NAND Testing - 05-25 Update: New LK, Recovery.img, Kernel Updates through Recovery

[Lmiller1708]

Originally Posted by [ACL]

Loading wozzers last info about the partition info on cmd line i see this on /proc/mtd

dev: size erasesize name
mtd0: 10000000 00020000 "nand"
mtd1: 01500000 00020000 "root"
mtd2: 066c0000 00020000 "system"
mtd3: 06f60000 00020000 "userdata"

However i dont see any /dev/mtd so maybe we arent even compiling this crap in.

That's the problem I was running into, but didn't know how to compile them or even check to see if they were getting complied.

You might want to look at the Vogue for more information about MTD, I know they have updated a lot of stuff since we copied the files from them. (Assuming that's how we got started)

I really like the setup the Vogue community used for updating there NAND. If we do decide to go that route our initramfs will need a total rewriting, but we should be able to copy what they have done!

Wish I had more time to help with this... But work comes first.

[[ACL]]

Originally Posted by Lmiller1708

That's the problem I was running into, but didn't know how to compile them or even check to see if they were getting complied.

You might want to look at the Vogue for more information about MTD, I know they have updated a lot of stuff since we copied the files from them. (Assuming that's how we got started)

I really like the setup the Vogue community used for updating there NAND. If we [/SIZE]do decide to go that route our initramfs will need a total rewriting, but we should be able to copy what they have done!

Wish I had more time to help with this... But work comes first.

I need to double check the defconfig to see if we have evetything neded for this. i havent tried actually putting stuff in there, but we should be able to see the mtd in the /dev directory. Right now i dont.

I had a chat with cotulla about it. Looks like there is a lot to do still to get it to work. i'm confused about how wince does things vs native devices. alot of reading to do.

[natemcnutty]

Originally Posted by [ACL]

Loading wozzers last info about the partition info on cmd line i see this on /proc/mtd

dev: size erasesize name
mtd0: 10000000 00020000 "nand"
mtd1: 01500000 00020000 "root"
mtd2: 066c0000 00020000 "system"
mtd3: 06f60000 00020000 "userdata"

However i dont see any /dev/mtd so maybe we arent even compiling this crap in.

Well, maurice.green had something going here: http://forum.ppcgeeks.com/tp2-androi...ml#post2020361

Also, I have attached my dumps of several ROM's as far as partitions go. Finally, I know others reference mtd through cmdline like this:
mtdparts=msm_nand:0x00300000@0x1600000(boot),0x000 40000@0x01900000(root),0x066C0000@0x01940000(syste m),0x08000000@0x08000000(userdata)

[[ACL]]

Originally Posted by natemcnutty

Well, maurice.green had something going here: http://forum.ppcgeeks.com/tp2-androi...ml#post2020361

Also, I have attached my dumps of several ROM's as far as partitions go. Finally, I know others reference mtd through cmdline like this:
mtdparts=msm_nand:0x00300000@0x1600000(boot),0x000 40000@0x01900000(root),0x066C0000@0x01940000(syste m),0x08000000@0x08000000(userdata)

Yeah i was using wozzers line but thats different than what you posted. Where did u get that ?

mtdparts=msm_nand:0x10000000@0x0(nand),0x01500000@ 0x04000000(root),0x066C0000@0x05540000(system),0x0 6f60000@0x0Bc00000(userdata)

However one thing no one knows is if ous msm_nand.c is actually working. We cna pass all these params down to the kernel but if they arent showing up on linux, then something might be wrong with the driver.

Cotulla recommended we try to flash to the last portion of the partition. This way we dont kill anything important. Makes sense but i dont know enough about the partitioning to do that.

[Lmiller1708]

Originally Posted by [ACL]

Yeah i was using wozzers line but thats different than what you posted. Where did u get that ?

mtdparts=msm_nand:0x10000000@0x0(nand),0x01500000@ 0x04000000(root),0x066C0000@0x05540000(system),0x0 6f60000@0x0Bc00000(userdata)

However one thing no one knows is if ous msm_nand.c is actually working. We cna pass all these params down to the kernel but if they arent showing up on linux, then something might be wrong with the driver.

Cotulla recommended we try to flash to the last portion of the partition. This way we dont kill anything important. Makes sense but i dont know enough about the partitioning to do that.

Can just compile the msm_nand.c and add it to the NAND manually without the flash?
Not too sure we have to mess around with the mtdparts until we get the drivers working correctly.

[[ACL]]

Originally Posted by Lmiller1708

Can just compile the msm_nand.c and add it to the NAND manually without the flash?
Not too sure we have to mess around with the mtdparts until we get the drivers working correctly.

how will we know if its working right without passing mtd parts? we have to pass it right or else we wont be able to tell if its actually visible.

thoughts ?

[Lmiller1708]

Originally Posted by [ACL]

how will we know if its working right without passing mtd parts? we have to pass it right or else we wont be able to tell if its actually visible.

thoughts ?

Let me restate... We still need the mtdparts in the cmd line, But I'm not sure the values will matter that much (currently). They can always be changed later. Just my thoughts.

[[ACL]]

Originally Posted by Lmiller1708

Let me restate... We still need the mtdparts in the cmd line, But I'm not sure the values will matter that much (currently). They can always be changed later. Just my thoughts.

The mtdparts is important since we can overwrite and brick .. i dont want to brick myself :-p .. I double checked with wozzer about his values

Those values I used for mtd parts were far enough in that it couldn't overwrite the radio and hspl.

As a backup, you will probably want to do a hard reset, flash the os.nb (android) then take a full MTD backup of nand.

There you have it folks.. more reading to do

[[ACL]]

ok so i finally see the /dev/mtd . Damn rootfs didnt call my modprobe like expected.

Now i was trying to dump stuff but lot of weird errors on dmesg

[ 351.681854] msm_nand_read_oob 31a1000 1000 0 failed -74, corrected 0
[ 351.683624] msm_nand_read_oob 31a2000 1000 0 failed -74, corrected 0
[ 351.685424] msm_nand_read_oob 31a3000 1000 0 failed -74, corrected 0
[ 351.687225] msm_nand_read_oob 31a4000 1000 0 failed -74, corrected 0
[ 351.689331] msm_nand_read_oob 31a5000 1000 0 failed -74, corrected 0

i dont have the balls to attempt a write yet. but if read fails i doubt write will be any better.

[guruvince]

Do it!