freebsd-dev/sys/arm
Ian Lepore 45cd4c8cef Revert the recent armv6 changes to ALIGNED_POINTER(), restoring the
fully-pessimized implementation that requires a type to be aligned to
its natural size.

On armv6+ the compiler might generate load-/store-multiple instructions
which require 4-byte alignment even though the source code is only
accessing individual uint32_t values in a way that doesn't require any
particular alignment at all.  The compiler apparently feels free to
combine multiple accesses into a single instruction that requires a
more-strict alignment, and no set of compiler flags seems to disable
this behavior (at least in clang 3.8).

This fixes alignment faults on arm systems using wifi adapters.  The
wifi code uses ALIGNED_POINTER(p, uint32_t) to decide whether it needs
to copy-align tcp headers.  Because clang is combining several uint32_t
accesses into a single ldm instruction, we need to say that accessing a
uint32_t requires 4-byte alignment.

Approved by:	re(gjb)
2016-06-21 17:53:42 +00:00
..
allwinner
altera/socfpga
amlogic/aml8726
annapurna/alpine
arm Update comments for the MD functions managing contexts for new 2016-06-16 12:05:44 +00:00
at91
broadcom/bcm2835
cavium/cns11xx
conf
freescale Postpone allocation of IRQ resource to the time when interrupt 2016-06-03 11:05:55 +00:00
include Revert the recent armv6 changes to ALIGNED_POINTER(), restoring the 2016-06-21 17:53:42 +00:00
lpc
mv Use nitems() macro instead of re-inventing it 2016-06-03 18:52:57 +00:00
nvidia tegra124: use roundup/rounddown macros from <sys/param.h>. 2016-06-03 21:11:34 +00:00
qemu
rockchip
samsung/exynos
ti
versatile
xilinx
xscale