09:23 <inode> i've compiled linux-5.1 for an RK3229 target using a stripped down version of the multi_v7_defconfig (including CONFIG_MMC_DW=y, CONFIG_MMC_DW_ROCKCHIP=y and CONFIG_EFI_PARTITION=y), which is booted from eMMC using the "rk3229-evb.dts" modified to enable UART2's RX. the kernel detects the eMMC, but only recognizes "mmcblk0", "mmcblk0boot0", "mmcblk0boot1", and "mmcblk0rpmb" but not

09:23 <inode> mmcblk0p{0,1,2,3,4,...} that the shipped android kernel acknowledges

09:28 <inode> any ideas on what i might be missing? this is what i see from UART2: https://pastebin.com/32x4Hy0W

09:42 <DuClare> CONFIG_MTD_CMDLINE_PARTS?

09:43 <inode> yes i've got it set: CONFIG_MTD_CMDLINE_PARTS=y

09:43 <inode> i didn't have CONFIG_EFI=y set, so i'm giving that a try at the moment

09:46 <inode> no luck there either

09:49 <sphalerite> inode: how is it partitioned?

09:50 <sphalerite> inode: if it's GPT, you'll want CONFIG_EFI_PARTITION, if it's MBR you'll want CONFIG_MSDOS_PARTITION

09:52 <inode> i've been under the assumption that it is GPT and used CONFIG_EFI_PARTITION, but i'll check

09:55 <sphalerite> inode: you could also try `busybox fdisk -l` to see if busybox can find a partition table

09:58 <inode> unfortunately this build of busybox doesn't contain that applet, so i'm restoring the android kernel and a somewhat usable rootfs atm to hopefully check mmcblk0 with fdisk and sgdisk

09:59 <sphalerite> ah ok

10:16 <inode> fdisk -l, and sgdisk -p display empty partition tables

10:16 <DuClare> cat /proc/cmdline?

10:17 <DuClare> Did you modify the (u-)boot script in order to boot the mainline kernel?[

10:17 <inode> line 108 https://pastebin.com/32x4Hy0W

10:18 <inode> no, i didn't modify u-boot at all

10:18 <inode> i modified the CMDLINE in the parameters file (via rkflashtool P)

10:18 <inode> i haven't touched mtdparts=... though

10:19 <DuClare> I think mtdparts is wrong

10:19 <DuClare> Specifically, rk29xxnand

10:20 <inode> ok, it is what shipped with the device

10:20 <DuClare> Yeah that's an identifier you'd find in the vendor fork

10:21 <inode> thanks

10:21 <DuClare> I never actually used mtdparts myself but I think you can just replace rk29xxnand with 0

10:22 <DuClare> (Or just switch over to gpt partitions or whatever)

10:22 <DuClare> Or not, I'm actullay not sure how the ids work

10:27 <mmind00> DuClare: if you get mmcblk0bootx though it's not nand but a real emmc

10:28 <inode> looks like drivers/mtd/cmdlineparts.c might have the answers about the mtd-id

10:28 <mmind00> inode: from what I remember Rockchip uses their own partition system (similar how they do nand partitions) on emmc devices

10:28 <mmind00> but I don't believe that is part of mainline

10:29 <sphalerite> using amarula's u-boot with u-boot SPL on my nanopi m4, cold boots work fine, but reboots don't — the SPL doesn't seem to be able to read from the SD card: mmc_load_image_raw_sector: mmc block read error

10:29 <sphalerite> any ideas how this might be fixed?

10:29 <inode> mmind00: http://rockchip.wikidot.com/partitions / http://rockchip.wikidot.com/partitions ?

10:30 <inode> sorry, last one should have been: http://opensource.rock-chips.com/wiki_Partitions

10:32 <mmind00> inode: was thinking more of https://github.com/rockchip-linux/kernel/blob/develop-4.4/block/partitions/rk.c

10:32 <DuClare> mmind00: Then I'm glad I didn't chase it and went straight to MBR partitions :)

10:33 <mmind00> DuClare: yep ... that old system is hacky and should not be used

12:19 <atopuzov[m]> wens i managed to compile and get a working uboot from an old PR by hanetzer https://github.com/atopuzov/c201/releases/download/20190513/uboot.kpart works with emmc/microsd

15:06 <inode> thanks for the pointers

15:09 <inode> mmind00: as a test, i've just ripped block/partitions/rk.{c,h} from develop-4.4 and included a pointer to rkpart_partition() in block/partitions/check.c's check_part array, amended the directory's Makefile and Kconfig for CONFIG_RK_PARTITION

15:10 <inode> and it appears that all of the partitions are detected