a pretty big change
Hello Paul,
thanks for the quick answer. I am sorry but I did not understand what I see in System.map:
excerpt 1 for System.Map:c0049f90 t $CODE$c0049f90 t put_unused_buffer_headc0049fd8 t L$1856c004a004 t L$1854c004a018 t $CODE$c004a018 t get_unused_buffer_headc004a110 t $CODE$c004a110 t create_buffersc004a320 t $CODE$c004a320 t end_buffer_io_async
So I thought 0xc004a078 is in get_unused_buffer_head which is in fs/buffer.c.
excerpt 2 from System.map:c0111f14 D blkdev_inode_operationsc0111f64 D def_chr_fopsc0111fa0 D chrdev_inode_operationsc0111ff0 d L$C0008c0111ffc d L$C0009c011200c d L$C0010c0112018 d L$C0011c0112028 d L$C0012c0112038 D nr_filesc0112038 d lock___76
So 0xc0112000 points into the data segment(?) as a return address for that subroutine.
Unfortunately I have no anonymous ftp server here at work, where I can put the kernel for you. And puffin.external.hp.com has no Incoming directory...
So is it ok to send Image + System.map as a tar.gz-file via mail to you?
If this problem is not of general interest and other big changes will happen soon (maybe the 2.3 merge) where the problem might go away, it is ok that you simply say "wait for that to happen" to me.
Regards
Ulrich
Quoted 154 lines · --- Paul Bame
= Hi all,
=
= I have build a new kernel with the CVS version
= 19991205 and I get the following console output
when I
= try to boot it on my 715/33:If you'll place that kernel somewhere I can get it, I'll check
this, or you can. Look into the kernel and/or System.map
for what's at0xc004a078, which is inside a procedure which will
return to0xc0112000. It looks like a null (or small) pointer
dereference because we're trying to create a virtual mapping for address 0.-P= PARISC/Linux Bootstrap Version 0.6 (interactive) = By Helge Deller & Jason Eckhardt
= Built Mon Dec 6 11:13:34 MET 1999 by root@be5359 =
= Reading parameters...done.
=
= Loading PA-RISC/Linux Kernel...
=
= Please enter the command-line:
= /stand/vmlinux HOME=/ TERM=LINUX console=tty =
= SOM-Kernel:
=aux_header_location: 00000080
= som : 00200080
=exec_dfile: 000CE000
=exec_dsize: 00081000
=exec_dmem: C00A0000
=exec_tfile: 0003E000
=exec_tsize: 0008F008
=exec_tmem: C0010000
= Code at0x00010000, size=0x0008F008
= Data at0x000A0000, size=0x00081000
= BSS at0x00121000.
=
= Transferring control to kernel. (At entry point =0x00010000)
=kernel(0x146880, 0x504010, 0x146880, 0x12c30)= Boot loader: PA/Linux, maybe PALO
= kernel free mem0x00146880
=realmode_setupexiting
= Can't insert TLB miss handler pa0x40000000va =0x00000000
= r0-30x0004fb0b0xc01120000xc004a170=
0x00000000
= r4-70xc01120000x000000000x00000400=
0x00000000
= r8-110xc01db0000x000000000x00000400=
0x00000100
= r12-150xc01a49880x000000000x00000000=
0x00000000
= r16-190x000000000x000004000x00001000=
0x0000001f
= r20-230x000000200xc01a001c0x00000001=
0xc01268b0
= r24-270x000000000x000004000x00000000=
0xc009f000
= r28-310xc01db0000x0000000b0xc01a4f40=
0xc007a39c
= sr0-30x000000000x000000000x00000000=
0x00000000
= sr4-70x000000000x000000000x00000000=
0x00000000
= iasq0x000000200x00000003
= iaoq0xc004a0780xc004a07c
= cr300x00000000ksp0x00000001kpc0xc01a4f80sar =0x00000012
= iir0x48730038
=
= The last successful booting kernel was from the CVS
= version 19991123.
=
= Any ideas ?
= Ulrich
=
= --- Paul Bame <[email protected]> wrote:
= >
= > I just committed the changes to embed real-mode code
= > in a separate
= > executable inside vmlinux. It should cause no = > functional changes
= > in theory. Read all about it in
= >arch/parisc/real/READMEand
= > ee.html. If you have terrible new problems, = > I've attached a list of CVS commands to check out
= > the files before I
= > started hacking on them. I haven't seen a shell = > prompt from this
= > set of hacks so beware...
= >
= > -Paul Bame
= >
= > prumpf- I movedkernel_threadand friends from
= > real/<something> tokernel/head.SI also
= > re-enabled BTLB in
= >real/setup.cso I could get further in the
boot.
= > Left some
= > debugging printks in thehandle_real_interrupt
= > handler too.
= >
= > cvs update -r1.25arch/parisc/Makefile
= > cvs update -Aarch/parisc/Makefile
= > cvs update -r1.28arch/parisc/kernel/Makefile
= > cvs update -Aarch/parisc/kernel/Makefile
= > cvs update -r1.35arch/parisc/kernel/head.S
= > cvs update -Aarch/parisc/kernel/head.S
= > cvs update -r1.39arch/parisc/kernel/irq.c
= > cvs update -Aarch/parisc/kernel/irq.c
= > cvs update -r1.14arch/parisc/kernel/pdc.c
= > cvs update -Aarch/parisc/kernel/pdc.c
= > cvs update -r1.25arch/parisc/kernel/pdc_cons.c
= > cvs update -Aarch/parisc/kernel/pdc_cons.c
= > cvs update -r1.73arch/parisc/kernel/setup.c
= > cvs update -Aarch/parisc/kernel/setup.c
= > cvs update -r1.11arch/parisc/kernel/traps.c
= > cvs update -Aarch/parisc/kernel/traps.c
= > cvs update -r1.2arch/parisc/real/Makefile
= > cvs update -Aarch/parisc/real/Makefile
= > cvs update -r1.2arch/parisc/real/head.S
= > cvs update -Aarch/parisc/real/head.S
= > cvs update -r1.2arch/parisc/real/interruption.S
= > cvs update -Aarch/parisc/real/interruption.S
= > cvs update -r1.2arch/parisc/real/lib.c
= > cvs update -Aarch/parisc/real/lib.c
= > cvs update -r1.2arch/parisc/real/pdc.c
= > cvs update -Aarch/parisc/real/pdc.c
= > cvs update -r1.2arch/parisc/real/real.h
= > cvs update -Aarch/parisc/real/real.h
= > cvs update -r1.2arch/parisc/real/real_s.S
= > cvs update -Aarch/parisc/real/real_s.S
= > cvs update -r1.2arch/parisc/real/setup.c
Quoted 9 lines
= > cvs update -A
arch/parisc/real/setup.c
= > cvs update -r1.1arch/parisc/real/traps.c
= > cvs update -Aarch/parisc/real/traps.c
= > cvs update -r1.1include/asm-parisc/real.h
= > cvs update -Ainclude/asm-parisc/real.h
= >
= >
=
---------------------------------------------------------------------------
Quoted 12 lines
= > To unsubscribe: send e-mail to
= > [email protected] with = > `unsubscribe' as the subject.
= >
= >
=
= __________________________________________________ = Do You Yahoo!?
= Thousands of Stores. Millions of Products. All in one place.
= Yahoo! Shopping: http://shopping.yahoo.com
__________________________________________________
Do You Yahoo!?
Thousands of Stores. Millions of Products. All in one place. Yahoo! Shopping: http://shopping.yahoo.com
Quoted 2 lines
So I thought
0xc004a078is inget_unused_buffer_headwhich is infs/buffer.c.
I think the bug you are seeing is due to the inline definition for xchg in asm/system.h. The problem occurs in exchanges with 0 when the address used in the ldcws instruction is not aligned on a 16 byte boundary. Change the define for xchg to just use the spinlock method in all cases.
The other option is to align all the addresses that are used in xchg's with 0.
Dave
Signature · 2 lines
J. David Anglin [email protected] National Research Council of Canada (613) 990-0752 (FAX: 952-6605)