ChanServ changed the topic of #msm8937-mainline to: Boot Linux on your MSM8917/37/40 and QM215 mobile! | GitHub: https://github.com/msm89x7-mainline | Logs: https://oftc.irclog.whitequark.org/msm8937-mainline | Bridged to #msm8937-mainline:kde.org on Matrix
<barni2000[m]> it just backlight gpio, if you off/on the display it will work
<M0xCAFEBABE[m]> <M0xCAFEBABE[m]> "for unknown reason it doesn't..." <- oh well, it's due to how I'm appending the dtbs
<M0xCAFEBABE[m]> maybe the order is wrong
<M0xCAFEBABE[m]> if I only append santoni dtb, it boots
<barni2000[m]> check the board-id and msm-id maybe it is not correct in the kernel dt
<barni2000[m]> maybe ugg is matching when you boot on santoni because it is simple mtp board
<barni2000[m]> at least in board-id
<M0xCAFEBABE[m]> downstream have qcom,pmic-id too
<M0xCAFEBABE[m]> and some other qcom soc in mainline has that property too
<M0xCAFEBABE[m]> why don't we have it...
<barni2000[m]> <M0xCAFEBABE[m]> "downstream have qcom,pmic-id..." <- we don't need the pmic-id
<barni2000[m]> lk2nd not uses it
<barni2000[m]> maybe QCDT
<M0xCAFEBABE[m]> barni2000[m]: really? there's still parsing code for that in lk2nd
<barni2000[m]> for QCDT
<M0xCAFEBABE[m]> * really? there's still code to parse it in lk2nd
<barni2000[m]> upstream not using it lk2nd ADTB also not
<barni2000[m]> boot.img also worked with out it at the begining
<barni2000[m]> it seems redmi 7a charging without driver :(
<M0xCAFEBABE[m]> ???
<barni2000[m]> there is no qcom-qg and the charging driver in the tree but it was charged 30% from 10% to 40%
<barni2000[m]> it is scary
<M0xCAFEBABE[m]> oh
<barni2000[m]> i had similar issues with redmi 7
<barni2000[m]> that is why it was never discharged
<barni2000[m]> but my redmi 7 battery has defects now
<M0xCAFEBABE[m]> <barni2000[m]> "maybe ugg is matching when you..." <- ... yea it's loading ugg dtb
<M0xCAFEBABE[m]> just checked pstore
<barni2000[m]> i wonder how hard boot android with extlinux config
<M0xCAFEBABE[m]> damn, msm8916-mainline lk2nd also fails to boot my downstream recovery.img
<M0xCAFEBABE[m]> which is used to be fine on msm8953-mainline lk2nd
<M0xCAFEBABE[m]> says dtb not found
<M0xCAFEBABE[m]> (dtb not found happens too if I add qcom,pmic-id to mainline dts)
<barni2000[m]> i've said pmic-id is not necessary
<M0xCAFEBABE[m]> lol, msm8916-mainline lk2nd can't even boot itself for twice on santoni, on the second boot it doesn't have the proper dtb loaded it seems
<M0xCAFEBABE[m]> s/for/from/, s/twice/itself/
<barni2000[m]> Yes there will be FIXME message
<M0xCAFEBABE[m]> well, the right one comes from msm8953-mainline lk2nd, and the left one comes from the new msm8916-mainline lk2nd
<M0xCAFEBABE[m]> apparently board_hardware_id() and board_hardware_subtype() returns wrong value now
<M0xCAFEBABE[m]> 11 is QRD, 8 is MTP
<barni2000[m]> qcom,msm-id = <QCOM_ID_MSM8940 0>;
<barni2000[m]> qcom,board-id = <0x1000b 1>, <0x2000b 1>;
<barni2000[m]> board-id comes from the downstream dt
<M0xCAFEBABE[m]> yeah
<M0xCAFEBABE[m]> but the problem is
<M0xCAFEBABE[m]> lk2nd reads board-id from smem, to match dtbs
<M0xCAFEBABE[m]> however smem isn't reporting the correct value right now
<M0xCAFEBABE[m]> so, matching fails
<barni2000[m]> so board_id is <11 1>
<barni2000[m]> let me try with my santoni
<M0xCAFEBABE[m]> with the same value
<M0xCAFEBABE[m]> printing with %d shows 11
<M0xCAFEBABE[m]> printing with %x shows 100b
<M0xCAFEBABE[m]> s/100b/1000b/
<barni2000[m]> btw there are multiple variant from santoni
<M0xCAFEBABE[m]> <M0xCAFEBABE[m]> "image.png" <- here msm8916-mainline lk2nd is detecting `<8 0>`
<M0xCAFEBABE[m]> * `<8 0>`, which should be `<11 1>`
<M0xCAFEBABE[m]> * here you can see msm8916-mainline lk2nd, * `<8 0>`, which should be `<11 1>`
<barni2000[m]> 8 0 is ugg
<M0xCAFEBABE[m]> i'm sure both of the logs comes from the same device
<M0xCAFEBABE[m]> * i'm sure the log on both sides comes from the same santoni
<M0xCAFEBABE[m]> and yeah, santoni was loading ugg dtb with msm8916-mainline lk2nd
<barni2000[m]> it seems fixing the boot issue, lk2nd from lk2nd
<barni2000[m]> btw 0xb is 11
<barni2000[m]> is there any chance santoni have buggy bl?
<barni2000[m]> or lk2nd
<barni2000[m]> i am thinking maybe that is why i cannot make a proper lk2nd dt for tiare
<M0xCAFEBABE[m]> <barni2000[m]> "i am thinking maybe that is..." <- you may need to add a dedicated target for it...
<barni2000[m]> i was tried to use special match properties
<barni2000[m]> panel-match should be enough but order is matter
<M0xCAFEBABE[m]> or, well
<M0xCAFEBABE[m]> there's really a good cmdline parameter from stock which you could use for matching
<barni2000[m]> nvm i will care about tiare when i will have time for that
<barni2000[m]> match-cmdline also good for this
<barni2000[m]> but match panel also should work
<M0xCAFEBABE[m]> tiare shares panels with riva iirc
<barni2000[m]> yes
<barni2000[m]> but it have different board_id
<M0xCAFEBABE[m]> btw, that logic would work on all xiaomi msm89x7 except ugg,ugglite,prada as well
<barni2000[m]> mtp boards
<M0xCAFEBABE[m]> the main thing is, different ODM vendor
<M0xCAFEBABE[m]> only wingtech ones has the board_id= parameter
<M0xCAFEBABE[m]> all begins with S885
<M0xCAFEBABE[m]> * all the values begins with S885
<M0xCAFEBABE[m]> well, we can make use of the Unexpected DTB selected by bootloader, need to override logic
<M0xCAFEBABE[m]> have to keep only one entry on qcom,board-id
<M0xCAFEBABE[m]> s/entry/item/, s//`/, s//`/
<M0xCAFEBABE[m]> s/entry/component/, s//`/, s//`/
<M0xCAFEBABE[m]> barni2000: could u append all the dtbs and try this?
<barni2000[m]> cat Image.gz dt1... dtn?
<M0xCAFEBABE[m]> yep
<M0xCAFEBABE[m]> the order I used is msm8917-xiaomi-* msm8937-xiaomi-* msm8940-xiaomi-*
<barni2000[m]> modem is up on rolex
<M0xCAFEBABE[m]> what functions are working right now?
<M0xCAFEBABE[m]> * what modem functions are
<barni2000[m]> data, call, sms i will try gps
<M0xCAFEBABE[m]> nice
<barni2000[m]> i have 4G
<M0xCAFEBABE[m]> ah also btw, have you tried booting the recovery.img which I sent you earlier on rolex and riva?
<barni2000[m]> no
<barni2000[m]> or i dont remember :D
<M0xCAFEBABE[m]> ah, try it please...
<M0xCAFEBABE[m]> <M0xCAFEBABE[m]> "barni2000: could u append all..." <- cuz i've also done for rolex and riva on this commit
<M0xCAFEBABE[m]> if the recovery.img boots, then that's unnecessary change and I'll revert it
<M0xCAFEBABE[m]> and if that commit is proven to be working fine, i'll submit it
<barni2000[m]> btw maybe these ids should be changed in the kernel also
<M0xCAFEBABE[m]> booted the kernel without changes
<barni2000[m]> ig for gps i will need bind memshare drive
<barni2000[m]> s/drive/driver/
<barni2000[m]> GPS is up and running :D
<barni2000[m]> i think you have sent me an lk2nd.img as a recovery.img
<barni2000[m]> now i have booted from lk and it is lk2nd
<barni2000[m]> * have booted it from lk
<M0xCAFEBABE[m]> <barni2000[m]> "i think you have sent me an..." <- the real recovery image is at 512KB offset
<M0xCAFEBABE[m]> if you don't want to flash the entire image
<M0xCAFEBABE[m]> dd if=recovery.img of=out.img bs=524288 skip=1