Date: Sat, 28 Jun 2014 20:38:23 -0700 From: John-Mark Gurney <jmg@funkthat.com> To: arm@FreeBSD.org Subject: arm alignment faults... Message-ID: <20140629033823.GN1560@funkthat.com>
index | next in thread | raw e-mail
So, one of the little projects I'd like to see is the removal of
ETHER_ALIGN from the tree.. This bogosity can (and does) cause the use
of bouncing durning DMA ops on all ethernet frames...
In investigating this issue, I found CPU_CONTROL_AFLT_ENABLE... The
interesting part is that it looks like most of the code in arm/cpufunc.c
has dead code:
cpuctrlmask = CPU_CONTROL_MMU_ENABLE | CPU_CONTROL_32BP_ENABLE
| CPU_CONTROL_32BD_ENABLE | CPU_CONTROL_SYST_ENABLE
| CPU_CONTROL_IC_ENABLE | CPU_CONTROL_DC_ENABLE
| CPU_CONTROL_WBUF_ENABLE | CPU_CONTROL_ROM_ENABLE
| CPU_CONTROL_BEND_ENABLE | CPU_CONTROL_AFLT_ENABLE
| CPU_CONTROL_LABT_ENABLE | CPU_CONTROL_VECRELOC
| CPU_CONTROL_ROUNDROBIN;
#ifndef ARM32_DISABLE_ALIGNMENT_FAULTS
cpuctrl |= CPU_CONTROL_AFLT_ENABLE;
#endif
If you define ARM32_DISABLE_ALIGNMENT_FAULTS, you'll still get alignment
faults since that is what is specified above... If anything, the code
needs to be changed to:
#ifdef ARM32_DISABLE_ALIGNMENT_FAULTS
cpuctrl &= ~CPU_CONTROL_AFLT_ENABLE;
#endif
Though mv_pj4b is the only one to get this correct...
It looks like it was this way when the original code was imported..
Shall we just delete the ifndef ARM32_DISABLE_ALIGNMENT_FAULTS sections
then?
--
John-Mark Gurney Voice: +1 415 225 5579
"All that I will do, has been done, All that I have, has not."
help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140629033823.GN1560>