changelog

开放源码实时操作系统源码.

2004-04-22  Jani Monoses <[e-mail]>

	 * cdl/hal_mips_upd985xx.cdl :
	 Invoke tail with stricter syntax that works in latest coreutils. 

2003-04-10  Nick Garnett  <[e-mail]>

	* src/hal_mips_upd985xx.ld: 
	Added libsupc++.a to GROUP() directive for GCC versions later than
	3.0.

2002-04-02  Hugo Tyson  <[e-mail]>
2002-04-02  Anssi Pulkkinen <[e-mail]>

	* src/var_misc.c (cyg_hal_interrupt_acknowledge): Remove the read
	of the read-clear ISR register - it loses other pending interrupt
	sources.  Thanks to ASCOM for spotting this.  It should have been
	removed when the soft-copy and arbitration ISR were added, because
	those changes mean the hardware register would already be cleared
	down.

2002-01-22  Hugo Tyson  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl: CYGIMP_HAL_COMMON_INTERRUPTS_CHAIN
	must not be enabled because the way the arbitration ISR is
	attached does not work with chained interrupts.

2001-10-30  Jonathan Larmour  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl (CYGHWR_HAL_MIPS_UPD985XX_DIAG_BAUD): 
	This is the same as CYGNUM_HAL_VIRTUAL_VECTOR_CONSOLE_CHANNEL_BAUD
	which RedBoot uses, so define it.
	And also reimplement CYGINT_HAL_VIRTUAL_VECTOR_COMM_BAUD_SUPPORT
	as this is what makes it work.

2001-10-30  Hugo Tyson  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl: Platform does *not* implement VV baud
	operations, in fact the implements
	CYGINT_HAL_VIRTUAL_VECTOR_COMM_BAUD_SUPPORT prevents it building.

2001-10-09  Hugo Tyson  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl (..._MIPS_UPD985XX_HARDWARE_BUGS...):
	New CDL options to control workarounds for System Controller bugs
	S1 and S2.  Shorthand for S1 requires the fixes elsewhere.  That
	for S2 selects alternate versions of the interrupt code.

	* include/var_intr.h (HAL_INTERRUPT_ACKNOWLEDGE,...): Alternative
	versions which call the new routines in src/var_misc.c to deal
	with System Controller interrupts.

	* src/var_misc.c (cyg_hal_interrupt_unmask,...): New routines to
	manage the System Controller interrupts by software rather than
	S_ISR/S_IMR.  We never mask an interrupt once unmasked, but let
	the interrupt happen, fielding it silently.  When/if it becomes
	unmasked, then we call the ISR &c.

2001-08-17  Jonathan Larmour  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl: Platform implements VV baud operations.

2001-08-08  Hugo Tyson  <[e-mail]>

	* include/variant.inc (hal_intc_decode): Do not decode the S_ISR
	interrupts after all - they must be handled by arbitration.  So
	any S_ISR interrupt is reported as CYGNUM_HAL_INTERRUPT_SYSCTL -
	Who are they?  They are *all* number 6.

	* src/var_misc.c (_arbitration_isr): New routine to arbitrate
	between - and call all of - the interrupt sources that hang off
	the system controller S_ISR register, because the S_ISR register
	is read-clear, and the interrupt sources are edge-triggered so
	they do not re-assert themselves - so we must address multiple
	sources per actual interrupt.
	(hal_variant_init): Install _arbitration_isr() on the SYSCTL
	interrupt at startup.

	* include/var_intr.h: Commented the change in interrupt usage -
	the SYSCTL is occupied from time zero.

2001-08-01  Hugo Tyson  <[e-mail]>

	* src/hal_mips_upd985xx.ld (hal_interrupt_sr_mask_shadow_base):
	This must be placed statically so it is shared with RedBoot.  And
	it must be an array so that we can address it from afar.

	* include/var_intr.h (HAL_DISABLE_INTERRUPTS): Implement this
	macro and its fellows as well as the individual MASK/UNMASK
	routines; these work together to keep the SR IM bits "true" to
	what is intended by means of a shadow variable.  Otherwise a race
	condition in the vanilla HAL_DISABLE_INTERRUPTS() can discard an
	interrupt mask change made by an ISR.
	(hal_interrupt_sr_mask_shadow): declare this and define it to
	entry zero of hal_interrupt_sr_mask_shadow_base[].
	
	* src/var_misc.c (hal_interrupt_sr_mask_shadow): initialize this
	new variable.

	* cdl/hal_mips_upd985xx.cdl (CYGPKG_HAL_MIPS_UPD985XX): set
	CYGINT_HAL_MIPS_INTERRUPT_RETURN_KEEP_SR_IM because we use bits
	within the SR for interrupt control.

2001-07-25  Hugo Tyson  <[e-mail]>

	* include/var_intr.h (CYGNUM_HAL_INTERRUPT_SOFT_ZERO)
	(CYGNUM_HAL_INTERRUPT_SOFT_ONE): New interrupt numbers for
	software interrupts 0 and 1.  Also mask, unmask, and ack them
	correctly; ack clears the R/W bit in the cause regsiter.

	* include/variant.inc (hal_intc_decode) : Decode Software
	interrupts into numbers 0 and 1 - all else are moved up a bit.

2001-07-20  Hugo Tyson  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl: Demand that CYGPKG_LIBM includes
	-fno-strict-aliasing in its CFLAGS_ADD to workaround a tools issue
	with access to double via pointer-to-union casts.

2001-07-18  Hugo Tyson  <[e-mail]>

	* include/var_arch.h (UARTLCR_8N1): Add more divers definitions of
	UART control bits for implementing all the controls in the serial
	device that nobody ever uses.

2001-07-17  Hugo Tyson  <[e-mail]>

	* include/variant.inc: If RAM startup, don't blow away the
	contents of cache - it might contain things that matter such as
	debug connection state.

2001-07-09  Hugo Tyson  <[e-mail]>

	* include/var_arch.h (SDMDR_INIT): Change one of the numbers to
	match what's in the hardware when set up by customer code.

	* include/variant.inc: Remove dependency on temporary development
	config point CYGBLD_HAL_STARTUP_ROM_POST_OMIT_INITIALIZATION - the
	initialization now works OK.  So hal_memc_setup_table gets called.

	* src/variant.S (hal_memc_setup_table): ROM startup now works, so
	this code now gets called.

2001-07-06  Hugo Tyson  <[e-mail]>

	* src/hal_diag.c (hal_uart_init): Ensure that we use the internal
	baud clock because there is no external one.

	* include/var_arch.h (UARTCLOCK): Clock is now 50MHz for "new"
	boards, not 18.xMHz any more.

2001-07-03  Hugo Tyson  <[e-mail]>

	* include/var_intr.h (HAL_INTERRUPT_[UN]MASK): Better manipulate
	these with interrupts disabled for atomicity.

2001-07-02  Bart Veer  <[e-mail]>

	* src/var_misc.c (hal_variant_init):
	Move the GPIO0 manipulating (flash programming voltagE)
	to the platform-init.

2001-06-27  Hugo Tyson  <[e-mail]>

	This all fixes the serial interrupt problem of 2001-06-26.
	
	* src/var_misc.c (hal_variant_init): Initialize the system
	controller's ISR/IMR registers to mask and ack all external
	interrupts and then enable the underlying interrupt source that
	controls them all.
	Also provide hal_interrupt_handlers[CYGNUM_HAL_ISR_COUNT]; &c for
	use by system in vectors.S, and initialize them.

	* src/hal_diag.c (cyg_hal_plf_serial_isr): Do not paranoically
	acknowledge spurious interrupts.  No need.

	* include/variant.inc: Add complete redefinitions of interrupt
	decoding to handle extended interrupt sources from the system
	controller's ISR/IMR registers.

	* include/var_intr.h: Add complete redefinitions of interrupt
	management to handle extended interrupt sources from the system
	controller's ISR/IMR registers.

	* include/var_arch.h (S_ISR_ADR): Add ISR and IMR register
	addresses for use by assembly files.

	* include/plf_stub.h: Provide proto for cyg_hal_plf_comms_init()
	to reduce warning.

2001-06-26  Hugo Tyson  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl: Remove src/var_stub.c - functionality
	is duplicated in the common HAL.

	* include/plf_stub.h: Remove a load of unneccessary cruft which is
	duplicated in the common HAL.

	* src/var_stub.c: Removed.

2001-06-26  Hugo Tyson  <[e-mail]>

	* src/var_misc.c (hal_variant_init): Unmask the UART interrupt in
	the system controller - this is before the interrupt system in the
	MIPS core which is used for dynamic control.

	Also removed a load of duplicate hal_ctrlc_isr/HAL_CTRLC_ISR stuff
	that fought with the common HAL,

	Asynchronous CTRL-C still does not work.  The interrupt asserts
	forever.  This is similar to System Controller known bug S2
	"Interrupt Mask Restriction" from the NEC docs.  Hence this code
	in this file:	
	// *******************FIXME
	// This causes an interrupt loop from the UART as soon as you do IO.
	//    *S_IMR |= S_ISR_UARTIS; // unmask UART
	// *******************NB we mask it here so that the status is in control
	// *******************   of the application code whatever RedBoot does.
	    *S_IMR &=~S_ISR_UARTIS;
	// *******************FIXME

	* src/hal_diag.c (cyg_hal_plf_serial_isr): This now contains all
	it should need to.
	(cyg_hal_plf_serial_control): Ditto, all cases now supported.

	CYGHWR_HAL_GDB_PORT_VECTOR use now unconditional, other tidyups.
	Bugfix to interrupt ack in nonblocking read, it was after the
	return.
	
	* cdl/hal_mips_upd985xx.cdl: We want to support ^Cm, so do not
	implement CYGINT_HAL_DEBUG_GDB_CTRLC_UNSUPPORTED.

	* include/var_intr.h (HAL_READ_INTR_REGS): New macro, handy for
	debugging MIPS.  Also remove duplicate hal_ctrlc_isr/HAL_CTRLC_ISR
	stuff that fought with the common HAL, just leaving
	CYGHWR_HAL_GDB_PORT_VECTOR defined.

	* src/var_stub.c (hal_var_stub_init): Use symbols for entries in
	vector tables and the like.

2001-06-22  Hugo Tyson  <[e-mail]>

	* include/variant.inc: Use the proper target-agnostic config
	include CYGBLD_HAL_PLATFORM_H, and do not initialize the memory
	controller if you're told not to by the platform configuration
	CYGBLD_HAL_STARTUP_ROM_POST_OMIT_INITIALIZATION.

2001-06-22  Hugo Tyson  <[e-mail]>

	* src/var_misc.c (hal_variant_init): Enable write-access to the
	flash area, and power-up the programming voltage via GPIO0, so
	that flash drivers can work.  Also enable the IBUS Arbiter so that
	internal peripherals can work.

	* src/variant.S (hal_memc_setup): Remove hal_memc_setup for RAM
	start; the work is done elsewhere.

	* include/variant.inc (hal_memc_init): Don't call hal_memc_setup
	for RAM start.
	
2001-06-07  Hugo Tyson  <[e-mail]>

	* include/variant.inc (hal_memc_init): Always do the memc call
	even in RAM start, to enable...

	* src/variant.S (hal_memc_setup): Enable write-access to the flash
	area so that flash drivers can work.

2001-06-06  Hugo Tyson  <[e-mail]>

	* include/var_cache.h (HAL_DCACHE_ENABLE_DEFINED): Add enough NOPs
	after diddling the cache-enability that it works.  Also give
	correct (apparantly!) figures for cache size despite the
	documentation arguing with itself.

	* src/var_misc.c (hal_variant_init): Enable the caches during
	startup.

	* include/var_arch.h: Add a very few definitions for use by
	assembler code here, and tidy up a little so that it can be used
	from .S files.  Specifically this is to let us init the RAM and
	ROM access controllers from assembly.

	* src/variant.S: 
	* include/variant.inc: Cut out some stuff we don't need that I had
	blindly copied from another platform.  Specifically we don't need
	to set up the TLB *at all* because all its space are belong, um,
	all memory, IO and devices are accessible through kseg1 and kseg0.
	Added initial (untested) cut at setup of RAM/ROM controllers.

2001-06-06  Hugo Tyson  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl: Add implements statement for
	CYGINT_HAL_DEBUG_GDB_STUBS_BREAK, moved from the platform HAL.

2001-06-05  Hugo Tyson  <[e-mail]>

	* include/variant.inc: It all works rather better now...  fiddling
	with vector setup and cache initialization.

2001-06-05  Hugo Tyson  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl: to make GDB stubs work, implements
	CYGINT_HAL_MIPS_STUB_REPRESENT_32BIT_AS_64BIT.  Commented out
	implements CYGINT_HAL_VIRTUAL_VECTOR_SUPPORT_NOT_GUARANTEED.

	* include/hal_diag.h: Now it works, properly be dependent on
	CYGSEM_HAL_VIRTUAL_VECTOR_DIAG.

	* include/plf_stub.h: Choose the right init routine dependent on
	CYGSEM_HAL_VIRTUAL_VECTOR_DIAG.

	* src/hal_diag.c (hal_uart_init): Remove diagnostics and delays
	from initialization.

	* src/var_stub.c (hal_var_stub_init): Remove bogus definition of
	vsr_table and leave breakpoint VSR alone, the springboard will
	handle it AOK.

2001-06-04  Hugo Tyson  <[e-mail]>

	* cdl/hal_mips_upd985xx.cdl
	* include/hal_diag.h
	* include/plf_stub.h
	* include/var_arch.h
	* include/var_cache.h
	* include/var_intr.h
	* include/variant.inc
	* src/hal_diag.c
	* src/hal_mips_upd985xx.ld
	* src/var_misc.c
	* src/var_stub.c
	* src/variant.S
	New files; initial checkin.

	RAM startup works; kernel tests run, loaded via SRecords,
	including tm_basic, clockcnv, mutex3 and the exception tests.
	Virtual vector calling works, but all apps do not work with a ROM
	monitor, so they are all standalone.  Diag printf et all all work.
	Floating point emulation works.

//===========================================================================
//####ECOSGPLCOPYRIGHTBEGIN####
// -------------------------------------------
// This file is part of eCos, the Embedded Configurable Operating System.
// Copyright (C) 1998, 1999, 2000, 2001, 2002 Red Hat, Inc.
//
// eCos is free software; you can redistribute it and/or modify it under
// the terms of the GNU General Public License as published by the Free
// Software Foundation; either version 2 or (at your option) any later version.
//
// eCos is distributed in the hope that it will be useful, but WITHOUT ANY
// WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
// for more details.
//
// You should have received a copy of the GNU General Public License along
// with eCos; if not, write to the Free Software Foundation, Inc.,
// 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
//
// As a special exception, if other files instantiate templates or use macros
// or inline functions from this file, or you compile this file and link it
// with other works to produce a work based on this file, this file does not
// by itself cause the resulting work to be covered by the GNU General Public
// License. However the source code for this file must still be made available
// in accordance with section (3) of the GNU General Public License.
//
// This exception does not invalidate any other reasons why a work based on
// this file might be covered by the GNU General Public License.
//
// Alternative licenses for eCos may be arranged by contacting Red Hat, Inc.
// at http://sources.redhat.com/ecos/ecos-license/
// -------------------------------------------
//####ECOSGPLCOPYRIGHTEND####
//===========================================================================