failed to set interface sw00 into AP mode nl80211 driver initialization failed. Failed to start hostapd for phy0 Configuration file: /var/run/hostapd-phy1.conf nl80211: Failed to set interface sw10 into AP mode nl80211 driver initialization failed. Failed to start hostapd for phy1 The only other thing I see going wrong is this: [17308.449218] gw01: Trigger new scan to find an IBSS to join [17309.289062] ADDRCONF(NETDEV_CHANGE): gw01: link becomes ready [17315.261718 [5]] gw11: Trigger new scan to find an IBSS to join [17316.726562 [6]] ADDRCONF(NETDEV_CHANGE): gw11: link becomes ready [17382.464843] ath: Failed to stop TX DMA, queues=0x004! I just managed to rather quickly build iw-3.4 and am about to go looking at hostapd... :me flails: -- Dave Täht SKYPE: davetaht US Tel: 1-239-829-5608 [2] http://www.bufferbloat.net [3] -- Dave Täht SKYPE: davetaht US Tel: 1-239-829-5608 [7] http://www.bufferbloat.net [8] Another symptom is that adhoc interfaces are created but do not get ssids assigned. I can assign them manually, however... -- Dave Täht SKYPE: davetaht US Tel: 1-239-829-5608 http://www.bufferbloat.net [10] Links: -- [1] mailto:so...@guug.org [2] http://webmail.fami-braun.de/tel:1-239-829-5608 [3] http://www.bufferbloat.net [4] mailto:dave.t...@gmail.com [5] http://webmail.fami-braun.de/tel:%5B17315.261718 [6] http://webmail.fami-braun.de/tel:%5B17316.726562 [7] http://webmail.fami-braun.de/tel:1-239-829-5608 [8] http://www.bufferbloat.net [9] mailto:dave.t...@gmail.com [10] http://www.bufferbloat.net ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] initramfs support
Hi, Am 06.04.2012 10:16, schrieb Florian Fainelli: You are supposed to set the kernel command-line to contain init=/etc/preinit for initramfs to work. I tried init=/bin/sh as well as init=/sbin/init and it failed with PANIC: cannot mount root device. Though, looking through the code of linux 3.3 once again (init/main.c:302), I think I should have used rdinit=/sbin/init instead, as only this method will set ramdisk_execute_command which is checked on line init/main.c:827 with sys_access and thus triggers (or not) the kernel trying to mount root. Thanks, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] apr-util compile error
Hi, apr-utils misses a PKG_BUILD_DEPENDS:=util-linux so linking against libuuid fails. Regards, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] depending on util-linux fails for libuuid
Hi, for example apr-utils and e2fsprogs depend on util-linux proving libuuid, but util-linux does not install them into staging_dir/target.../. For libuuid, the following patch fixes this and makes apr-util build again. diff --git a/package/util-linux/Makefile b/package/util-linux/Makefile index aa45dd0..1f4eab2 100644 --- a/package/util-linux/Makefile +++ b/package/util-linux/Makefile @@ -372,6 +372,11 @@ define Package/wipefs/description libblkid. endef +define Build/InstallDev +$(INSTALL_DIR) $(1)/usr/lib + $(CP) $(PKG_BUILD_DIR)/libuuid/src/.libs/libuuid.* $(1)/usr/lib/ +endef + define Package/libblkid/install $(INSTALL_DIR) $(1)/usr/lib $(CP) $(PKG_BUILD_DIR)/libblkid/src/.libs/libblkid.so.* $(1)/usr/lib/ Regards, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCHv2] depending on util-linux fails for libuuid
Hi, for example apr-utils and e2fsprogs depend on util-linux proving libuuid, but util-linux does not install them into staging_dir/target.../. For libuuid, the following patch fixes this and makes apr-util build again. Patch rev 2: also exports uuid.h header and all other libraries included in util-linux. diff --git a/package/util-linux/Makefile b/package/util-linux/Makefile index aa45dd0..0626002 100644 --- a/package/util-linux/Makefile +++ b/package/util-linux/Makefile @@ -372,6 +372,18 @@ define Package/wipefs/description libblkid. endef +define Build/InstallDev + $(INSTALL_DIR) $(1)/usr/lib + $(CP) $(PKG_BUILD_DIR)/lib*/src/.libs/lib* $(1)/usr/lib/ + $(INSTALL_DIR) $(1)/usr/include + $(INSTALL_DIR) $(1)/usr/include/blkid + $(CP) $(PKG_BUILD_DIR)/libblkid/src/blkid.h $(1)/usr/include/blkid + $(INSTALL_DIR) $(1)/usr/include/libmount + $(CP) $(PKG_BUILD_DIR)/libmount/src/libmount.h $(1)/usr/include/libmount + $(INSTALL_DIR) $(1)/usr/include/uuid + $(CP) $(PKG_BUILD_DIR)/libuuid/src/uuid.h $(1)/usr/include/uuid +endef + define Package/libblkid/install $(INSTALL_DIR) $(1)/usr/lib $(CP) $(PKG_BUILD_DIR)/libblkid/src/.libs/libblkid.so.* $(1)/usr/lib/ Regards, M. Braun ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
