The single write to memory or ioport output are mostly
atomic operations already. The proper memory synchronization barrier
should be used around them to guarantee ordering (sync or eieio
on PowerPC for example) but because I have not found settable
portable primitive only compiler barrier is used.
It should be enough on x86 because the externally visible order
should be/is guaranteed to be preserved on x86 architecture.
Rename _ISR_Disable() into _ISR_Local_disable(). Rename _ISR_Enable()
into _ISR_Local_enable(). Remove _Debug_Is_owner_of_giant().
This is a preparation to remove the Giant lock.
Update #2555.
GCC 6.0 previews give the following warning on sonic.c:
../../../../../rtems/c/src/libchip/network/sonic.c:837:11: warning:
statement is indented as if it were guarded by... [-Wmisleading-indentation]
This was discussed here:
https://lists.rtems.org/pipermail/devel/2016-March/014004.html
and we concluded it would be better to rewrite the entire snippet.
This first step eliminates the following as public APIs for the pc386
BSP:
+ pcib_conf_read8
+ pcib_conf_read16
+ pcib_conf_read32
+ pcib_conf_write8
+ pcib_conf_write16
+ pcib_conf_write32
The if_fxp.c driver uses these enough where I provided local macros
to allow the code to be mostly unmodified. On other architectures
these names have been used privately. It will take multiple patches
to completely eliminate these symbols from the RTEMS source tree.
The focus of the first effort is just to eliminate these as a public
pc386 API so support can be added for systems without legacy PCI BIOS.
This was flagged by CodeSonar. It should be impossible to get an
incorrect baud number back but ensure this in debug mode. The _Assert()
keeps their scanner from evaluating for divide by 0 past this point.
This patch eliminates most of the warnings in this BSP but attempts
very little clean up. This BSP includes copies of a lot of code
from free NDS libraries and modifications should be kept to a minimum.
Call rtems_termios_dequeue_characters() only in case all characters of
the last write handler invocation have been transmitted. This avoids
problems with pppstart() since the start line discipline does not know
how many characters have been dequeued and assumes the maximum.
This patch allows the user to configure the PHY address for the DWMAC
driver by giving a pointer to a dwmac_user_cfg structure to network
stack via rtems_bsdnet_ifconfig::drv_ctrl.
Reduce number of packet transmitted interrupts by using the interrupt mechanism only
if we run out of DMA descriptors.
Under normal conditions regaining DMA descriptors, mbufs and clusters is handled
via a counter.
This partially reverts commit 1215fd4d94.
In order to support profiling of SMP locks and provide a future
compatible SMP locks API it is necessary to add an SMP lock destroy
function. Since the commit above adds an SMP lock to each chain control
we would have to add a rtems_chain_destroy() function as well. This
complicates the chain usage dramatically. Thus revert the patch above.
A global SMP lock for all chains is used to implement the protected
chain operations.
Advantages:
* The SAPI chain API is now identical on SMP and non-SMP
configurations.
* The size of the chain control is reduced and is then equal to the
Score chains.
* The protected chain operations work correctly on SMP.
Disadvantage:
* Applications using many different chains and the protected operations
may notice lock contention.
The chain control size drop is a huge benefit (SAPI chain controls are
66% larger than the Score chain controls). The only disadvantage is not
really a problem since these applications can use specific interrupt
locks and unprotected chain operations to avoid this issue.
Sebastian Huber
Pavel Pisa
Joel Sherrill
Chris Johns
Martin Galvan
Joel Sherrill
Josh Oguin
Christian Mauderer
Daniel Hellstrom
Ralf Kirchner
Pavel Pisa
Daniel Cederman