**** BEGIN LOGGING AT Mon Apr 04 02:59:58 2016
Apr 04 06:51:10 <diizzy_> Hmm...
Apr 04 06:51:52 <diizzy_> 1. Ripped out the SD-driver patch for the kernel (cleaned and prepared)
Apr 04 06:52:05 <diizzy_> 2. Uncommented the lines in config-4.4
Apr 04 06:52:16 <diizzy_> 3. No active packages in menuconfig
Apr 04 06:52:33 <diizzy_> The darned SD/MMC driver for MT7621 still gets built
Apr 04 06:53:02 <diizzy_> This is pretty much breaking all devices, can anyone who knows the inner workings have a look at where it gets pulled?
Apr 04 07:10:31 <russell--> diizzy_ why is it breaking things?
Apr 04 07:10:49 * russell-- had to specifically add it for apu2
Apr 04 07:11:23 <russell--> also, git grep is your friend
Apr 04 07:13:17 <diizzy_> russell--: it's broken and panics
Apr 04 07:13:34 <diizzy_> I've removed everything I've found and it still keeps getting added
Apr 04 07:13:48 <diizzy_> https://dev.openwrt.org/ticket/21834
Apr 04 07:13:54 <diizzy_> it's a long standing issue
Apr 04 07:13:54 <russell--> have you make target/linux/clean'd?
Apr 04 07:13:57 <diizzy_> yes
Apr 04 07:15:11 <diizzy_> and I've used grep, no go :(
Apr 04 07:15:44 <russell--> what are you grepping for?
Apr 04 07:16:09 <diizzy_> mtk-sd
Apr 04 07:16:12 <diizzy_> https://dev.openwrt.org/browser/trunk/target/linux/mediatek/config-4.4?rev=49098#L288
Apr 04 07:16:23 <diizzy_> this is also wrong afaik
Apr 04 07:16:41 <diizzy_> and it's an optional module (in menuconfig)
Apr 04 07:18:06 <russell--> CONFIG_MMC_SDHCI_PLTFM
Apr 04 07:18:32 <diizzy_> uhm...
Apr 04 07:18:34 <diizzy_> hold on
Apr 04 07:19:17 <russell--> and all the other CONFIG_MMC* stuff?
Apr 04 07:19:45 <diizzy_> I'm confused
Apr 04 07:20:01 <diizzy_> mt7621 (and friends) uses the ramips dir
Apr 04 07:20:11 <diizzy_> what's the mediatek dir?
Apr 04 07:21:27 <diizzy_> ramips has no config-4.4?
Apr 04 07:22:15 <russell--> Mediatek Ralink ARM
Apr 04 07:22:54 <diizzy_> so uhm...
Apr 04 07:23:02 <diizzy_> where's the kernel config for ramips?
Apr 04 07:23:23 <diizzy_> ahh... it's inside each directory
Apr 04 07:23:38 <diizzy_> and it doesn't define mmc
Apr 04 07:23:50 <Necrosporus> There are some router with 128 or more flash. What's the point of using such large roms if 16m seems to be more than enough for OpenWRT with all modules one may want?
Apr 04 07:24:40 <diizzy_> Necrosporus: becase some applications are larger?
Apr 04 07:24:57 <diizzy_> Samba 4.X comes to mind, DLNA-software is usually quite large too
Apr 04 07:25:08 <russell--> ask the manufacturer?
Apr 04 07:25:28 <Necrosporus> Is samba that big that it could fill 256m chip?
Apr 04 07:25:31 <russell--> also, NAND is cheap
Apr 04 07:25:42 <russell--> might be the cheapest part they can find
Apr 04 07:25:54 <Necrosporus> but is 256m nand cheaper than 16m SPI NOR?
Apr 04 07:26:15 <diizzy_> russell--: hmm....
Apr 04 07:26:53 <russell--> diizzy_: are you using git?
Apr 04 07:27:04 <diizzy_> Yeah
Apr 04 07:27:18 <russell--> you know about "git grep"
Apr 04 07:27:28 <diizzy_> no?
Apr 04 07:27:44 <russell--> worth readin up on
Apr 04 07:27:56 <diizzy_> I'm not sure how that would help
Apr 04 07:28:05 <diizzy_> there's some leakage in the buildsystem
Apr 04 07:28:12 <russell--> finding things in the git tree
Apr 04 07:28:26 <russell--> ignores the stuff not added
Apr 04 07:28:46 <diizzy_> Guess I could do a new checkout, rip-out the driver patch
Apr 04 07:29:05 <diizzy_> if something doesn't get cleaned correctly
Apr 04 07:29:20 <russell--> checkout a branch, make changes in your branch
Apr 04 07:29:36 <russell--> git checkout -b my-carnage
Apr 04 07:30:18 <russell--> that way, you can track master, rebase your changes, etc
Apr 04 07:33:27 <diizzy_> true
Apr 04 07:38:55 <russell--> diizzy_: target/linux/ramips/mt7620/profiles/zbt.mk:             kmod-scsi-core kmod-scsi-generic kmod-mmc kmod-sdhc
Apr 04 07:40:23 <diizzy_> russell--: yeah, but even if you rip that out...
Apr 04 07:40:37 <diizzy_> I'm using the witi board (mqmaker.mk)
Apr 04 07:44:00 <diizzy_> Hmm...
Apr 04 07:44:01 <russell--> cd target/linux/ramips ; git grep sdhci
Apr 04 07:44:08 <diizzy_> now it fails...
Apr 04 07:44:12 <diizzy_> wtf
Apr 04 07:44:16 <russell--> what fails
Apr 04 07:44:17 <russell--> ?
Apr 04 07:44:44 <diizzy_> http://paste.ubuntu.com/15608167/
Apr 04 07:45:54 <russell--> what have you changed?  git diff
Apr 04 07:46:36 <diizzy_> From what I undertstanddeleted:    target/linux/ramips/patches-4.4/0046-mmc-MIPS-ralink-add-sdhci-for-mt7620a-SoC.patch
Apr 04 07:46:39 <diizzy_> wee
Apr 04 07:46:40 <diizzy_> From what I undertstanddeleted:    target/linux/ramips/patches-4.4/0046-mmc-MIPS-ralink-add-sdhci-for-mt7620a-SoC.patch
Apr 04 07:46:45 <diizzy_> err
Apr 04 07:46:58 <diizzy_> deleted:    target/linux/ramips/patches-4.4/0046-mmc-MIPS-ralink-add-sdhci-for-mt7620a-SoC.patch
Apr 04 07:47:06 * russell-- guesses you don't need to touch the patches, just don't build the mmc stuff
Apr 04 07:47:07 <diizzy_> paste fail
Apr 04 07:47:19 <diizzy_> as I said, it's leaking somewhere
Apr 04 07:47:56 <diizzy_> and I fail to see why that should break stuff
Apr 04 07:52:34 <russell--> do you know about searching in menuconfig?
Apr 04 07:53:14 <russell--> slash then search word, e.g. "sdhci" will show various dependencies and their state
Apr 04 07:53:44 <diizzy_> Yes
Apr 04 07:54:29 <diizzy_> Anyhow... I reverted the changes, removed the shdci-stuff out of the profile and verified that the .config does not include any mmc/sdhc(i) driver
Apr 04 07:55:40 <russell--> for example, http://paste.ubuntu.com/15608231/
Apr 04 07:56:48 <russell--> that implies the only thing pulling in PACKAGE_kmod-sdhci is TARGET_ramips_mt7620 [=y]
Apr 04 08:03:15 <russell--> or, better (i found your target board): http://paste.ubuntu.com/15608287/
Apr 04 08:03:43 <russell--> and the look for the =y's
Apr 04 08:03:45 <diizzy_> russell--: yes, all that is disabled
Apr 04 08:04:36 <russell--> oh, and check out the device tree file, it has a reference to sdhci as well
Apr 04 08:05:58 <diizzy_> https://dev.openwrt.org/browser/trunk/target/linux/ramips/dts/WITI.dts#L18
Apr 04 08:08:58 <russell--> with it all disabled, does it build?
Apr 04 08:10:05 <diizzy_> Yes
Apr 04 08:10:50 <diizzy_> I just did a fresh checkout, it'll probably take 30min or so to compile everything
Apr 04 08:11:42 <russell--> look at the openwrt and kernel .config files
Apr 04 08:12:37 <diizzy_> there's the kernel .config files?
Apr 04 08:12:44 <russell--> yeah, in the build_dir
Apr 04 08:12:50 <diizzy_> where*
Apr 04 08:12:52 <diizzy_> :)
Apr 04 08:13:03 <diizzy_> ahh, I'l have a look
Apr 04 08:13:14 <diizzy_> thanks
Apr 04 08:14:59 <russell--> build_dir/target-mipsel_24kec+dsp_musl-1.1.14/linux-ramips*/linux-4.4.6/.config or something like that
Apr 04 08:31:53 <diizzy_> Hmm.. doesn't that patch do the exact opposite? https://patchwork.ozlabs.org/patch/605647/
Apr 04 08:39:56 <russell--> maybe the minus signs mean something
Apr 04 08:41:10 <diizzy_> it adds more modules that it removes
Apr 04 08:41:23 <diizzy_> unless I'm reading it wrong
Apr 04 08:41:30 <diizzy_> than*
Apr 04 08:41:59 <diizzy_> so....
Apr 04 08:42:06 <diizzy_> make[3] -C target/linux install
Apr 04 08:43:57 <russell--> diizzy_: note the difference between "kmod-foo" and "-kmod-foo"
Apr 04 08:44:24 <diizzy_> uhm.... ahh
Apr 04 08:46:59 <diizzy_> moment of truth =)
Apr 04 08:47:17 <russell--> do you have a serial console?
Apr 04 08:47:52 <diizzy_> yes
Apr 04 08:48:03 <russell--> good
Apr 04 08:51:22 <diizzy_> hmm...
Apr 04 08:51:25 <diizzy_> hah
Apr 04 08:51:28 <diizzy_> So...
Apr 04 08:51:33 <diizzy_> I think I have found the issue
Apr 04 08:59:27 <diizzy_> Yey
Apr 04 08:59:41 <diizzy_> So... a few notes
Apr 04 08:59:51 <diizzy_> tftpboot is broken
Apr 04 09:00:23 <diizzy_> bootm does not imply that you actually want to boot from memory by default
Apr 04 09:00:57 <diizzy_> and I'm not sure if clean actually removes the module
Apr 04 09:01:03 <diizzy_> the clean build booted "fine"
Apr 04 10:24:11 <diizzy_> russell--: I've submitted a patch to disable the sd driver by default
Apr 04 10:56:56 <builder_> build #241 of rb532 is complete: Failure [failed shell_18 compile_9]  Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/241
Apr 04 14:53:16 <OpenWrt-Commits> rmilecki 49115 branches/chaos_calmer/target/linux/bcm53xx/patches-3.18/ (10 files) bcm53xx: update BCM5301X patches
Apr 04 14:53:22 <OpenWrt-Commits> rmilecki 49116 branches/chaos_calmer/tools/firmware-utils/ src/oseama.c Makefile firmware-utils: add oseama tool for creating Seama entities
Apr 04 14:53:26 <OpenWrt-Commits> rmilecki 49117 /branches/chaos_calmer/target/linux/bcm53xx/image/Makefile bcm53xx: prepare for building D-Link DIR-885L image
Apr 04 21:01:11 <neoraider> nbd, any idea what's going on in this dmesg? http://paste.debian.net/plain/423938 System is based on current CC HEAD
Apr 04 21:49:10 <diizzy_> neoraider: ath9k is having a fit
Apr 04 21:49:19 <diizzy_> try trunk? :)
Apr 04 21:59:15 <neoraider> diizzy_, well, as mac80211-compat has been backported from the trunk a month ago, the versions aren't fundamentally different
Apr 04 22:03:02 <jow_laptop> neoraider: there's about ~100 tickets about the problem
Apr 04 22:03:19 <jow_laptop> there was one with a more detailled explanation but I didn't find it yet again
Apr 04 22:07:39 <neoraider> I see. I hadn't seen this one before, so I assumed it was something new, but you're right, there are tons of tickets...
Apr 04 22:12:35 <jow_laptop> I have one particular comment in mind where someone explained the exact reason for the MCS warning but I simply cannot find it again
Apr 04 22:12:51 <jow_laptop> too much metoo spam
Apr 04 22:16:32 <diizzy_> jow_laptop: can you perhaps have a look at the patch about disabling the SD card reader driver on MT7621 SoC for now?
Apr 04 22:19:12 <neoraider> jow_laptop, I think I found the comment you were refering to, https://dev.openwrt.org/ticket/6411#comment:106
Apr 04 22:19:55 <jow_laptop> neoraider: yes, that was it
Apr 04 22:20:52 <jow_laptop> we should interview nbd about this if he's back
Apr 04 22:21:06 <jow_laptop> this somehow went under the radar I believe
Apr 04 22:21:41 <jow_laptop> diizzy_: the pathc looks fine to me but blogic is doing the ramips targets so I will not interfere there
Apr 04 22:22:00 <diizzy_> I see, thanks :)
Apr 04 22:22:35 <neoraider> jow_laptop, at least it doesn't seem to kill hostapd anymore (like some earlier comments mentioned)
Apr 04 22:24:47 <jow_laptop> I think I had it a few times in my log as well but the radios are rock stabler otherwise
Apr 04 22:26:20 <neoraider> I wouldn't call it rock stable (we regularly get reports of WLAN breaking from the Gluon community), but it has vastly improved over the last years
Apr 04 22:26:42 <jow_laptop> its a phantom bug
Apr 04 22:27:21 <jow_laptop> nbd stated a few times that he is simply unable to reproduce it so its likely an obscure environmental thing / sequeunce of events or whatever
Apr 04 22:28:35 <jow_laptop> as much as I hate the legacy bcm crap I cannot recall such an issue with wl.o, but we did have wifi restart watchdogs in our ancient ff firmwares as well so maybe its just a fact of life
Apr 04 22:28:45 <jow_laptop> blame solar spots :)
Apr 04 23:47:59 <builder_> build #216 of xburst is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/216
Apr 05 00:09:37 <builder_> build #195 of ep93xx is complete: Failure [failed shell_13]  Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/195
**** ENDING LOGGING AT Tue Apr 05 02:59:58 2016