By default, pfkey allocates a 2MB buffer that is used for SPD entries.
This size is a good choice for a server system where a lot of clients
should be handled. But on our embedded systems, an application with that
much clients is unlikely and 2MB is a lot of space. So reduce that to
the default value of 128kB which should be enough for a small number of
ipsec connections.
See https://bugzilla.redhat.com/show_bug.cgi?id=607361 for more details
why the upstream project originally increased the size.
If someone really needs a bigger size, there is a option in the
configuration file of pfkey called `pfkey_buffer` that can overwrite
this value.
Closes#4621
The pppstart expected that a driver write would somehow magically
process all data passed to the write function. Because ppp disables all
buffering that originally has been in termios, that assumption is not
true for all but polled drivers.
With this patch, the pppstart now gets and processes the feedback that
is returned from the driver via rtems_termios_dequeue_characters.
Fixes#4493
When the cmpset and fcmpset functions were refactored, the return value
of the operation was discarded instead of being returned for SMP builds
outside of gcc 4.x. This had the effect of turning these functions into
a long busywait loop that eventually failed due to integer overflow.
This patch restores the use of the return value of the atomic
operations.
There was a typo in the patch that added the HAVE_<LIBRARY> definition
that prevented "./waf configure" from succeeding. This adds the missing
character.
Conditionally compiled tests (consisting only of debugger01) were not
compiling as expected when libdebugger was present. This appears to have
occurred during the transition from header detection or due to an
intervening change in the waf internal libraries. The result of
check_cc() is the only location this information is reported, so library
checks now set HAVE_<LIBRARY> as appropriate when the library is found
so existing code for header configuration correctly recognize that a
library is present.
- The script's use of sort proved to not be portable
- No need to check the commits as symbols are only added
- Regenerated kernel header to reset the sort order
Update #4475
- Wait for a slow PHY to bring the link UP. If the IP address is
static the test can start before the link is up and the test
fails.
- Make 2 tests wait. Others will need to be added.
- Add PCI IO region support
- Add support map buffers to PCI address space
- Add BSP conditional IO space support. Some PC implementations
have PCI IO space mapped differently to memory space and this needs
to be reflected in the busspace.
- Include bsp.h to pick per BSP configuration.
Closes#4245
The default implementation was completely broken. Use the GCC specific
__alignof__ if compiling for C99 or C++03. If not C++11, C11, or
GCC, then it is an error.
When the MII bus is unterminated on unused interfaces, it results in PHY
read timeouts which manifest as spurious PHYs during the attach call.
Detect these timeouts during the probe so the device can be ignored.
This alters the selection of the 4 Cadence GEM interfaces on the Zynq
Ultrascale+ MPSoC BSP to be provided by config.inc instead of being
provided by options in the RTEMS BSP itself since those options appear
to be dead code when not used in conjunction with LibBSD.
