On 10/20/2015 04:13 PM, Pierre-Marie de Rodat wrote:
Sorry about the slow review on these patches. In future please feel
free to ping me as often as once a week.
+ /* DWARF operations all work on signed integers.
Note that this will no longer be the case in DWARF 5, where stack
elements have associated types. Let's add a comment about that so
perhaps we can remove the workaround in a future release.
+ /* ??? Set of all DW_OP_nop operations we remove: is it really a good thing
+ to free them, or should we instead let the garbage collect do it? */
Might as well free them if we know they're garbage, it lets us reuse
that memory sooner.
+ /* Trailing nops from loc_descritor_from_tree (if any) cannot be removed
missing 'p' in loc_descriptor_from_tree.
+ /* When translating a function into a DWARF procedure, contains the frame
+ offset *before* evaluating this operation. It is -1 when not yet
+ initialized. */
+ int dw_loc_frame_offset;
+ /* For DW_OP_call* operations: contains the number of stack slots that were
+ added overall when returning from the procedure (so it's negative if the
+ procedure removes slots). */
+ int dw_loc_frame_offset_increment;
I'm not excited about adding another couple of words to every loc insn
for uses that occur so rarely.
dw_loc_frame_offset seems to be primarily used for checking, can we make
it conditional on ENABLE_CHECKING and use a hash_set to remember already
visited nodes?
Instead of dw_loc_frame_offset_increment, can we look up the number of
arguments from the callee?
+ /* Return a type to used in the debug info instead of TYPE, or NULL_TREE to
"to use"
+ /* Arbitrary scale factors cannot be describe in standard DWARF,
"described"
Jason
