[#3832] kernel build with Debug option enabled fails to boot up
Submitted By: Mingquan Pan
Open Date
2008-01-16 00:41:52 Close Date
2008-01-21 20:10:07
Priority:
Medium Assignee:
Michael Hennerich
Status:
Closed Fixed In Release:
N/A
Found In Release:
N/A Release:
Category:
N/A Board:
STAMP
Processor:
BF537 Silicon Revision:
Is this bug repeatable?:
Yes Resolution:
Fixed
Uboot version or rev.:
Toolchain version or rev.:
Oct_07
App binary format:
N/A
Summary: kernel build with Debug option enabled fails to boot up
Details:
kernel build with Debug option enabled fails to boot up since Michael's gpio patch r4088.
U-Boot 1.1.6-svn997 (ADI-2007R2-pre) (Nov 1 2007 - 14:45:30)^M
^M
CPU: ADSP bf537-0.2 (Detected Rev: 0.2)^M
Board: ADI BF537 stamp board^M
Support: blackfin.uclinux.org/^M
Clock: VCO: 500 MHz, Core: 500 MHz, System: 100 MHz^M
SDRAM: 64 MB^M
FLASH: 4 MB^M
In: serial^M
Out: serial^M
Err: serial^M
Net: Blackfin EMAC^M
I2C: ready^M
Hit any key to stop autoboot: 2 ^H^H^H 0 ^M
bfin> set bootargs root=/dev/mtdblock0 rw earlyprintk=serial,uart0,57600^M
bfin> set serverip 10.100.4.174^M
bfin> set ipaddr 10.100.4.50^M
bfin> save^M
Saving Environment to Flash...^M
. done^M
Un-Protected 1 sectors^M
Erasing Flash...^M
. done^M
Erased 1 sectors^M
Writing to Flash... done^M
. done^M
Protected 1 sectors^M
bfin> tftpboot 0x2000000 linux^M
Using Blackfin EMAC device^M
TFTP from server 10.100.4.174; our IP address is 10.100.4.50^M
Filename 'linux'.^M
Load address: 0x2000000^M
Loading: *^H#################################################################^M
#################################################################^M
#################################################################^M
#################################################################^M
#################################################################^M
#################################################################^M
##############################################################
...
#################################################################^M
#################################################################^M
#################################################################^M
#################################################################^M
#######################^M
done^M
Bytes transferred = 22745618 (15b1212 hex)^M
Image size is 15b1212
bfin> bootelf^M
Loading .text @ 0x00001000 (1030848 bytes)^M
Loading .rodata @ 0x000fd000 (230040 bytes)^M
Loading __ksymtab @ 0x00135298 (14400 bytes)^M
Loading __ksymtab_gpl @ 0x00138ad8 (4016 bytes)^M
Loading __ksymtab_strings @ 0x00139a88 (44100 bytes)^M
Loading __param @ 0x001446cc (260 bytes)^M
Loading .data @ 0x00145000 (69632 bytes)^M
Loading .init.text @ 0x00156000 (94692 bytes)^M
Loading .init.data @ 0x0016d1e4 (3164 bytes)^M
Loading .init.setup @ 0x0016de40 (612 bytes)^M
Loading .initcall.init @ 0x0016e0a4 (436 bytes)^M
Loading .con_initcall.init @ 0x0016e258 (4 bytes)^M
Loading .init.ramfs @ 0x0016e25c (3180969 bytes)^M
Loading .text_l1 @ 0xffa00000 (8368 bytes)^M
sh_addr: FFA00000, p_paddr: 00476C05^M
Loading from: 02477000 to 00476C05, size: 8368^M
Loading .data_l1 @ 0xff800000 (192 bytes)^M
sh_addr: FF800000, p_paddr: 00478CB5^M
Loading from: 0247A000 to 00478CB5, size: 192^M
Clearing .bss @ 0x00479000 (66912 bytes)^M
## Starting application at 0x00156000 ...^M
Failed to reboot the linux kernel
~
Follow-ups
--- Sonic Zhang 2008-01-18 01:27:30
kernel of r4087 works with Debug option enabled.
--- Sonic Zhang 2008-01-18 01:29:19
Only replicate on bf537-stamp.
--- Sonic Zhang 2008-01-18 03:28:16
To be more specific, after selecting following options, kernel fails to boot.
Kernel hacking --->
[*] Kernel debugging
[*] Compile the kernel with frame pointers.
--- Michael Hennerich 2008-01-18 03:36:33
I will have a look - now.
-Michael
--- Michael Hennerich 2008-01-18 05:19:49
The problem seems to happen due to stack corruption in functions called by
_bf53x_cache_init.
In bf53x_cache_init after UNLINK, RTS we jump into the nirvana.
I suspect bfin_icache_init() or bfin_dcache_init() is causing the problem.
Further down I think we imported a bug during the recent no-mpu, mpu
restructuring.
I’ll dig further.
-Michael
--- Michael Hennerich 2008-01-18 08:28:33
Fixed -
trunk/arch/blackfin/kernel/cplb-nompu/cacheinit.c (4115 => 4116)
writes to I/DMEM_CONTROL must be followed by SSYNC.
-Michael
--- Mingquan Pan 2008-01-21 22:12:38
Yes,close.
Files
Changes
Commits
Dependencies
Duplicates
Associations
Tags
File Name File Type File Size Posted By
No Files Were Found