**** BEGIN LOGGING AT Mon Apr 15 02:59:57 2019 Apr 15 08:13:51 jow: any idea about https://github.com/openwrt/packages/pull/8569 ? I don't understand opkg that well. Apr 15 09:48:57 mangix: pong (if still needed) Apr 15 09:59:25 Hi guys have you ever seen a demo unit of a fritzbox 7490? Looks like that they smash with a hammer the chips to prevent the demo unit to work! A friend of mine just gave me his demo unit just do see what was missing... but the solution is quite dramatic! Apr 15 09:59:57 https://photos.app.goo.gl/x4QD7cFTHt6sgSJA7 Apr 15 10:00:06 https://photos.app.goo.gl/6gJ4GYWbK8hgaDGNA Apr 15 10:13:33 mangix: yes, a package name may never be the same as the virtual provided name Apr 15 10:24:33 is there someone that know how is the boot process of the fritzbox done? Apr 15 10:24:40 it have both nand and spi Apr 15 10:39:17 mkresin, can I ask you some information about a lantiq router? Apr 15 10:40:28 is there here a qualcom expert? how the hell can another CPU (Lantiq) boot another qualcom CPU ? (there is no secondary spi splash) Apr 15 10:52:43 hitech95: e.g. pinstrap to boot from UART, then supply a firmware through hardwired UART. Or maybe the qualcomm chip supports EP mode, and provides a pci interface for loading a firmware (quantenna wifi cards work that way) Apr 15 10:54:55 KanjiMonster, its a QCA9558. I was wondering how the hell that Soc was booting. I have a dead 7490 and I'm tring to find out how it works :D looks like that the QCA have SPI testpoints! Apr 15 10:55:08 I don't remember if the fritzbox boot from spi and than from nand Apr 15 10:55:37 in that case I could also make the Lantiq soc working! (SPI pads are fine... but the nand one are ripped off) Apr 15 10:58:50 hitech95: are you sure the lantiq chip is the main SoC? They could also boot the QCA9558 which then uses the lantiq chip as a pcie connected modem (I have seen setups like that before) Apr 15 11:01:09 KanjiMonster, yup, ethernet, NAND and NOR is connected to the lantiq chip. Apr 15 11:02:55 https://photos.app.goo.gl/yfyKhaVhzbo7Hbyx5 I ahve reworked the board to remove the shorts and the broken chips... Apr 15 11:05:57 https://forum.openwrt.org/t/support-fritzbox-7490/4112/17 Apr 15 11:06:43 hitech95: I see Apr 15 11:07:42 "Boot from external CPU via PCIE, USB, xMII, eliminating need for external flash" says AR9344; the same is probably true for the QCA9558 Apr 15 11:10:26 KanjiMonster, Really interestng I had no idea that it was possible! Apr 15 11:11:49 I wonder if they did something funny and have VRX288 <=> (xMII) QCA9558 <=> (PCIe) QCA9880, and basically use the QCA9558 as a full-mac simultaneous dual band wifi device Apr 15 11:14:21 KanjiMonster, yup thay probably have done exactly that! Apr 15 11:15:24 but the pcie an the qca9558 lines don't have continuity on the one of the qca9880 so I have no idea if the fpga is also doing something in the middle! Apr 15 11:24:49 hitech95: regardless on how it is setup, it sounds like it will be a huge PITA to get wifi working, especially with AVM always only releasing the bare minimum (and lol @ gpl-wlan.tar.gz from the gpl tarball) Apr 15 11:26:43 KanjiMonster, yup i've seen that gpl tarball! ANyway this is was a demo unit from a friend of mine that work on a electronic shop. AVM give them the new units. He tried to turn on the demo and it wasn't woking and he gived it to me. When i have opened it i've seen the smashed chips! Apr 15 11:28:03 Would be cool to have a bootable wca and lantiq on that shitty board! Apr 15 11:29:18 wondering if the qca have a bootrom like the lantiq with the "console" Apr 15 14:40:13 ping jow Apr 15 17:19:27 cotequeiroz: pong (but will be away again in a few minutes) Apr 15 18:09:46 jow: I'm working on the circular dependency problems affecting perl. Apr 15 18:11:45 Hopefully I have caught you before your leave. Have you seen my latest (4 hours ago) comment in https://github.com/openwrt/packages/issues/8535? Apr 15 18:13:19 I noticed that scripts/package-metadata.pl does not generate proper dependencies when they are OR'ed (||). Apr 15 18:14:31 TARGET_x86||TARGET_x86_64: gets coded as $(if $(CONFIG_(TARGET_x86||TARGET_x 86_64)),...) Apr 15 18:18:22 ... in tmp/.packagedeps. I have corrected that, but since it has been "working" this way for a while, I would like to make sure this is not "by design". **** BEGIN LOGGING AT Mon Apr 15 18:56:52 2019 Apr 15 19:34:13 nbd around? Apr 15 19:36:26 https://pastebin.com/Wuu8HUjc Apr 15 19:36:37 or anyone who wants to fix conntrack in 4.19 really Apr 15 19:46:52 out of curiosity, what's the issue? Apr 15 20:00:57 all established sessions have their timer set incorrectly (to the retrans timeout) at 3600 seconds Apr 15 20:01:04 or 300? Apr 15 20:01:06 i dont remember the default Apr 15 20:01:28 300 Apr 15 20:01:42 quite fun when your ssh sessions timeout after 5 minutes of inactivity lol Apr 15 20:02:48 Netfilter/Conntrack issues in 4.19 with 613-netfilter_optional_tcp_window_check.patch https://pastebin.com/uapWz27s Apr 15 20:04:40 moep Apr 15 20:05:47 simple change to the added logic in 4.19 seems to fix it. not sure if it's the most ideal way to short circuit all of that stuff in 4.19, though Apr 15 20:59:04 4.19 is going to be hillarious. Going to need to add broken-flash-reset for a bunch of boards Apr 15 21:01:24 well, just wait to 4.19.100, problems fixed :) Apr 15 22:12:43 ynezz: no broken-flash-reset is a new feature Apr 16 02:11:22 what does this mean needs reviewer https://github.com/openwrt/openwrt/pull/1359 **** ENDING LOGGING AT Tue Apr 16 02:59:57 2019