Re: memtest failed for rk3588

[Alexander Shiyan <eagle.alexander923@gmail.com>]

Hello All!

Update: I have start the second board and got similar errors.
Analysis of the code (at the 0x3fc000000), before the memory area
that is not available (at 0x3fc000200) is very similar to the code
from sdram-init.bin.
Maybe we are missing a description of some necessary memory area?

Thanks!

ср, 11 дек. 2024 г. в 18:52, Alexander Shiyan <eagle.alexander923@gmail.com>:
>
> So far I have isolated several memory pages using reserved-memory:
> <0x3 0xfc000000 0x0 0x500000>
> and
> <0x3 0xfff00000 0x0 0x100000>
> So it looks like there is a memory chip bug.
>
> barebox@Diasom DS-RK3588-BTB-EVB:/ memtest -t
> Start iteration 1 of 1.
> Testing memory space: 0x0000000000a00000 -> 0x00000000afcfffff:
> Starting data line test.
> Check for address bits stuck high.
> Check for address bits stuck low or shorted.
> Starting moving inversions test of RAM:
> Fill with address, compare, fill with inverted address, compare again
>         [################################################################]
> done.
>
> Testing memory space: 0x0000000100000000 -> 0x00000003fbffffff:
> Starting data line test.
> Check for address bits stuck high.
> Check for address bits stuck low or shorted.
> Starting moving inversions test of RAM:
> Fill with address, compare, fill with inverted address, compare again
>         [################################################################]
> done.
>
> Testing memory space: 0x00000003fc500000 -> 0x00000003ffefffff:
> Starting data line test.
> Check for address bits stuck high.
> Check for address bits stuck low or shorted.
> Starting moving inversions test of RAM:
> Fill with address, compare, fill with inverted address, compare again
>         [################################################################]
> done.
>
> Memtest successful.
>
> вт, 10 дек. 2024 г. в 14:47, Sascha Hauer <s.hauer@pengutronix.de>:
> >
> > On Tue, Dec 10, 2024 at 02:29:59PM +0300, Alexander Shiyan wrote:
> > > Hello.
> > >
> > > The dump does not cause a crash, but the output is very strange:
> > >
> > > barebox@Diasom DS-RK3588-BTB-EVB:/ md 0x3fc000100
> > > 3fc000100: 940033fd b0000060 39477ea1 91274c00               .3..`....~G9.L'.
> > > 3fc000110: 17ffffe2 91404001 b9428020 37f80180               .....@@. .B....7
> > > 3fc000120: b9428020 32010000 b9028020 52800020                .B....2 ... ..R
> > > 3fc000130: b9420822 32000042 b9020822 b9420822               ".B.B..2"...".B.
> > > 3fc000140: 521c0042 b9020822 d65f03c0 52800000               B..R"....._....R
> > > 3fc000150: 17fffff8 91404000 34000081 b9428001               .....@@....4..B.
> > > 3fc000160: 12007821 b9028001 b9420801 121f7821               !x........B.!x..
> > > 3fc000170: b9020801 b9420801 521c0021 b9020801               ......B.!..R....
> > > 3fc000180: d65f03c0 d37b7c21 71000c5f 8b010000               .._.!|{._..q....
> > > 3fc000190: 528aaaa1 72bdddc1 f9400400 b900ac01               ...R...r..@.....
> > > 3fc0001a0: 529dddc1 72aaaaa1 b90a7c01 b94a2401               ...R...r.|...$J.
> > > 3fc0001b0: 321c0021 b90a2401 b94a2403 12147463               !..2.$...$J.ct..
> > > 3fc0001c0: 54000300 52808001 1ac22022 2a030042               ...T...R" ..B..*
> > > 3fc0001d0: b90a2402 b94a2401 32000021 b90a2401               .$...$J.!..2.$..
> > > 3fc0001e0: b94a2401 360fffe1 b94a2401 121f7821               .$J....6.$J.!x..
> > > 3fc0001f0: b90a2401 b94a2401 12147421 b90a2401               .$...$J.!t...$..
> > > barebox@Diasom DS-RK3588-BTB-EVB:/ md 0x3fc000200
> > > 3fc000200: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000210: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000220: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000230: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000240: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000250: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000260: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000270: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000280: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc000290: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc0002a0: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc0002b0: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc0002c0: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc0002d0: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc0002e0: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> > > 3fc0002f0: xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx               ................
> >
> > See:
> >
> > | commit d8a6e1c2662aad2f5d9b4050a276b8e6b9fbb3b1
> > | Author: Jan Luebbe <jlu@pengutronix.de>
> > | Date:   Sun Aug 3 17:51:38 2014 +0200
> > |
> > |     ARM/mem: handle data aborts gracefully for md
> > |
> > |     Sometimes memory ranges contain inaccessible registers which trigger a
> > |     data abort when accessed. To handle this gracefully, we extend the data
> > |     abort exception handler to ignore the exception when configured to do
> > |     so.
> > |
> > |     This allows detecting inaccessible memory from the md command. It will
> > |     show XX for unreadable bytes instead.
> >
> > I don't know why you get a data abort here. The answer is probably
> > somewhere in the DMC section of the RK3588 reference manual.
> >
> > Sascha
> >
> > --
> > Pengutronix e.K.                           |                             |
> > Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
> > 31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
> > Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |