changelog

eCos/RedBoot for勤研ARM AnywhereII(4510) 含全部源代码

1999-11-04  John Dallaway  <[e-mail]>

	* cdl/hal_arm.cdl:

	Tidy up custom rules.

1999-11-04  John Dallaway  <[e-mail]>

	* cdl/hal_arm.cdl:

	Output custom rule dependency information to .deps files in
	the current directory.

	Dispense with the need to create a 'src' sub-directory.

1999-11-01  Jonathan Larmour  <[e-mail]>

	* src/hal_misc.c (exception_handler): Check for CYGDAT_CYGMON_ENABLE
	rather than just CYGPKG_CYGMON
	Include <pkgconf/cygmon.h> to get this define

1999-10-29 Simon FitzMaurice <[e-mail]>

	* cdl/hal_arm.cdl:

	Use fgrep instead of grep in custom rules.  The latter seems
	to behave weirdly under win32.

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

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

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

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

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

	* src/arm_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]>

	* src/arm_stub.c (cyg_hal_gdb_place_break): Add this to the
	ARM-specific GDB stub code.  See hal/common/.../hal_stub for
	details.

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

	* cdl/hal_arm.cdl:

	Invoke 'sh' version of 'mkdir' explicitly in custom rules.

1999-10-26  Jesper Skov  <[e-mail]>

	* src/vectors.S: Cleaned up LED code a bit.

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

	* src/hal_misc.c (exception_handler): If building for CygMon, defer
	exception processing to it, not here.

	* src/arm_stub.c (__install_breakpoints, __clear_breakpoints): 
	Leave out these [dummy] functions if building CygMon.

	* include/pkgconf/hal_arm.h: Add new configuration option to let
	programs which do not include GDB stubs still process "illegal
	instruction" exceptions.

	* src/vectors.S: Reorder context restore so stack changes are
	protected (interrupts off).

1999-10-26  Jesper Skov  <[e-mail]>

	* src/arm.ld: Added support for merging RODATA sections into DATA
	* include/hal_io.h: Added BE support.
	
	* src/vectors.S (LED): Added definition for PID board.

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

	* cdl/hal_arm.cdl:

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

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

	* src/arm_stub.c (cyg_hal_gdb_running_step):
	(cyg_hal_gdb_break_is_set): Provide dummies for these two for now.
	See changelog for hal/common/ hal_stub and generic_stub for info.

1999-10-21  Jonathan Larmour  <[e-mail]>

	* include/hal_arch.h: Tidy a bit so the breakpoint instruction is only
	defined in one place
	Change thumb breakpoint instruction to 0xbebe

1999-10-18  John Dallaway  <[e-mail]>

	* cdl/hal_arm.cdl:

	Add cdl_option CYGHWR_THUMB.

1999-10-14  Jesper Skov  <[e-mail]>
        From James Ingham <[e-mail]>:
	* src/PKGconf.mak: Make sed invocation Solaris safe.

1999-10-11  Jesper Skov  <[e-mail]>

	* src/vectors.S (reset_vector): Always execute PLATFORM_SETUP1.

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

	* include/pkgconf/hal_arm.h: 
	* src/vectors.S: 
	* src/arm_stub.c: Add thread debug support via Multi-ICE.

1999-10-08  Jonathan Larmour  <[e-mail]>

	* src/arm.ld (SECTION_bss): Make *(.bss*) rather than just .bss, to
	cater for nickc's BSS gc improvements.

1999-10-06  Jesper Skov  <[e-mail]>

	* include/basetype.h: Set endian to BE if required.

1999-10-05  Jesper Skov  <[e-mail]>

	* src/arm.ld: Force output format to BE if required.

1999-09-30  Jesper Skov  <[e-mail]>

	* src/vectors.S: Add thumb_func before thumb code where switching
	so GDB/objdump knows how to disassemble it.
	* include/hal_arch.h: Change thumb breakpoint to undefined
	instruction.
	* src/arm_stub.c (target_thumb_ins): Corrected branch offsets.

1999-09-23  Gary Thomas  <[e-mail]>

	* src/vectors.S (call_exception_handler): Change on 09-19 messed up
	status register wrt breakpoints, single stepping, etc.

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

	* include/hal_arch.h: Remove FIXME about GDB register layout.

	* src/context.S (hal_thread_save_context): Change function prologue
	to match APCS more closely (keeps GDB happier).

1999-09-19  Gary Thomas  <[e-mail]>

	* src/vectors.S (spurious_IRQ): Make sure interrupts are masked
	when switching to IRQ mode.

	* src/hal_misc.c (exception_handler): GDB stubs only take over
	memory exceptions.  This allows proper handling when an application
	has it's own IRQ/FIQ VSR.

1999-09-16  Jesper Skov  <[e-mail]>

	* src/arm.ld: Handle thumb glue sections.

1999-09-16  Jesper Skov  <[e-mail]>

	* src/vectors.S: Always include thumb exception support.

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

	* src/context.S: Don't trash r0 which holds the thread's initial
	argument.

	* include/hal_arch.h: 
	* src/arm_stub.c: 
	Added thumb-aware GDB interrupt functions.
	
1999-09-14  Jesper Skov  <[e-mail]>

	* src/vectors.S (undefined_instruction): Adjusted wrong register.
	* src/arm_stub.c (target_thumb_ins): Set thumb flag when required.
	(target_ins): Added BX handling.

	* src/vectors.S:
	* include/hal_arch.h: 
	* src/hal_mk_defs.c:
	* src/arm_stub.c (__single_step): 
	Determine thumb mode by looking at T bit.

1999-09-10  Jesper Skov  <[e-mail]>

	* src/arm_stub.c (target_thumb_ins): BL is a double instruction.

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

	* src/vectors.S (handle_IRQ_or_FIQ): Allow spurious interrupts
	to be ignored.  Some platforms (CL7xxx) have non-debounced 
	interrupt sources which can vanish before the hardware can be
	queried.

1999-09-11  Jonathan Larmour  <[e-mail]>

	* src/arm_stub.c (__computeSignal): Synchronise return value with
	what libbsp does and GDB expects
	(__get_trap_number): Don't shift right 8 - it's already the right number

1999-09-10  Jesper Skov  <[e-mail]>

	* src/vectors.S (start): Moved stub vector changeover to after
	stub init.

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

	* include/hal_arch.h (HAL_BREAKPOINT):
	* src/arm_stub.c:
	* src/vectors.S:
	Added thumb stub support (except C-c support).
	
1999-08-24  Gary Thomas  <[e-mail]>

	* src/vectors.S: Fix misc THUMB procedures.

1999-08-24  Hugo Tyson  <[e-mail]>

	* src/vectors.S (handle_IRQ_or_FIQ): Set up the pointer-to-regset
	in r2 for calling the ISR as well as the DSR.  This is needed for
	the cyg_hal_gdb_isr() for those platforms that handle ^C's
	incoming that way - the ISR needs a regset to know where to plant
	a breakpoint to stop the system in the interrupted context.

1999-08-24  Jesper Skov  <[e-mail]>

	* src/vectors.S (LED): Check XBUS setting jumper and behave
	accordingly.

1999-08-23  Jesper Skov  <[e-mail]>

	* src/arm_stub.c 
	(target_ins): Check whether branches execute.
	(ins_will_execute): Fixed HI and LS.

1999-08-13  John Dallaway  <[e-mail]>

	* cdl/hal_arm.cdl: Add path to CYGBLD_LINKER_SCRIPT value.

1999-08-12  John Dallaway  <[e-mail]>

	* cdl/hal_arm.cdl: Define CYGBLD_LINKER_SCRIPT.

1999-08-12  Jesper Skov  <[e-mail]>

	* include/hal_io.h: Include plf_io.h when appropriate.

1999-08-06  Jesper Skov  <[e-mail]>
        CR 101032
        * src/vectors.S (handle_IRQ_or_FIQ): Moved setup of register_frame
        argument to just before the interrupt_end call.

1999-07-28  Hugo Tyson  <[e-mail]>

	* src/vectors.S (_eCos_id): Remove bogus version string.

1999-07-26  Hugo Tyson  <[e-mail]>

	* src/hal_misc.c (hal_default_isr): Call the special platform HAL
	routine to see whether there is a ^C lurking in the serial device
	through a macro, if it's defined.  Problem is that
	CYGDBG_HAL_DEBUG_GDB_CTRLC_SUPPORT *is* defined for platforms that
	do not support it.

1999-07-23  Hugo Tyson  <[e-mail]>

	* src/hal_misc.c (hal_default_isr): If we are including
	CYGDBG_HAL_DEBUG_GDB_CTRLC_SUPPORT, then poll a special platform
	HAL routine to see whether there is a ^C lurking in the serial
	device.  This has to be done from the default ISR so that others
	(which must know about ^Cs is they are still to be handled) can
	install over the top.  Not all platforms support this, so enabling
	CYGDBG_HAL_DEBUG_GDB_CTRLC_SUPPORT will not necessarily work.

1999-07-21  Hugo Tyson  <[e-mail]>

	* src/vectors.S (vectors): Re-organize startup a little to support
	ROMming of the image and stubrom creation.  Withdraw publication
	of cyg_startup_stack; it's necessary to do this a different way,
	using...

	* include/hal_intr.h (hal_dram_size): reserve location 0x40 for
	communicating DRAM size between boot ROMs and app during startup,
	and remembering it within ROM startup.  Platforms that don't need
	this should zero the location.

1999-07-19  Hugo Tyson  <[e-mail]>

	* src/vectors.S (cyg_startup_stack): Publish this object for
	passing some DRAM size info around during startup.
	Remove the unused [well, ignored] array hal_vsr_table - the VSR
	table is defined to be at 0x20 by the code here in vectors.S

	* include/hal_intr.h (HAL_VSR_SET): define hal_vsr_table to be at
	0x20 where the hardware vectors actually read VSR entries from.
	Renumber VSR symbols to match.  Do not use the external
	"hal_vsr_table" symbol.

1999-07-15  Hugo Tyson  <[e-mail]>

	* src/vectors.S (reset_vector): Invalidate and disable caches for
	EBSA-285 startup is now in PLATFORM_SETUP1 as intended.  Cache
	enabling is in hal_hardware_init; so, same results as before, but
	tidier.

1999-07-14  Hugo Tyson  <[e-mail]>

	* src/vectors.S: Add some startup debug via a LED macro - which
	has different content depending on platform, of course.
	For the EBSA-285, flush caches when writing vectors and the like,
	since RAM startup occurs with caches enabled, otherwise you get
	mysterious traps as CygMon is entered regardless at the first
	clock tick, for example.  This may be taken out again as ROM
	startup is dealt with via PLATFORM_SETUP1, but for now...

1999-06-22  Gary Thomas  <[e-mail]>

	* src/vectors.S: Fix names of 'report' functions (debug dumps
	for various exceptions).  These needed changes for two reasons:
	name pollution and naming clashes after new ABI adopted.
	Improve support for FIQ interrupts.

1999-06-16  Gary Thomas  <[e-mail]>

	* src/vectors.S: Add ".pool" directives which cause assembler 
	to dump it's literal pool early.  This is necessary because
	of the limited addressing available in THUMB mode.

1999-06-15  Gary Thomas  <[e-mail]>

	* src/vectors.S (vectors): Some boards have an MMU, but the
	reset vector needs to have an unmapped address!

1999-06-09  Gary Thomas  <[e-mail]>

	* src/vectors.S (reset_vector): Streamline ROM/STUBS code.

1999-05-25  Gary Thomas  <[e-mail]>

	* src/vectors.S (_eCos_id): Add version string.  Useful for
	tracking when a system was configured (especially ROMs).
	Add initial pure-thumb support - only for kernel+application
	in thumb mode.
	
	* include/hal_intr.h: Update interrupt support macros for
	thumb mode system.

	* src/context.S: Add initial pure-thumb mode support.

1999-05-21  Hugo Tyson  <[e-mail]>

	* include/hal_intr.h: Define HAL_INTERRUPT_STACK_BASE and
	HAL_INTERRUPT_STACK_TOP so that stack usage macros in
	kernel/.../stackmon.hxx can work.

	* src/vectors.S (cyg_interrupt_stack_base): Define this symbol for
	the interrupt stack and its friend for the stack top so that we
	can publish them with nice names.

1999-05-19  Gary Thomas  <[e-mail]>

	* src/arm.ld: Change symbols to match new no-underscore ABI.

1999-05-19  Gary Thomas  <[e-mail]>

	* include/hal_intr.h: Define interrupt support functions for 
	Thumb mode.

1999-05-16  Gary Thomas  <[e-mail]>

	* include/hal_intr.h (HAL_INTERRUPT_STACK_CALL_PENDING_DSRS): 
	Add macro for new DSR handling mechanism (was override of a
	"weak" symbol in kernel).	

1999-05-13  Hugo Tyson  <[e-mail]>

	* include/hal_arch.h (HAL_THREAD_INIT_CONTEXT): Align stack
	(rather conservatively) before use.

1999-05-11  Gary Thomas  <[e-mail]>

	* src/vectors.S: No separate interrupt stack if no kernel.

1999-05-10  Gary Thomas  <[e-mail]>

	* src/vectors.S: 
	* src/hal_mk_defs.c: 
	* include/hal_arch.h: Fix up machine states, initialize stacks.

1999-04-28  Bart Veer  <[e-mail]>

	* src/PKGconf.mak:
	Remove the -n argument to tail, it does not appear to be required
	on any supported host and causes problems with some
	implementations of tail.

1999-04-25  Gary Thomas  <[e-mail]>