Ticket #799 (assigned defect)

Opened 7 years ago

Last modified 6 years ago

data abort while reading from SD Cards

Reported by: obri@… Owned by: openmoko-kernel
Priority: high Milestone:
Component: u-boot Version: 2007.2
Severity: major Keywords:
Cc: mickey@…, buglog@…, werner@…, Tuukka.Hastrup@…, openmoko@… Blocked By:
Blocking: Estimated Completion (week):
HasPatchForReview: no PatchReviewResult:
Reproducible:

Description

GTA01Bv4 # version
U-Boot 1.2.0-moko11_r0 (Sep 1 2007 - 16:49:19)
GTA01Bv4 #

============ using a SD Card from uboot ============

GTA01Bv4 # mmcinit
trying to detect SD Card...
Manufacturer: 0x03, OEM "SD"
Product name: "SU512", revision 8.0
Serial number: 2957172175
Manufacturing date: 3/2007
CRC: 0x76, b0 = 1
READ_BL_LEN=14, C_SIZE_MULT=7, C_SIZE=4095
size = 0
GTA01Bv4 # ext2ls mmc 0
data abort
pc : [<33fa0c28>] lr : [<33fa1434>]
sp : 33f14a2c ip : 00000400 fp : 33f14a90
r10: 33f14a40 r9 : 00000000 r8 : 33f17fdc
r7 : 33f25420 r6 : 00000001 r5 : 067e2948 r4 : 3400687c
r3 : 00000001 r2 : 33f25424 r1 : 33f14e38 r0 : 00000001
Flags: nZCv IRQs on FIQs off Mode SVC_32
Resetting CPU ...

============ booting from SD Card ==============
GTA01Bv4 # setenv bootargs root=/dev/mmcblk0p1 rootdelay=5
console=ttySAC0,115200 console=tty0 loglevel=8 ${mtdparts} ro
GTA01Bv4 # mmcinit
trying to detect SD Card...
Manufacturer: 0x03, OEM "SD"
Product name: "SU512", revision 8.0
Serial number: 2957172175
Manufacturing date: 3/2007
CRC: 0x76, b0 = 1
READ_BL_LEN=14, C_SIZE_MULT=7, C_SIZE=4095
size = 0
GTA01Bv4 # ext2load mmc 0 0x32000000 /boot/uImage
data abort
pc : [<33fa0c28>] lr : [<33fa1434>]
sp : 33f1495c ip : 00000400 fp : 33f149c0
r10: 33f14970 r9 : 00000000 r8 : 33f17fdc
r7 : 33f25420 r6 : 33f25420 r5 : 067e292e r4 : 33f14cfe
r3 : 00000001 r2 : 33f25424 r1 : 33f14d68 r0 : 00000001
Flags: nZCv IRQs on FIQs off Mode SVC_32
Resetting CPU ...

Attachments

799_workaround.patch (314 bytes) - added by mail@… 7 years ago.
Workaround to allow SD booting
799_workaround_2.patch (313 bytes) - added by mail@… 7 years ago.
Workaround to allow SD booting #2
om799-workaround.patch (321 bytes) - added by mail@… 7 years ago.
Previous patch, diffed from a higher directory level so it can be applied by bitbake's do_patch

Change History

comment:1 Changed 7 years ago by Tuukka.Hastrup@…

  • Cc Tuukka.Hastrup@… added

I have a system on the SD card (ext3 file system, thus valid ext2 too) which I
was able to boot fine with the U-Boot that shipped with my phone (U-Boot
1.2.0-moko9_r8_70124c2602ae2d4c5d3dba05b482d91548242de8_0_2040 (May 19 2007 -
06:59:48)).

U-Boot now crashes when I try to access the SD card, this is after I upgraded
U-Boot to the 2007.8 snapshot version (U-Boot 1.2.0-moko9_r0 (Aug 19 2007 -
19:17:12)) Now I realize the version number is actually lower, did I in effect
downgrade instead of upgrade?

In any case, here is the output of what happens with the 2007.8 snapshot U-Boot:

GTA01Bv4 # version
U-Boot 1.2.0-moko9_r0 (Aug 19 2007 - 19:17:12)

GTA01Bv4 # mmcinit
trying to detect SD Card...
Manufacturer: 0x03, OEM "SD"
Product name: "SU512", revision 8.0
Serial number: 2957172264
Manufacturing date: 3/2007
CRC: 0x27, b0 = 1
READ_BL_LEN=14, C_SIZE_MULT=7, C_SIZE=4095
size = 0
GTA01Bv4 # ext2ls mmc 0

<no output, USB serial console detaches, U-Boot restarts>

GTA01Bv4 # mmcinit
trying to detect SD Card...
Manufacturer: 0x03, OEM "SD"
Product name: "SU512", revision 8.0
Serial number: 2957172264
Manufacturing date: 3/2007
CRC: 0x27, b0 = 1
READ_BL_LEN=14, C_SIZE_MULT=7, C_SIZE=4095
size = 0
GTA01Bv4 # ext2load mmc 0 0x32000000 /boot/uImage
undefined instruction
pc : [<33f14ab8>] lr : [<33f9b3d0>]
sp : 33f1497c ip : 00000000 fp : 33f149c0
r10: 33f14970 r9 : 00000000 r8 : 33f17fdc
r7 : 33f26190 r6 : 33f26190 r5 : 067e292e r4 : 00000000
r3 : 00000001 r2 : 33f14d68 r1 : 00000000 r0 : 00000000
Flags: nZCv IRQs on FIQs off Mode SVC_32
Resetting CPU ...

comment:2 Changed 7 years ago by laforge@…

  • Cc mickey@… added

yes, you might well have downgraded your u-boot. AFAIR the 2007.2 developers
froze the patchset, while the 2007.1 continues to build the current u-boot head
+ patchset head.

mickey: ?

comment:3 Changed 7 years ago by mickey@…

No, we just broke adding the SRCREV into U-Boot again :/ I'll add it again and
ping here, so you can test again, Daniel.

comment:4 Changed 7 years ago by mickey@…

Does a brand-new U-Boot still exhibit this problem, Obri?

comment:5 Changed 7 years ago by scaredycat@…

Found a u-boot that seems to work 1.2.0-moko9_r13_0_2632_0 (copy available at
http://buildhost.automated.it/u-boot-good-for-sd-boot-r13_0_2632_0.bin )

comment:6 Changed 7 years ago by obri@…

U-Boot 1.2.0-moko9_r13_0_2632_0 is OK here with the bundled 512M Card, but i get
timeouts with a hama 2G Micro SD Card.

comment:7 Changed 7 years ago by willie_chen@…

  • Owner changed from laforge@… to michael@…

comment:8 Changed 7 years ago by openmoko@…

  • Cc openmoko@… added

comment:9 Changed 7 years ago by mickey@…

  • Owner changed from michael@… to willie_chen@…

reassigning to owner of selected component. Willie, that's important for our
developers as well as for people who want to try new images without overwriting
their "production" one.

Changed 7 years ago by mail@…

Workaround to allow SD booting

comment:10 Changed 7 years ago by mail@…

  • attachments.isobsolete changed from 0 to 1

comment:11 Changed 7 years ago by mickey@…

Obri, does it fix it for you as well?

Changed 7 years ago by mail@…

Previous patch, diffed from a higher directory level so it can be applied by bitbake's do_patch

comment:12 Changed 7 years ago by mail@…

  • attachments.isobsolete changed from 0 to 1

comment:13 Changed 7 years ago by obri@…

Unfortunately not :-( is there a new official u-boot image?

comment:14 Changed 7 years ago by werner@…

  • Cc werner@… added

It would be good if someone could find out why not using a dynamically
allocated array (sometimes ?) makes the problem go away. E.g., is it

  • a compiler bug ?
  • an error in the array size ?
  • something entirely different that we just happen to touch ?

From what I see in the code, it should have the intended behaviour.
So we should find out what's really happening there.

comment:15 Changed 6 years ago by roh

  • Owner changed from willie_chen@… to willie_chen

comment:16 Changed 6 years ago by john_lee

  • Owner changed from willie_chen to openmoko-kernel
  • Status changed from new to assigned
  • HasPatchForReview unset

willie_chen doesn't work for us anymore.

Note: See TracTickets for help on using tickets.