URL
https://opencores.org/ocsvn/or1k/or1k/trunk
Subversion Repositories or1k
[/] [or1k/] [trunk/] [ecos-2.0/] [packages/] [hal/] [mips/] [upd985xx/] [v2_0/] [ChangeLog] - Rev 1765
Compare with Previous | Blame | View Log
2003-04-10 Nick Garnett <nickg@balti.calivar.com>
* src/hal_mips_upd985xx.ld:
Added libsupc++.a to GROUP() directive for GCC versions later than
3.0.
2002-04-02 Hugo Tyson <hmt@redhat.com>
2002-04-02 Anssi Pulkkinen <anssi.pulkkinen@ascom.com>
* 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 <hmt@redhat.com>
* 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 <jlarmour@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <jlarmour@redhat.com>
* cdl/hal_mips_upd985xx.cdl: Platform implements VV baud operations.
2001-08-08 Hugo Tyson <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* include/var_intr.h (HAL_INTERRUPT_[UN]MASK): Better manipulate
these with interrupts disabled for atomicity.
2001-07-02 Bart Veer <bartv@redhat.com>
* src/var_misc.c (hal_variant_init):
Move the GPIO0 manipulating (flash programming voltagE)
to the platform-init.
2001-06-27 Hugo Tyson <hmt@redhat.com>
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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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 <hmt@redhat.com>
* include/variant.inc: It all works rather better now... fiddling
with vector setup and cache initialization.
2001-06-05 Hugo Tyson <hmt@redhat.com>
* 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 <hmt@redhat.com>
* 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####
//===========================================================================