| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
* shared/comm/uart.c, shared/comm/uart.h: Reflect termios API
changes.
|
| |
|
|
|
|
|
| |
* console/console.c, console/ps2_mouse.c: Reflect termios API
changes.
|
|
|
|
|
|
|
| |
* make/custom/pc386.cfg: EXEEXT is not defined in application Makefile
support. All other BSPs just use .exe directly on objcopy commands.
This makes it consistent and allows external examples (network demos,
examples-v2, etc to link.
|
|
|
|
| |
* make/custom/pc386.cfg: Revert the PR1495 change.
|
|
|
|
|
|
|
| |
PR 1495/bsp
* clock/ckinit.c, make/custom/pc386.cfg: Calling
rtems_clock_get_uptime() in a tight loop sometimes showed time moving
backwards.
|
|
|
|
| |
* include/bsp.h: Add BSP_HAS_FRAME_BUFFER.
|
|
|
|
| |
* clock/ckinit.c: Fix warning.
|
|
|
|
|
| |
* startup/bspgetworkarea.c: Make bsp_mem_size available for page table
initialization.
|
|
|
|
|
| |
PR 1483/cpukit
* shared/pci/pcibios.c: Rename pci_config into rtems_pci_config_t.
|
|
|
|
|
|
|
|
| |
* include/bsp.h, startup/bspgetworkarea.c: Rework bsp_size_memory() to
ensure that multiboot information regarding memory size is used as
the primary source. This was broken in the move to supporting a
unified workspace. It worked this way in 4.9 so this was a
regression.
|
|
|
|
| |
* console/console.c: Eliminate casts.
|
|
|
|
| |
* shared/comm/uart.c, shared/comm/uart.h: Adjust prototypes.
|
|
|
|
|
|
| |
* startup/bspgetworkarea.c: Add debug printk's.
* startup/linkcmds: Using "-Wl,--defsym -Wl,RamSize=0x1000000"
results in _RamSize being defined. Handle this properly.
|
|
|
|
|
| |
* include/bsp.h: BSP should not define
CONFIGURE_NUMBER_OF_TERMIOS_PORTS.
|
| |
|
| |
|
| |
|
|
|
|
|
| |
* ne2000/ne2000.c: Logic on defaulting port number was inverted when
adding command line arguments.
|
|
|
|
|
| |
* pc386/make/custom/pc586-sse.cfg, pc386/start/start.S,
shared/irq/irq_asm.S: Added experimental SSE support.
|
|
|
|
|
|
| |
* clock/ckinit.c: Remove nested externs.
Add missing prototypes.
Fix Clock_isr prototype.
|
|
|
|
|
| |
* clock/ckinit.c: Fix Clock_isr prototype.
* network/network.c: Use PRI*32 to print uint32_t's.
|
| |
|
| |
|
|
|
|
| |
* console/console.c: Fixed compiler warning.
|
|
|
|
|
|
|
|
|
| |
* start/start.S, console/console.c: Separated code
that switched console and printk port based on info
gathered from the commandline and/or a BSP provided
routine into new BSP_console_select() which is called
from 'start.S' so that early messages make it to the
redirected printk port.
|
|
|
|
|
|
|
|
| |
* shared/irq/irq_asm.S: Make sure stack is aligned to CPU_STACK_ALIGNMENT
for *all* C-routines (including _Thread_Dispatch() and
_ThreadProcessSignalsFromIrq()) not only C_dispatch_isr().
* shared/irq/irq.c: Added IRQ statistics counters.
|
|
|
|
|
| |
* shared/irq/irq_asm.S: Beautification; ajusted margins and
spaces to make the whole thing more readable.
|
|
|
|
|
|
|
| |
* shared/irq/irq_asm.S, shared/irq/irq.c: removed
(fake) exception frame argument to _ThreadProcessSignalsFromIrq().
This has never been used and removing it simplifies the assembly
code.
|
|
|
|
|
| |
* start/start.S: Make sure initial stack is aligned
to CPU_STACK_ALIGNMENT before calling C-code.
|
|
|
|
|
| |
* shared/irq/irq_asm.S: Make sure stack is aligned to
CPU_STACK_ALIGNMENT before invoking C-code.
|
|
|
|
|
| |
* include/bsp.h, console/console.c: Provide weak-alias
hook so that the application may choose the console.
|
|
|
|
| |
* make/custom/ts_386ex.cfg: Remove RTEMS_BSP_FAMILY.
|
|
|
|
| |
* make/custom/pc386.cfg: Remove RTEMS_BSP_FAMILY.
|
|
|
|
| |
* make/custom/i386ex.cfg: Remove RTEMS_BSP_FAMILY.
|
|
|
|
| |
* configure.ac: Don't add -ansi -fasm to CFLAGS.
|
|
|
|
|
|
|
| |
* shared/irq/idt.c: Check is the irq handlers are present before
calling.
* shared/comm/i386-stub-glue.c: Revert the change of 2009-05-06 as
the handlers need to be present. Fixed the warnings.
|
|
|
|
|
|
|
|
|
|
| |
* ne2000/ne2000.c: Add --ne2k-irq and --ne2k-port boot command
line configure options.
* ide/ide.c: Fix a bug which left 4 words in the buffer of the
disk. Some devices do not follow the standard and terminate the
command which a new command occurs and/or low data ready when data
is still to be read.
|
|
|
|
| |
* make/custom/ts_386ex.cfg: New (relocated from /make/custom).
|
|
|
|
|
|
|
|
| |
* make/custom/pc386.cfg: New (relocated from /make/custom).
* make/custom/pc386dx.cfg: New (relocated from /make/custom).
* make/custom/pc486.cfg: New (relocated from /make/custom).
* make/custom/pc586.cfg: New (relocated from /make/custom).
* make/custom/pc686.cfg: New (relocated from /make/custom).
|
|
|
|
| |
* make/custom/i386ex.cfg: New (relocated from /make/custom).
|
| |
|
|
|
|
|
|
|
|
| |
* ide/ide.c: Fixed more IDE driver bugs. The driver can transfer
64K in a single write on my test PC and qemu with this driver. The
driver only checks the data ready bit at the start of each 512
block being transfered. This has speed the transfer loop up. Fixed
the timing so the BSP timer is used until RTEMS has started.
|
|
|
|
|
| |
PR 1445/BSPs
* shared/pci/pcibios.c: Rename indirect_pci_* into BSP_pci_*.
|
|
|
|
|
| |
* shared/comm/tty_drv.c: Reflect termios_baud_to_number having been
renamed to rtems_termios_baud_to_number.
|
|
|
|
|
| |
* console/console.c: Reflect termios_baud_to_number having been
renamed to rtems_termios_baud_to_number.
|
| |
|
|
|
|
|
| |
* console/console.c: Reflect termios_baud_to_number having been
renamed to rtems_termios_baud_to_number.
|
|
|
|
|
|
| |
* console/console.c, console/serial_mouse.c: Reflect
termios_baud_to_number having been renamed to
rtems_termios_baud_to_number.
|
|
|
|
|
|
|
| |
PR 1344/bsps:
* clock/ckinit.c: Fix more bugs (applied PR1344/attachment 702):
TSC handler was used when 8254 was requested and vice versa.
Also, nanosecond handler was never installed to RTEMS clock.
|