Remove some unnecessary inferior_ptid setting/restoring when fetching/storing registers
authorSimon Marchi <simon.marchi@ericsson.com>
Thu, 23 Mar 2017 17:37:06 +0000 (13:37 -0400)
committerSimon Marchi <simon.marchi@ericsson.com>
Thu, 23 Mar 2017 17:37:06 +0000 (13:37 -0400)
commit3e00d44febb8093d8dc0e6842b975afb194c4fd1
tree328c147408e951650ebe8247a794592bcb6e8625
parent077ae656a69fcf144f68848a3f6f2cb559b62987
Remove some unnecessary inferior_ptid setting/restoring when fetching/storing registers

Now that the to_fetch_registers, to_store_registers and
to_prepare_to_store target methods don't rely on the value of
inferior_ptid anymore, we can remove a bunch of now unnecessary setting
and restoring of inferior_ptid.

The asserts added recently in target_fetch_registers and
target_store_registers, which validate that inferior_ptid matches the
regcache's ptid, must go away.  It's the whole point of this effort, to
not require inferior_ptid to have a particular value when calling these
functions.

One thing that I noticed is how sol-thread.c's ps_lgetregs and friends
use the current value of inferior_ptid instead of what's passed as
argument (ph->ptid), unlike proc-service.c's versions of the same
functions.  Is it expected?  I left it like this in the current patch,
but unless there's a good reason for it to be that way, I guess we
should make it use the parameter.

gdb/ChangeLog:

* fbsd-tdep.c (fbsd_corefile_thread): Don't set/restore
inferior_ptid.
* proc-service.c (ps_lgetregs, ps_lsetregs, ps_lgetfpregs,
ps_lsetfpregs): Likewise.
* regcache.c (regcache_raw_update, regcache_raw_write): Likewise.
* sol-thread.c (ps_lgetregs, ps_lsetregs, ps_lgetfpregs,
ps_lsetfpregs): Likewise.
* target.c (target_fetch_registers, target_store_registers):
Remove asserts.
gdb/ChangeLog
gdb/fbsd-tdep.c
gdb/proc-service.c
gdb/regcache.c
gdb/sol-thread.c
gdb/target.c