Skip to contentSkip to conversation

PATCH m68knommu clear frame-pointer in start_thread

DISCUSSION Philippe De Muyter 3 participants 3 messages Mar 15, 2006 – Mar 16, 2006

lore
Share & tools
lore
#0 lore raw

When trying to print the calltrace of a user process on m68knommu targets gdb follows the frame-pointer en falls on unreachable adresses, because the frame pointer is not properly initialised by start_thread. This patch initialises the frame pointer to NULL in start_thread.

---

the only change is the initialisztion of a6, the rest is spacing fixes

Index: linux-2.6.x/include/asm-m68knommu/processor.h
===================================================================
RCS file: /newdev6/cvsrepos/uClinux-dist/linux-2.6.x/include/asm-m68knommu/processor.h,v retrieving revision 1.1.1.1
diff -r1.1.1.1 processor.h
92,99c92,100
< #define start_thread(_regs, _pc, _usp) \
< do { \
< set_fs(USER_DS); /* reads from user space */ \
< (_regs)->pc = (_pc); \
< if (current->mm) \
< (_regs)->d5 = current->mm->start_data; \
< (_regs)->sr &= ~0x2000; \
< wrusp(_usp); \
---

Quoted 9 lines

#define start_thread(_regs, _pc, _usp) \

do {							\
	set_fs(USER_DS); /* reads from user space */	\
	(_regs)->pc = (_pc);				\
	((struct switch_stack *)(_regs))[-1].a6 = 0;	\
	if (current->mm)				\
		(_regs)->d5 = current->mm->start_data;	\
	(_regs)->sr &= ~0x2000;				\
	wrusp(_usp);					\
#1 lore raw
Quoted 6 lines · Philippe De Muyter

When trying to print the calltrace of a user process on m68knommu targets gdb follows the frame-pointer en falls on unreachable adresses, because the frame pointer is not properly initialised by start_thread. This patch initialises the frame pointer to NULL in start_thread.

You've posted an ed-style diff. Quite a safe method to make sure nobody will read it. Pleae repost a unified diff.

Ralf

#2 lore raw
Quoted 7 lines · Ralf Baechle
Quoted 5 lines · Philippe De Muyter

When trying to print the calltrace of a user process on m68knommu targets gdb follows the frame-pointer en falls on unreachable adresses, because the frame pointer is not properly initialised by start_thread. This patch initialises the frame pointer to NULL in start_thread.

Quoted 6 lines
Quoted 1 line

You've posted an ed-style diff. Quite a safe method to make sure nobody will read it. Pleae repost a unified diff.

Ralf

Oops

Thanks for your attention and your reply

I will resend it.

Philippe

Signature · 2 lines
Philippe De Muyter  phdm at macqel dot be  Tel +32 27029044
Macq Electronique SA  rue de l'Aeronef 2  B-1140 Bruxelles  Fax +32 27029077