kwz has quit [Quit: The Lounge - https://thelounge.chat]
Misanthropos has quit [Ping timeout: 480 seconds]
Misanthropos has joined #openwrt-devel
filimonic has joined #openwrt-devel
rmilecki has quit [Quit: Konversation terminated!]
rmilecki has joined #openwrt-devel
rmilecki has quit []
rmilecki has joined #openwrt-devel
rmilecki has quit []
rmilecki has joined #openwrt-devel
Danct12 has joined #openwrt-devel
Danct12 has quit [Ping timeout: 480 seconds]
tSYS has quit [Quit: *squeak*]
Danct12 has joined #openwrt-devel
kwz has joined #openwrt-devel
floof58 has quit [Ping timeout: 480 seconds]
floof58 has joined #openwrt-devel
sjao2 has joined #openwrt-devel
torv has quit [Remote host closed the connection]
torv has joined #openwrt-devel
<ynezz> aparcar: sure, feel free to take over it, I would be really grateful for that, but there is much more work needed, then just simply rebasing it
<ynezz> aparcar: current approach is naive and doesn't work, just a simple test using diff comparison of diffconfig changes for each target reveals that
<ynezz> aparcar: and it doesnt work properly in image builder either, so in other words it needs to be reworked :)
valku has quit [Quit: valku]
Danct12 has quit [Quit: WeeChat 4.0.4]
<russell--> inspection doesn't reveal any obviously unpopulated pads
Danct12 has joined #openwrt-devel
Danct12 has quit [Quit: WeeChat 4.0.4]
Danct12 has joined #openwrt-devel
hanetzer has joined #openwrt-devel
hanetzer1 has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
iocampomx has quit [Ping timeout: 480 seconds]
hitech95 has joined #openwrt-devel
Tapper has quit [Read error: Connection reset by peer]
fioriceta has joined #openwrt-devel
<fioriceta> I'm looking for some help with a T56/T48 programmer and NAND
robimarko has joined #openwrt-devel
<fioriceta> so, my current situation is that I've bricked some board, desoldered the tsop-48 package, have reprogrammed it with a modified jffs2 structure (cferam.000 was wrong initially, hence this shit), and then have resoldered it onto the board (my work should be good enough, I can't see any bridged pins)
<fioriceta> but I'm still not seeing anything on the uart, there is no way to jtag this board (no schematic/visible pads, though, there were some 10k resistor pads that were... interesting, but it was next to a pull-down)
<fioriceta> I've always had this issue with NAND and that programmer, no issues with NOR (though, I used a clip, and that was serial)
<fioriceta> I think it's something to do with the spare data, not sure
goliath has joined #openwrt-devel
<KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_ath79.html has been updated. (98.7% images and 99.9% packages reproducible in our current test framework.)
rua has quit [Quit: Leaving.]
<russell--> fioriceta: does your data read back correctly?
<fioriceta> russell--: when I originally dumped it? yeah
<russell--> i mean, after you wrote it
<fioriceta> passed verification, that's for sure
<fioriceta> though, I will desolder it and check
<russell--> it seems unlikely you reprogrammed it while resoldering. are you sure you to the orientation right?
<fioriceta> yeah, everything was okay
<russell--> you got something on the uart before your NAND shenanigans?
<fioriceta> yeah, it'd get stuck at loading cferam
<fioriceta> let me see if I took a capture
<fioriceta> nah, don't have a capture of after that whole shit, but I definitely would the cferom stuff on the uart
<fioriceta> LEDs on the board, etc., would change, so it's nothing like that
<russell--> what is the device?
<fioriceta> * see
<fioriceta> it's similar to the dgnd3700v2
<fioriceta> but not close enough, I accidentally put the cferom image as the cferam
<fioriceta> dgnd4000v1
<fioriceta> same SoC, ramdisk image worked out of the box (I tried to port over the DSA stuff to v19.x.x, but gave up)
<fioriceta> sercomm made modifications to cfe :/
hanetzer has quit [Quit: WeeChat 4.0.4]
<russell--> cfe sounds like broadcom
<fioriceta> it is
<fioriceta> what's weird is that you should also be able to flash the .w file over nand/nor, and have it boot up
<fioriceta> but I can't even get that going on another board of mine that uses nand
<fioriceta> maybe rmilecki has an idea
<russell--> you saw this i presume: "DSL will not work at all on devices with BCM63xx DSL chipset"
<fioriceta> oh, I don't care about DSL
<fioriceta> this SoC actually has 4 mips cores :P
<fioriceta> one is for gigabit accel, one is for DSL, and the other two are for linux
<rmilecki> fioriceta: sorry, I have no idea about such level stuff
<rmilecki> my development skill being with a working bootloader at minimum ;)
hanetzer has joined #openwrt-devel
<fioriceta> curious, is your openocd experience as bad as mine?
<russell--> dual band n routers are a dime a dozen these days, probably the most fruitful path would be to get something already supported
<fioriceta> *was
<fioriceta> russell--: yeah, but bmips has no backdoors
<fioriceta> maybe I'm losing my mind
<russell--> broadcom will do that to you
<fioriceta> it's a nice device tbh, if only I had the schematic ._.
<fioriceta> see, basically, I did have another idea
<fioriceta> there's a 3k pull down for nand and the spi pad is unpopulated
<russell--> i guess i'd suggest re-reading the NAND chip and confirming what you put on it. even writing back what was there originally (presuming you have a backup)
<fioriceta> there's basically only two pull downs from what I could tell
<fioriceta> yeah, I took a backup
<russell--> what voltage is the flash chip?
<fioriceta> 3.3V
<fioriceta> well, 3-3.3V, anyway
<russell--> do you have an oscilloscope?
<fioriceta> nope
<russell--> or a logic analyser?
<fioriceta> nope, none of that
<fioriceta> there's a few 6362 ref boards, so I'm thinking
<fioriceta> maybe there is a cfe bin in the gpl dumps I can use
<fioriceta> the nor pinout seemed standard
<russell--> can you get another sacrificial dgnd4000v1?
<fioriceta> all of the listings I saw were v2/would cost more than this board
<fioriceta> I originally wanted dgnd3700v1 tbh
<fioriceta> that was the real dream board of mine cause it was 6368
<fioriceta> 6362 is 6368 with shit taken from cm/DDR2, anyway
<russell--> all i know is when i hear broadcom, i think i'm back in vietnam or iraq or something ... and i was never in any of those places.
<fioriceta> I think the weirdest bmips board I've seen is one with three SoCs on it
<fioriceta> or maybe it was two, I've only seen a dual SoC board as a product anyway
<fioriceta> it is annoying tbh
<fioriceta> I don't think SPI is pulled down as a slave by default, but that'd assume that this SoC is even new enough to support that feature
<fioriceta> hmm, it does
<fioriceta> hmm, unfortunate
<fioriceta> it's master by default
<fioriceta> I fucking hate the vendors, fucking sercomm
<fioriceta> russell--: I do wonder now
<fioriceta> maybe I do need the spare data included in the dump
<fioriceta> well, fuck
Danct12 has quit [Quit: WeeChat 4.0.4]
<fioriceta> fuck
<fioriceta> going to try one more thing, think this might work
<fioriceta> russell--: reading back the flash atm, hmm
<fioriceta> seems okay, must be an ecc thing
\x has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
Danct12 has joined #openwrt-devel
<fioriceta> I don't even fucking know
<fioriceta> it is some sort of programming issue, but wtf?
Danct12 has quit [Quit: WeeChat 4.0.4]
<fioriceta> NAND is fucking cursed
<fioriceta> NOR just fucking works...
Danct12 has joined #openwrt-devel
Daaanct12 has joined #openwrt-devel
Danct12 has quit [Ping timeout: 480 seconds]
goliath has joined #openwrt-devel
Daaanct12 has quit [Quit: WeeChat 4.0.4]
hanetzer has quit [Remote host closed the connection]
hanetzer has joined #openwrt-devel
ptudor has joined #openwrt-devel
rua has joined #openwrt-devel
ptudor has quit [Ping timeout: 480 seconds]
<hitech95> does anyone know how to troubleshoot an uboot spitting UBI CRC errors in the SPI NAND? The initramfs image boot sfine and is able to correctly flash the nand.
<robimarko> Make sure U-Boot is properly detecting the NAND
<robimarko> And also make sure you are using the correct block and page sizes for the image generation
<hitech95> Loading Environment from UBI... spi-nand: spi_nand spi_nand@0: Macronix SPI NAND was found.
<hitech95> spi-nand: spi_nand spi_nand@0: 512 MiB, block size: 256 KiB, page size: 4096, OOB size: 128
<hitech95> but then:
<hitech95> ubi0 error: check_corruption: PEB 1571 contains corrupted VID header, and the data does not contain all 0xFF
<hitech95> on another device:
<hitech95> Loading Environment from UBI... spi-nand: spi_nand spi_nand@0: Micron SPI NAND was found.
<hitech95> spi-nand: spi_nand spi_nand@0: 512 MiB, block size: 256 KiB, page size: 4096, OOB size: 256
<hitech95> This one has a different NAND chip, but it is the same router model
<hitech95> different OOB size
<robimarko> Well
<robimarko> Well, OOB size is not hardcoded in the UBI image so that shouldnt matter
<robimarko> But, check against the values that Linux detects
<hitech95> Just in case this is the full bootloghttps://pastebin.com/TXE39SbD
valku has joined #openwrt-devel
<rmilecki> i've a weird change in U-Boot behaviour after switching from 21.02 to master
<rmilecki> it's about uboot-bcm4908 which is basically the same for 21.02 and master
<rmilecki> it's about "appending extra boot args to linux boot command line:"
<rmilecki> or what goes after it anyway
<rmilecki> 21.02: mtdparts=brcmnand.0:1048576(loader),534773760@1048576(image) root=/dev/ubiblock0_6 ubi.mtd=image ubi.block=0,6 rootfstype=squashfs cma=0M
<rmilecki> master: mtdparts=brcmnand.0:1048576(loader),534773760@1048576(image) root=/dev/ubiblock0_4 ubi.mtd=1 ubi.block=0,4 rootfstype=squashfs cma=0M
<rmilecki> (see ubi.mtd=...)
<rmilecki> can someone tell me why 21.02 uses mtd name and master uses mtd index?
<rmilecki> 21.02: U-Boot 2019.07-OpenWrt-unknown (Jun 19 2023 - 21:49:53 +0000)
<rmilecki> master: U-Boot 2019.07-OpenWrt-r24046+4-abceef120d (Sep 28 2023 - 10:53:15 +0000)
<fioriceta> rmilecki: did you look into the rdp stuff?
<fioriceta> I think it might only help with WAN -> LAN perf, though
<rmilecki> fioriceta: if I did it was years ago and I don't remember anymore
<fioriceta> yeah, well, it's not exactly OSS
<fioriceta> I wonder what's going to happen when v4.19 goes EOL
<robimarko> rmilecki: I doubt its U-Boot related as only change after 21.02 is YYLOC global declaration to build with GCC10
<rmilecki> exactly...
<rmilecki> it makes no sense
* rmilecki is debugging U-Boot now for the first time ever
<robimarko> You tried booting 21.02 images with the master U-Boot?
<fioriceta> rmilecki: cfe is a lot better than u-boot, mm...
<fioriceta> rip my board, no way I'm buying another one lmao
<rmilecki> 5.4:
<rmilecki> # CONFIG_CMDLINE_FORCE is not set
<rmilecki> # CONFIG_BLK_CMDLINE_PARSER is not set
<rmilecki> 6.1:
<rmilecki> CONFIG_CMDLINE_FROM_BOOTLOADER=y
<rmilecki> # CONFIG_CMDLINE_FORCE is not set
<rmilecki> new symbol
<robimarko> Well that would explain it
<rmilecki> in both cases CONFIG_CMDLINE_FORCE is disabled though
<rmilecki> so there's sth fishy i'd say
<rmilecki> fioriceta: u-boot is open source ;)
<rmilecki> fioriceta: you can at least try fixing bugs
<fioriceta> cfe chokes on loading anything bigger than a 4MiB elf file
<robimarko> Hm, it is fishy as CONFIG_CMDLINE_FROM_BOOTLOADER uses bootloader args and if not specified falls back to the default one
<fioriceta> biggest meme bootloader tbh
<rmilecki> ok, I found it...
<rmilecki> it's some stupid U-Boot addition from Broadcom:
<rmilecki> https://git.openwrt.org/?p=project/bcm63xx/u-boot.git;a=blob;f=board/broadcom/bcmbca/sdk_test_commands.c;h=3724f5e1f93e18fd8866b58f9b1942d593b6c585;hb=refs/heads/generic#l1203
<rmilecki> env set -f rootfs_opts (...)
<rmilecki> under some conditions u-boot overwrites "rootfs_opts" env variable
<rmilecki> i have to find out why that condition gets triggered with master and does not with 21.02
<rmilecki> of course Broadcom had to break sth
<fioriceta> isn't there anything that they haven't broken?
<fioriceta> rmilecki: speaking of which, I was looking at other boards of mine...
<fioriceta> so, it's got to be a programming issue
goliath has quit [Quit: SIGSEGV]
filimonic has quit [Ping timeout: 480 seconds]
<fioriceta> I am so fucking done
<fioriceta> all of my fucking bmips boards are in the fucking bin nearly, holy fuck
<fioriceta> I must've gone through like 20+ bcm boards
goliath has joined #openwrt-devel
EqUaTe has quit [Read error: Connection reset by peer]
EqUaTe has joined #openwrt-devel
Daanct12 has quit [Quit: What if we rewrite the code?]
<KGB-0> https://tests.reproducible-builds.org/openwrt/openwrt_omap.html has been updated. (11.1% images and 99.8% packages reproducible in our current test framework.)
Danct12 has joined #openwrt-devel
robimarko has quit [Remote host closed the connection]
robimarko has joined #openwrt-devel
danitool has joined #openwrt-devel
dangole has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
Danct12 has quit [Quit: What if we rewrite the code?]
goliath has joined #openwrt-devel
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
floof58 has quit [Ping timeout: 480 seconds]
floof58 has joined #openwrt-devel
robimarko has quit [Remote host closed the connection]
dangole has quit [Quit: Leaving]
<fioriceta> anyone have experience with nand and OOB?
<hitech95> <Finally I've managed to get this thing to work. some other nand chip is till not booting tho!
<fioriceta> rmilecki: I'm going to `buy big` again, and get another board
<fioriceta> I will try to not brick this one, I think
<fioriceta> I want to try upgrading the BGA DDR chip, too
<hitech95> fioriceta, hood luck.I've finally managet to have a working (partially) uboot on my stipid zyxel router.
<fioriceta> did you desolder the chip?
<fioriceta> I can't fucking get the ecc shit to work right, too many variables tbh, probably easier to do it from openocd
<hitech95> lucky for me I didn't had to. the oem bootload allowed me to boot .bin files so I've managed to test the binary.
<hitech95> But a friend of mine didnt and he had to flesh the chip multiple times with a clip
<hitech95> SPINAND
<fioriceta> lucky you, I have had to desolder tsop-48 like four times in the last few hours
<fioriceta> it's actually very easy, but I don't know wtf is happening even though I have strapped down ecc disable
<hitech95> OMG is that a RAW nand?
<fioriceta> yes
<hitech95> damm that is annoing as hell
<fioriceta> thinking it'd be easier to just buy another board and dump nand with OOB in the process
<fioriceta> and then maybe fuck with the DDR chip later on
<fioriceta> I really just want vendor kernel + some distro
<fioriceta> vendor kernel still has support, so I don't see the real issue
<hitech95> you dont have any JTAG tool to flash the NAND from there?
<fioriceta> the test pads are <0.5mm in size
<fioriceta> I don't have any wire for that atm
<hitech95> can I ask what hardware it is?
<fioriceta> some bcm board :S
<fioriceta> this one isn't bmips, though, ARM
<fioriceta> mainline openwrt really needs a separate target for ARM brcm SoCs with subtargets (similar to bmips)
<fioriceta> but not sure about actual mainlined support, I think they just upstreamed literal stub code
<hitech95> no idea I just dont buy stuff with BRCM chips in it. Except for the RPI
<fioriceta> rpi 5 is the biggest joke
<jakllsch> Rpi is the biggest joke
<fioriceta> rpi is just a fucking rebadged stb
<fioriceta> I miss having rpi zeros for €10/board tbh
<hitech95> my last pi is the PI1 rev b (the one with the mounting holes)
<hitech95> But I want to buy a CM$/5 for DIY
<hitech95> *CM4/CM5
<fioriceta> I just want a cheap ARMv8a platform with vt-x already, and the best option I found was a fucking rack-mount NAS
<fioriceta> so, no...
<fioriceta> cause I want RAID, that is
<fioriceta> why not opt for riscv/any other vendor?
<fioriceta> you're supporting DRM mafia by buying rpi
<hitech95> I've seen that is that case is cheaper to go with a oldish x86 small box PC
<fioriceta> a router does not fucking need secboot, why is the EU bothering with fucking removable batteries and shit?
<hitech95> fioriceta, I want to go with RPI so I dont have to mess with missing drivers and bad documentation by Asian vendors.
<fioriceta> heard of armbian?
<fioriceta> `raspbian` has its own patches over it, just as bad as vendor kernel
<fioriceta> SPL is a blob, anyway
<fioriceta> at least with sunxi, you can see the source
<hitech95> yea but I dont have to run basic stuff. Last time I had to write a f**ing alsa driver from scratch for the sunxi target. Just becouse I wasnt using the easy ports/function on the chip.
<fioriceta> well, tbh, my sunxi board never is reliable
<fioriceta> but I just buy the cheapest shit...
<fioriceta> secboot shit really fucking pisses me off
<hitech95> I buy lot of shit and is all in the shelfs, at least with the pis I used to to something cool back in the day
<fioriceta> I guess if you opt for emmc, you could write an out-of-tree module to load in a kernel image, but good luck when uart is disabled
<fioriceta> though, I doubt otp actually disables it permanently, might be possible to re-enable it through registers
<fioriceta> evil US vendors smh
<fioriceta> I think once shit really starts to get worse with rust, etc., I'm going to lose it
<fioriceta> but I really doubt the US will be around in a decade
<fioriceta> (china war)
<hitech95> well its time to sleep 1AM here. By!
hitech95 has quit [Quit: Leaving]
<fioriceta> tbh, anyway, I was thinking
<fioriceta> smartphones might be a better value, though, the mainlining effort might not be worth it
<fioriceta> I've given up on mainline, though, anyway
<fioriceta> really fucking sucks that only two individuals are working on bmips
Mangix has quit [Read error: No route to host]
<slh> fioriceta: could you cut down the expletives a bit, no I don't object on principle or moral grounds, but every third line ending in a fuckfest is getting boring
<fioriceta> I'll try, I guess, no guarantees
Mangix has joined #openwrt-devel
Slimey is now known as Guest1622
Slimey has joined #openwrt-devel
Slimey is now known as SlimeyX
Guest1622 is now known as Slimey