changelog

基于ecos的redboot

	Use wildcard in the custom rule for extras.o to
	accommodate 'make -n'.

1999-11-23  Simon Fitzmaurice <[e-mail]>
	* cdl\common.cdl: Correct example of illiteracy

1999-11-20  Gary Thomas  <[e-mail]>
	
	* cdl/common.cdl: Add MMU tables options.

	* include/pkgconf/hal.h (CYGSEM_HAL_INSTALL_MMU_TABLES): 
	(CYGSEM_HAL_STATIC_MMU_TABLES): New options to control MMU table
	flavour and placement.

1999-11-15  Jonathan Larmour  <[e-mail]>

	* include/dbg-threads-api.h, include/generic-stub.h,
	include/hal_stub.h: Add comments to explain these header files should
	not be included by user programs

1999-11-09  Gary Thomas  <[e-mail]>

	* cdl/common.cdl: Add CYGDBG_HAL_DIAG_DISABLE_GDB_PROTOCOL.
	
	* include/pkgconf/hal.h (CYGDBG_HAL_DIAG_DISABLE_GDB_PROTOCOL): 
	New option which [explicitly] controls use of GDB protocol
	for diagnostic I/O.

1999-11-03  Jesper Skov  <[e-mail]>

	* cdl/hal.cdl: More crud out + mn10300 fixes.

1999-11-02  Jesper Skov  <[e-mail]>

	* cdl/hal.cdl: Cleaned out some of the CPU config entries.

1999-10-29  Gary Thomas  <[e-mail]>

	* tests/PKGconf.mak: Add cache tests for Cirrus Logic boards.

1999-10-29  Jesper Skov  <[e-mail]>

	* tests/cache.c: Made safe for targets with no cache.

	* tests/PKGconf.mak: Don't build cache test for ARM unless
	explicitly enabled.

1999-10-28  Jesper Skov  <[e-mail]>

	* tests/cache.c: Fix array size.

1999-10-27  Gary Thomas  <[e-mail]>

	* src/hal_stub.c: Fix typo (CYGDAT_CYGMON_ENABLE).  Also need
	to include <pkgconf/cygmon.h>

1999-10-27  Gary Thomas  <[e-mail]>

	* src/hal_stub.c: Add explicit enable for CygMon package.

1999-10-27  Hugo Tyson  <[e-mail]>

	* src/hal_stub.c (cyg_hal_gdb_interrupt): Bugfix to previous: ^Cs
	were being ignored when received by the diag output code, because
	there was already a break in place.  Solution is to let
	cyg_hal_gdb_interrupt() override any previous break, and tidy up
	state of course.

1999-10-26  Hugo Tyson  <[e-mail]>

	* include/hal_stub.h (CYG_HAL_GDB_ENTER_CRITICAL_IO_REGION): Place
	the break at the address of a label within the calling context
	[normally hal_diag_write_char()]; this is neater and more
	portable; it's supported for more targets.  Also use
	cyg_hal_place_break() to place it, define that, &c.

	* src/hal_stub.c (cyg_hal_gdb_place_break): New routine to
	specially handle the inline breakpoint.  It basically does the
	same as cyg_hal_gdb_interrupt(), except on some platforms where
	you must do different things to set a bp from a non-interrupt
	context.

1999-10-25  Gary Thomas  <[e-mail]>

	* src/hal_stub.c (putDebugChar, getDebugChar): When building with
	CygMon, these functions are defined by CygMon.

1999-10-26  John Dallaway  <[e-mail]>

	* cdl/hal.cdl:

	Modify custom make rule to specify file paths relative
	to the package version directory.

1999-10-22  Hugo Tyson  <[e-mail]>

	* include/hal_stub.h (CYG_HAL_GDB_ENTER_CRITICAL_IO_REGION):
	(CYG_HAL_GDB_LEAVE_CRITICAL_IO_REGION): new macros to be invoked
	in hal_diag routines to make atomic the code that outputs a
	complete GDB $O packet.  This is required so that watching
	variables (which is implemented by stepping) and stepping itself
	does not try to ss through that $O output code.  CASE 102327.

	* src/hal_stub.c (cyg_hal_gdb_break_is_set): new routine to query
	whether a break is set: used by step in the stubs to decide to
	continue instead, because we are in a critical region.
	(cyg_hal_gdb_running_step): communication between generic stub and
	here: if set, a step has been converted into a continue, so we do
	NOT obey cyg_hal_gdb_remove_break until it is cleared, in
	handle_exception_cleanup().

	* src/generic-stub.c (__process_packet): If there is a break set
	by the hal_stub.c, convert a step to continue, and set
	cyg_hal_gdb_running_step to say we did so.

1999-10-22  Jesper Skov  <[e-mail]>
        Case 102379
	* src/generic-stub.c (__handle_exception): Flush and clear caches
	after cleaning up after single-stepping.

	(memcpy, memset): Added to prevent stub hanging if user puts
	breakpoints in either function.

1999-10-19  Nick Garnett  <[e-mail]>

	* include/pkgconf/hal.h: Added monitor configury for Brother
	board.
	
1999-10-10  Gary Thomas  <[e-mail]>

	* src/thread-packets.c (stub_pkt_currthread): Fix typo in comment.

	* src/generic-stub.c: Reorg to expose register get/set functions.

1999-10-08  Simon Fitzmaurice <[e-mail]>
	* cdl\interrupts.cdl: Quoted '[' and ']' (significant in TCL and thus CDL)

1999-10-08  Jesper Skov  <[e-mail]>

	* cdl/interrupts.cdl: Added
	CYGIMP_HAL_COMMON_INTERRUPTS_IGNORE_SPURIOUS.

1999-09-22  Simon FitzMaurice <[e-mail]>

	* cdl\hal.cdl : add priority field to "extras.o" rule

1999-09-21  Gary Thomas  <[e-mail]>

	* src/thread-packets.c: Fix some typos in comments.

1999-09-15  Jesper Skov  <[e-mail]>

	* src/hal_stub.c: Override generic GDB interrupt functions for
	ARM. Need them to be thumb aware. 

1999-09-13  Gary Thomas  <[e-mail]>

	* include/pkgconf/hal.h: Add new config option to allow spurious
	interrupts to be ignored by the HAL.

	* src/generic-stub.c (__process_packet): Build in version string
	(date/time) for stubs in ROM to be returned via 'd' packet.

1999-09-09  Jonathan Larmour  <[e-mail]>

	* include/pkgconf/hal.h (CYGDBG_HAL_DEBUG_GDB_INCLUDE_STUBS): Rename
	CYGPKG_HAL_MN10300_SIM to CYGPKG_HAL_MN10300_AM31_SIM

1999-09-07  Nick Garnett  <[e-mail]>

	* src/dbg-threads-syscall.c (dbg_currthread_id): Fixed to return
	zero if dbg_currthread() returns false. Previously it could have
	returned a random value which would have made GDB think there had
	been a thread switch when there hadn't. It was particularly prone
	to do this in configurations where the kernel was absent.

1999-09-01  Gary Thomas  <[e-mail]>

	* include/pkgconf/hal.h: Always include target and platform package
	include (pkgconf) files.

1999-08-26  Jonathan Larmour  <[e-mail]>

	* src/generic-stub.c (__hex2mem_helper):
	Use target_register_t when casting to address type, rather than long
	Since the compiler can choose arbitrary alignment of unions, force
	to use byte array only
	Restore use of loop termination test for *hexMemSrc, just in case
	since that's how it used to be
	Increment destination, not source at end of loop
	
	(__mem2hex_helper):
	Use target_register_t when casting to address type, rather than long.
	Since the compiler can choose arbitrary alignment of unions, force
	to use byte array only. Therefore use single __read_mem_safe() call
	
	* src/hal_stub.c (__do_copy_mem): Reassign back from dst/src at start
	of each	"if" clause in case we go round the loop again.

1999-08-24  Gary Thomas  <[e-mail]>

	* src/hal_stub.c (__do_copy_mem): Move temp variables to top of
	function - potentially safer if there is an error.

1999-08-23  Jonathan Larmour  <[e-mail]>

	* src/generic-stub.c (__process_packet): For 'g' and 'G' packets,
	when the register size differs from sizeof(target_register_t) we must
	adjust according to the endianness 

	* include/hal_stub.h: Ensure __LITTLE_ENDIAN__ is defined on
	little-endian targets

	* src/hal_stub.c (__build_t_packet): Use the correct register sizes
	for PC and SP rather than assuming they are the same as target_register_t

1999-08-23  Gary Thomas  <[e-mail]>

	* include/generic-stub.h: Change prototypes for internal functions.

	* src/hal_stub.c (__do_copy_mem): 
	* src/generic-stub.c (__mem2hex_helper): Rework to perform aligned,
	multi-byte operations when conditions allow.  This should handle
	most "hardware register" access cases.  Note: there currently is
	no good way to force GDB to require/use such accesses.

1999-08-17  Jonathan Larmour  <[e-mail]>

	* tests/intr.c (cyg_start): Use CYG_TEST_NA rather than a pass saying
	it's inapplicable

1999-08-16  Jonathan Larmour  <[e-mail]>

	* include/pkgconf/hal.h: Rename CYG_HAL_MN10300_AM32_STDEVAL1 to
	CYG_HAL_MN10300_AM31_STDEVAL1

1999-08-13  Nick Garnett  <[e-mail]>

	* src/hal_stub.c: Removed superfluous trace() extern.

1999-08-12  Nick Garnett  <[e-mail]>	

	Imported following changes from AM33 branch:
	
    1999-07-27  Nick Garnett  <[e-mail]>

	* src/hal_stub.c: Added support for 1 byte breakpoints.
	Removed	invalidates for cache flushes - I am not entirely sure why
	we have these here, since HAL_DCACHE_SYNC() should do what we
	need, and these invalidates are causing a problem on the MIPS and
	MN10300 targets.

	* include/pkgconf/hal.h: Added Monitor selection for AM33.

    1999-06-29  Nick Garnett  <[e-mail]>

	* include/pkgconf/hal.h: Modified to match new MN10300 HAL
	organization.

1999-08-06  Jesper Skov  <[e-mail]>

	* src/generic-stub.c (__process_packet): FLush and clear caches
	after call to __single_step.

1999-07-05  Jesper Skov  <[e-mail]>

	* tests/intr.c: Don't use kernel's RTC definitions.

1999-07-02  Jesper Skov  <[e-mail]>
        PR20157
	* tests/cache.c: Print stride values, preventing timeout on slow
	targets.

1999-06-28  Jesper Skov  <[e-mail]>

	* include/pkgconf/hal.h (CYGDBG_HAL_DEBUG_GDB_BREAK_SUPPORT):
	Fixed parent.

1999-06-24  Jonathan Larmour  <[e-mail]>

	* include/pkgconf/hal.h: Clarify descriptions of
	CYGDBG_HAL_DEBUG_GDB_BREAK_SUPPORT and
	CYGDBG_HAL_DEBUG_GDB_CTRLC_SUPPORT.

1999-06-17  Nick Garnett  <[e-mail]>

	* include/pkgconf/hal.h: Added CYGDBG_HAL_DEBUG_GDB_CTRLC_SUPPORT
	to enable inclusion of ^C interrupt handler. Added CDL to control
	it.

1999-06-10  Hugo Tyson  <[e-mail]>

	* include/pkgconf/hal.h: Add support for the PowerPC target MBX860
	platform, CYG_HAL_POWERPC_MBX.

1999-06-10  Gary Thomas  <[e-mail]>

	* src/hal_stub.c: Need to init register pointer.

1999-06-08  Jesper Skov  <[e-mail]>

	* include/pkgconf/hal.h: Added CYG_HAL_ROM_MONITOR for EDK7708.

1999-05-31  Jesper Skov  <[e-mail]>

	* src/hal_stub.c: Respect HAL_BREAKINST_SIZE setting.

1999-05-28  Jesper Skov  <[e-mail]>

	* include/pkgconf/hal.h: Renamed SH platform package to edk7708.

1999-05-28  Nick Garnett  <[e-mail]>

	* src/hal_stub.c (initHardware): Added call to
	HAL_STUB_PLATFORM_INIT if it is defined.

	* include/pkgconf/hal.h: Added a #undef for CYG_HAL_ROM_MONITOR,
	largely so that a stubrom permutation can define it.

1999-05-27  Nick Garnett  <[e-mail]>

	* src/hal_stub.c: Do not call HAL_ICACHE_INVALIDATE_ALL() and
	HAL_DCACHE_INVALIDATE_ALL() in __instruction_cache() and
	__data_cache() for MIPS targets. These macros are simply
	duplicates.

1999-05-27  Jesper Skov  <[e-mail]>

	* tests/cache.c:  Added handling of unified caches.

1999-05-25  Jonathan Larmour  <[e-mail]>

	* tests/cache.c: Use HAL_DCACHE_SYNC if available

1999-05-24  Jesper Skov  <[e-mail]>

	* include/pkgconf/hal.h: Added CPU specifier for EDK board.

1999-05-21  Jesper Skov  <[e-mail]>

	* src/hal_stub.c (__install_traps): Set exit vector to __reset.

1999-05-21  Nick Garnett  <[e-mail]>

	* include/pkgconf/hal.h: Switched VR4300 over to using GDB STUBS
	monitor from PMON.
	
1999-05-21  Jesper Skov  <[e-mail]>

	* tests/intr.c: Get RTC values from CYGBLD_HAL_PLATFORM_H.
	Ensure there are fallback definitions.

	* src/hal_stub.c (handle_exception_cleanup): Added optional call
	of HAL_STUB_PLATFORM_STUBS_FIXUP macro.

1999-05-18  Jesper Skov  <[e-mail]>

	* src/hal_stub.c (hal_output_gdb_string): Added missing ;

1999-05-13  Nick Garnett  <[e-mail]>

	The following changes were all made on a branch and imported into
	the main trunk later.
	
    1999-05-11  Nick Garnett  <[e-mail]>

	* include/pkgconf/hal.h:
	tidied up all MIPS targets to include CYGBLD_HAL_TARGET_H. This
	will eventually apply to all configurations.

    1999-05-06  Nick Garnett  <[e-mail]>

	* src/PKGconf.mak (COMPILE): Added dbg-threads-syscall.c.

	* include/dbg-thread-syscall.h: 
	* src/dbg-threads-syscall.c:
	Files added to enable support for thread-awareness in GDB stub
	ROMs. The .h file has been moved from kernel/src/debug.

    1999-04-28  Nick Garnett  <[e-mail]>

	* src/hal_stub.c (__build_t_packet): Changed sp variable to a
	non-pointer type to force it to be installed in the T packet with
	the right size.

	* src/drv_api.c: Added cyg_interrupt_call_pending_DSRs() to keep
	HAL happy in non-kernel configurations.

    1999-04-21  Nick Garnett  <[e-mail]>

	* include/pkgconf/hal.h:
	Added VR4300 configury for ROM monitor use and inclusion of
	variant-specific configuration header. This needs some further
	tidying up.
	
1999-04-28  Bart Veer  <[e-mail]>

	* src/PKGconf.mak:
	Use the new rules for generating libextras.a

1999-04-27  Jesper Skov  <[e-mail]>

	* include/pkgconf/hal.h: GDB_BREAK support now available for ARM
	targets.

1999-04-16  Bart Veer  <[e-mail]>

	* src/PKGconf.mak: 
	* src/dummy.c:
	New file that can always be added to libextras.a without
	affecting image sizes.

1999-04-14  Jonathan Larmour  <[e-mail]>

	* include/hal_stub.h: 
	Make sure all functions have C linkage even in C++

1999-04-14  Jesper Skov  <[e-mail]>
        PR 19845
	* src/generic-stub.c (__process_packet): Reset _register pointer
	to the exception state before attempting to set up single-step
	magic.

1999-04-14  Jesper Skov  <[e-mail]>

	* include/hal_stub.h: Added declaration for __stub_copy_registers.

1999-04-14  Jesper Skov  <[e-mail]>

	* src/hal_stub.c: Use HAL_xCACHE_IS_ENABLED if defined.

1999-04-09  Jesper Skov  <[e-mail]>

	* src/hal_stub.c:
	Moved get_register and put_register here. Made them access the
	currently selected register set.

1999-03-22  Jonathan Larmour  <[e-mail]>