When data cache snooping is not present the cache needs
flushing, the SPARC LEON CPUs does not have to ability
to flush individual cache rows and flushing all cache is
expensive. Instead the LDA instruction is used to force
cache miss on individual loads during the IP-align copy
operation required anyway.
GRETH GBIT non-snooping systems are still unsupported,
since it use zero-copy (can deal with unaligned DMA).
Let the bsp.h select if the GRETH driver is supported.
Currently only the LEON2/LEON3 platforms BSPs builds the
driver.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
Moved the intelligence whether the driver is supported or not
completely to the BSP, now that bsp.h is included (even though it
is a driver... hmm). The ERC32 was never supported, so HAS_SMC91111
is not added to erc32/include/bsp.h.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
The low level routines can be used in different occasions, it will be
required when accessing PCI.
Note the difference between byteorder.h (inlined functions) and access.S
where the functions will be declared in the library archive librtemscpu.a.
Function names starting with _ are in library and can be referenced by
function pointers.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
o Removed RTEMS_LIBIO_PERMS_SEARCH.
o Renamed RTEMS_LIBIO_PERMS_READ in RTEMS_FS_PERMS_READ.
o Renamed RTEMS_LIBIO_PERMS_WRITE in RTEMS_FS_PERMS_WRITE.
o Renamed RTEMS_LIBIO_PERMS_EXEC in RTEMS_FS_PERMS_EXEC.
o Renamed RTEMS_LIBIO_FOLLOW_HARD_LINK in RTEMS_FS_FOLLOW_HARD_LINK.
o Renamed RTEMS_LIBIO_FOLLOW_SYM_LINK in RTEMS_FS_FOLLOW_SYM_LINK.
o Renamed RTEMS_LIBIO_MAKE in RTEMS_FS_MAKE.
o Renamed RTEMS_LIBIO_EXCLUSIVE in RTEMS_FS_EXCLUSIVE.
o Renamed RTEMS_LIBIO_ACCEPT_RESIDUAL_DELIMITERS in
RTEMS_FS_ACCEPT_RESIDUAL_DELIMITERS.
o Renamed RTEMS_LIBIO_REJECT_TERMINAL_DOT in
RTEMS_FS_REJECT_TERMINAL_DOT.
New defaults rtems_filesystem_default_fsync_or_fdatasync() and
rtems_filesystem_default_fsync_or_fdatasync_success() for fsync_h and
fdatasync_h. The rtems_filesystem_default_fsync_or_fdatasync() sets now
errno to EINVAL according to POSIX.
o A new data structure rtems_filesystem_global_location_t was
introduced to be used for
o the mount point location in the mount table entry,
o the file system root location in the mount table entry,
o the root directory location in the user environment, and
o the current directory location in the user environment.
During the path evaluation global start locations are obtained to
ensure that the current file system instance will be not unmounted in
the meantime.
o The user environment uses now reference counting and is protected
from concurrent access.
o The path evaluation process was completely rewritten and simplified.
The IMFS, RFS, NFS, and DOSFS use now a generic path evaluation
method. Recursive calls in the path evaluation have been replaced
with iteration to avoid stack overflows. Only the evaluation of
symbolic links is recursive. No dynamic memory allocations and
intermediate buffers are used in the high level path evaluation. No
global locks are held during the file system instance specific path
evaluation process.
o Recursive symbolic link evaluation is now limited by
RTEMS_FILESYSTEM_SYMLOOP_MAX. Applications can retrieve this value
via sysconf().
o The device file system (devFS) uses now no global variables and
allocation from the workspace. Node names are allocated from the
heap.
o The upper layer lseek() performs now some parameter checks.
o The upper layer ftruncate() performs now some parameter checks.
o unmask() is now restricted to the RWX flags and protected from
concurrent access.
o The fchmod_h and rmnod_h file system node handlers are now a file
system operation.
o The unlink_h operation has been removed. All nodes are now destroyed
with the rmnod_h operation.
o New lock_h, unlock_h, clonenod_h, and are_nodes_equal_h file system
operations.
o The path evaluation and file system operations are now protected by
per file system instance lock and unlock operations.
o Fix and test file descriptor duplicate in fcntl().
o New test fstests/fsnofs01.
We increased PMEM to 8 kB but never updated SOFTUSB_PMEM_SIZE.
This didn't have much of an effect because all we used it for was
to zero unused program memory - which we should never touch
anyway.
But now that "usb load" in FN uses SOFTUSB_PMEM_SIZE, this caused
it to silently truncate the firmware, producing rather puzzling
effects.
Signed-off-by: Joel Sherrill <joel.sherrill@oarcorp.com>
Leave the task of providing a suitable firmware for the Milkymist softusb
core to the application instead of having an inflexible array in the driver.
Signed-off-by: Gedare Bloom <gedare@rtems.org>
Slightly adjust video timing registers since we had problems with a few screens and the previous values.
Signed-off-by: Gedare Bloom <gedare@rtems.org>
With the addition of dynamically registered libchip serial devices,
there is the need to be able to use printk() before the console driver
has initialized the indirect pointer table. This lets printk() support
routines call pass a control structure directly without a lookup through
the uninitialized indirect one.
PR 2015/bsps
Since the configuration struct is always present one can let
DATA initialize it to reduce footprint, at the same time it
is made weak to let the user able to configure the SHM driver
without editing the driver code.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
This PR was about a warning for no previous prototype for
rtems_interrupt_level_attribute. This method exists (like
a few others) to have real bodies for Classic API services
implemented as macros. These macros are not available from
anything but C and C++. The most explicit use was in the Ada
binding but these would be needed from assembly language
or any other non-C based language.
On top of needing a prototype, the methods were misnamed.
They were related to modes. This renames them, moves the
file, fixes test code, etc.
* console/console.c: char_ready() was never returning true so console
never processed input data
* console/console.c: added printk() support to default device
* include/bsp.h: Added #define for MOT_162BUG_VEC_ADDRESS
* startup/bspclean.c: Modified to use MOT_162BUG_VEC_ADDRESS
* startup/bspstart.c: Modified to use MOT_162BUG_VEC_ADDRESS
* make/custom/mvme162.cfg: Modified to use "RTEMS_CPU_MODEL=68lc040"
and "CPU_CFLAGS = -mcpu=68040 -msoft-float" so BSP will always
work with all board variations.
* README: Added notes on user required configuration changes and
information about board models and variants
* README.models: New file that contains a detailed list of MVME162
models and variants.
Signed-off-by: Vic Hoover <victor.hoover.ctr@navy.mil>
* console/console.c: char_ready() was never returning true so console
never processed input data
* console/console.c: added printk() support to default device
* include/bsp.h: Added #define for MOT_162BUG_VEC_ADDRESS
* startup/bspclean.c: Modified to use MOT_162BUG_VEC_ADDRESS
* startup/bspstart.c: Modified to use MOT_162BUG_VEC_ADDRESS
* make/custom/mvme162.cfg: Modified to use "RTEMS_CPU_MODEL=68lc040"
and "CPU_CFLAGS = -mcpu=68040 -msoft-float" so BSP will always
work with all board variations.
* README: Added notes on user required configuration changes and
information about board models and variants
* README.models: New file that contains a detailed list of MVME162
models and variants.
Signed-off-by: Vic Hoover <victor.hoover.ctr@navy.mil>
GRETH driver updated, 10-15% performance improvements for GBIT MAC,
unnecessary RX interrupts not taken which under heavy load saves approx.
1500 interrupts/s, one task removed saving about 5kb memory and 1 bug
solved.
BUG: RX interrupt was enabled before the RX-daemon was created which could
result in a faulty call to rtems_event_send.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
Deadlock may arise when the EDCL bug link is used to tunnel
console output over Ethernet, when Ethernet is down one should
avoid using console (only during debugging of LEON targets)
Author: Marko Isomaki <marko@gaisler.com>
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
PR 2010/bsps
The previous code only checked if d-cache snooping was implemented,
however snooping may be available but not enabled which may lead
to driver bugs.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
PR 2009/bsps
All LEON3/4 systems have a CPU-id, if on a single-CPU system the
ID is always zero. On a multicore system it ranges from 0 to 15.
The CPU index should always by updated even in a non-MP RTEMS OS
since the CPU running RTEMS may not always be CPU0. For example
when RTEMS runs on CPU1 and Linux on CPU0 in a mixed ASMP system.
The old code executed within the IRQ controller initialization code
makes no sense since the ASR register is a CPU register, it has
nothing to do with AMBA initialization either.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
The SHM code always wakes one CPU more that configured, however
this has never been a problem since RTEMS will be running on all CPUs
or only two cores were available.
PR 2006/bsps
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
