* NEWS: Mention ARM VFP support.

	* target-descriptions.c (tdesc_register_type): Make public.
	(tdesc_unnumbered_register): New function.
	(tdesc_register_reggroup_p): Allow missing
	pseudo_register_reggroup_p.
	* target-descriptions.h (tdesc_register_type): Declare.
	(tdesc_unnumbered_register): Declare.
	* arm-tdep.c (arm_neon_quad_read, arm_neon_quad_write): New functions.
	(arm_push_dummy_call): Use arm_neon_quad_write.
	(arm_neon_double_type, arm_neon_quad_type): New functions.
	(arm_register_type): Handle VFP and NEON registers.  Override the
	types of double-precision registers for NEON.  Disable FPA registers
	if they are not present.
	(arm_dwarf_reg_to_regnum): Add current VFP and NEON register numbers.
	(arm_return_value): Use arm_neon_quad_write and arm_neon_quad_read.
	(arm_register_name): Handle VFP single and NEON quad registers.
	(arm_pseudo_read, arm_pseudo_write): New functions.
	(arm_gdbarch_init): Check for VFP and NEON in the target description.
	Assign numbers to double-precision registers.  Register VFP and NEON
	pseudo registers.  Remove a shadowed "i" variable.
	* arm-tdep.h (enum gdb_regnum): Add ARM_D0_REGNUM and
	ARM_D31_REGNUM.
	(struct gdbarch_tdep): Add have_neon_pseudos, have_neon,
	have_vfp_registers, have_vfp_pseudos, neon_double_type,
	and neon_quad_type.

	* features/Makefile: Make expedite settings only architecture
	specific.
	(WHICH): Add new ARM descriptions.
	* features/arm-with-neon.xml, features/arm-with-vfpv2.c,
	features/arm-with-vfpv3.c, features/arm-vfpv2.xml,
	features/arm-vfpv3.xml, features/arm-with-vfpv2.xml,
	features/arm-with-vfpv3.xml, features/arm-with-neon.c: New files.
	* regformats/arm-with-neon.dat, regformats/arm-with-vfpv2.dat,
	regformats/arm-with-vfpv3.dat: Generate.

	doc/
	* gdb.texinfo (ARM Features): Document org.gnu.gdb.arm.vfp and
	org.gnu.gdb.arm.neon.

	gdbserver/
	* linux-low.c (linux_write_memory): Update debugging output.
	* Makefile.in (clean): Add new descriptions.
	(arm-with-vfpv2.o, arm-with-vfpv2.c, arm-with-vfpv3.o)
	(arm-with-vfpv3.c, arm-with-neon.o, arm-with-neon.c): New rules.
	* configure.srv: Add new files for arm*-*-linux*.
	* linux-arm-low.c: Add new declarations.
	(PTRACE_GETVFPREGS, PTRACE_SETVFPREGS): Define if undefined.
	(arm_hwcap, HWCAP_VFP, HWCAP_IWMMXT, HWCAP_NEON, HWCAP_VFPv3)
	(HWCAP_VFPv3D16): New.
	(arm_fill_wmmxregset, arm_store_wmmxregset): Check HWCAP_IWMMXT
	instead of __IWMMXT__.
	(arm_fill_vfpregset, arm_store_vfpregset, arm_get_hwcap)
	(arm_arch_setup): New.
	(target_regsets): Remove #ifdef.  Add VFP regset.
	(the_low_target): Use arm_arch_setup.

	testsuite/
	* gdb.base/float.exp: Handle VFP registers.

2 files changed, 9 insertions, 0 deletions

diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index c2ae2f8..aaaa84e 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,7 @@
+2009-07-28  Daniel Jacobowitz  <dan@codesourcery.com>
+
+	* gdb.base/float.exp: Handle VFP registers.
+
 2009-07-14  Michael Snyder  <msnyder@vmware.com>
 
 	* gdb.reverse/finish-reverse.exp: Do not expect reverse-finish
diff --git a/gdb/testsuite/gdb.base/float.exp b/gdb/testsuite/gdb.base/float.exp
index a14a915..09a02ca 100644
--- a/gdb/testsuite/gdb.base/float.exp
+++ b/gdb/testsuite/gdb.base/float.exp
@@ -51,6 +51,11 @@ if { [istarget "alpha*-*-*"] } then {
 	-re "Software FPU type.*mask:.*flags:.*$gdb_prompt $" {
 	    pass "info float (FPA)"
 	}
+	-re "fpscr.*s0.*s1.*s31.*$gdb_prompt $" {
+	    # Only check for single precision; d0 might be a vector register
+	    # if we have NEON.
+	    pass "info float (VFP)"
+	}
         -re "No floating.point info available for this processor.*" {
             pass "info float (without FPU)"
 	}