| Subject: | Re: [CFT] re-organize var-tracking frame_base |
|---|---|
| From: | Richard Henderson |
| Date: | Sat, 20 Aug 2005 19:25:43 -0700 |
On Fri, Aug 19, 2005 at 07:15:49PM -0400, David S. Miller wrote: > When GDB sees a DW_CFA_GNU_window_save, it records that > registers I0->I7 and L0->L7 are saved at: > > CFA + (reg_size * reg_slot_nr) > > It also records that the O0->O7 registers have become > the I0->I7 registers. Which is irrelevant in this case. The important part is that the window_save opcode does not affect the cfa. I got around to building a cross compiler and looking at the output; it seems plausible. You should give it a run through gdb to but sure though. r~ |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: PATCH: PR target/23485: [ia64]: Integer dvide by zero doesn't raise a signal, Richard Henderson |
|---|---|
| Next by Date: | Re: PATCH: PR target/23485: [ia64]: Integer dvide by zero doesn't raise a signal, H. J. Lu |
| Previous by Thread: | Re: [CFT] re-organize var-tracking frame_base, David S. Miller |
| Next by Thread: | Re: [RFC patch]: Skip EH_RETURN_STACKADJ_RTX assignments in create_pre_exit, Richard Henderson |
| Indexes: | [Date] [Thread] [Top] [All Lists] |