mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [OOPS] in kdb v2.3 on top of 2.5.44
@ 2002-10-20 18:46 David Mansfield
  2002-10-22  8:34 ` Keith Owens
  0 siblings, 1 reply; 2+ messages in thread
From: David Mansfield @ 2002-10-20 18:46 UTC (permalink / raw)
  To: Keith Owens, linux-kernel


Hi Keith, list,

I tried patching the 2.5.44 vanilla kernel with the 2.3 you released for
2.5.43.  The patching had only a few rejects, which were completely
trivial, but perhaps this mis-match of kernel to patch version caused the
oops, but here it is (starting from me successfully executing a few
commands inside the debugger):

[ snip kernel boot and beginning of kdb session ]

[1]kdb> bt
EBP        EIP        Function (args)
0xc0106ee0 0xc0106f0a default_idle+0x2a (0x0, 0x0, 0x0)
                               kernel .text 0xc0100000 0xc0106ee0 
0xc0106f20
           0xc0106fb2 cpu_idle+0x52
                               kernel .text 0xc0100000 0xc0106f60 
0xc0106fd0
           0xc0452dfd start_secondary+0x6d
                               kernel .init.text 0xc044c000 0xc0452d90 
0xc0452e00
[1]kdb> go

[ time passes here while I use the system a bit - then I do ctrl-a again ]

Entering kdb (current=0xc0420e20, pid 0) on processor 0 due to Keyboard 
Entry
Unable to handle kernel NULL pointer dereference at virtual address 
00000000
 printing eip:
c02b77c4
*pde = 00000000
Oops: 0002
 
CPU:    0
EIP:    0060:[<c02b77c4>]    Not tainted
EFLAGS: 00010006
EIP is at kdba_setjmp+0x4/0x50
eax: 00000000   ebx: c044a000   ecx: 00000001   edx: 00000000
esi: 00000004   edi: c034409f   ebp: c044a000   esp: c044be1c
ds: 0068   es: 0068   ss: 0068
Process swapper (pid: 0, threadinfo=c044a000 task=c0420e20)
Stack: c02084ec c04d4c40 00000000 00000000 00000000 00000000 00000000 
00000002 
       ae50ded0 1750ded0 00000000 00000000 c036952e ffffffff c0369532 
00000000 
       c02b7229 c036952e c036952e c044a000 c044a000 00000008 00000000 
c020872c 
Call Trace:
 [<c02084ec>] kdb_local+0x2cc/0x3e0
 [<c02b7229>] kdba_getregcontents+0x119/0x280
 [<c020872c>] kdb_main_loop+0xbc/0x220
 [<c02b76e4>] kdba_main_loop+0x54/0x60
 [<c0208e49>] kdb+0x5b9/0x7e0
 [<c0233c4c>] receive_chars+0x9c/0x290
 [<c0234098>] serial8250_interrupt+0x88/0x120
 [<c010ae9a>] handle_IRQ_event+0x3a/0x60
 [<c010b0e6>] do_IRQ+0xc6/0x160
 [<c0106ee0>] default_idle+0x0/0x40
 [<c0106ee0>] default_idle+0x0/0x40
 [<c01098d4>] common_interrupt+0x18/0x20
 [<c0106ee0>] default_idle+0x0/0x40
 [<c0106ee0>] default_idle+0x0/0x40
 [<c0106f0a>] default_idle+0x2a/0x40
 [<c0106fb2>] cpu_idle+0x52/0x70
 [<c0105000>] stext+0x0/0x60
 [<c0105051>] stext+0x51/0x60

Code: 89 58 00 89 70 04 89 78 08 8b 0c 24 89 48 0c 8d 4c 24 08 89 
 kdb: Debugger re-entered on cpu 0, new reason = 5
     Not executing a kdb command
     No longjmp available for recovery
     Cannot recover, allowing event to proceed
<0>Kernel panic: Aiee, killing interrupt handler!
In interrupt handler - not syncing
 
My system is a dual PII-450 with 1GB ram (highmem 4g) running all SCSI
with an adaptec aic7xxx.  System is redhat 7.3 based with official
updates.

David

-- 
/==============================\
| David Mansfield              |
| lkml@dm.cobite.com           |
\==============================/


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [OOPS] in kdb v2.3 on top of 2.5.44
  2002-10-20 18:46 [OOPS] in kdb v2.3 on top of 2.5.44 David Mansfield
@ 2002-10-22  8:34 ` Keith Owens
  0 siblings, 0 replies; 2+ messages in thread
From: Keith Owens @ 2002-10-22  8:34 UTC (permalink / raw)
  To: David Mansfield; +Cc: linux-kernel, kdb

On Sun, 20 Oct 2002 14:46:05 -0400 (EDT), 
David Mansfield <lkml@dm.cobite.com> wrote:
>EIP is at kdba_setjmp+0x4/0x50
>eax: 00000000   ebx: c044a000   ecx: 00000001   edx: 00000000
>esi: 00000004   edi: c034409f   ebp: c044a000   esp: c044be1c

A kbuild change in 2.5.41 moved the initial setting of CFLAGS before
the include of .config.  Now the global CFLAGS must not depend on any
config settings (totally undocumented, of course).  kdba_setjmp was
compiled with CONFIG_FRAME_POINTERS=y from config.h, but gcc flags were
set to -fomit-frame-pointer, the mismatch generated bad code.

Will be fixed in next kdb patch, in the meantime, apply this over
kdb-v2.3-2.5.53-common-1.

Index: 43.2/Makefile
--- 43.2/Makefile Thu, 17 Oct 2002 15:02:15 +1000 kaos (linux-2.5/I/d/12_Makefile 1.41.1.10 444)
+++ 43.2(w)/Makefile Tue, 22 Oct 2002 18:17:22 +1000 kaos (linux-2.5/I/d/12_Makefile 1.41.1.10 444)
@@ -159,9 +159,6 @@ CPPFLAGS	:= -D__KERNEL__ -Iinclude
 CPPFLAGS	+= $(patsubst %,-I%,$(CROSS_COMPILE_INC))
 CFLAGS 		:= $(CPPFLAGS) -Wall -Wstrict-prototypes -Wno-trigraphs -O2 \
 	  	   -fno-strict-aliasing -fno-common
-ifndef CONFIG_FRAME_POINTER
-CFLAGS		+= -fomit-frame-pointer
-endif
 AFLAGS		:= -D__ASSEMBLY__ $(CPPFLAGS)
 
 export	VERSION PATCHLEVEL SUBLEVEL EXTRAVERSION KERNELRELEASE ARCH \
@@ -212,6 +209,10 @@ include-config := 1
 
 endif
 
+ifndef CONFIG_FRAME_POINTER
+CFLAGS		+= -fomit-frame-pointer
+endif
+
 include arch/$(ARCH)/Makefile
 
 core-y		+= kernel/ mm/ fs/ ipc/ security/


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2002-10-22  8:28 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-10-20 18:46 [OOPS] in kdb v2.3 on top of 2.5.44 David Mansfield
2002-10-22  8:34 ` Keith Owens

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®