[odc] Daily src changes for 2009-02-05

[ODC]

OpenBSD src changes summary for 2009-02-05
==========================================

share/mk                                sys/arch/amd64/amd64
sys/arch/amd64/include                  sys/arch/i386/i386
sys/arch/i386/include                   sys/dev/pci
sys/uvm                                 usr.sbin/rtsold
usr.sbin/smtpd                          

== share ============================================================= 01/03 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/share

mk

  ~ bsd.lib.mk                            

  > Revert 1.54 and build debug libs on hp300 too. (miod@)

== sys =============================================================== 02/03 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/sys

arch/amd64/amd64

  ~ bus_dma.c                             ~ pmap.c

  > Mirroring the i386 commit just made. Add MD PMAP_NOCACHE flag to pmap,
  > and use it to implement BUS_DMA_NOCACHE for uncached mappings of dma
  > memory. Needed for some broken hardware.
  > Discussion with art, miod, kettenis and toby, ok miod. (oga@)

arch/amd64/include

  ~ bus.h                                 ~ pmap.h

  > Mirroring the i386 commit just made. Add MD PMAP_NOCACHE flag to pmap,
  > and use it to implement BUS_DMA_NOCACHE for uncached mappings of dma
  > memory. Needed for some broken hardware.
  > Discussion with art, miod, kettenis and toby, ok miod. (oga@)

arch/i386/i386

  ~ bus_dma.c                             ~ pmap.c

  > add MD PMAP_NOCACHE flag to i386 and use it to implement the
  > BUS_DMA_NOCACHE flag with guarantees that the dma memory will be mapped
  > uncached. Some broken/odd hardware needs this.
  > discussion with miod, toby, art and kettenis. ok miod. (oga@)

arch/i386/include

  ~ bus.h                                 ~ pmap.h

  > add MD PMAP_NOCACHE flag to i386 and use it to implement the
  > BUS_DMA_NOCACHE flag with guarantees that the dma memory will be mapped
  > uncached. Some broken/odd hardware needs this.
  > discussion with miod, toby, art and kettenis. ok miod. (oga@)

dev/pci

  ~ drm/drmP.h                            ~ drm/files.drm
  ~ drm/i915_dma.c                        ~ drm/i915_drv.c
  ~ drm/i915_drv.h                        ~ drm/i915_mem.c
  ~ drm/radeon_drv.c                      ~ drm/radeon_drv.h
  ~ drm/radeon_mem.c                      ~ drm/radeon_state.c
  + drm/drm_heap.c                        

  > Merge the static block allocation code from {i915,radeon}_mem.c into
  > non-static code that's shared between both.  While i'm here convert them
  > to TAILQ.
  > Eventually, both of these will die, but until then I'd rather shave the
  > space in the kernel.
  > Tested on radeon and intel. (oga@)

  ~ drm/i915_mem.c                        

  > KNF, no change to behaviour. (oga@)

  ~ drm/radeon_mem.c                      

  > KNF, and remove a debug printf I missed. (oga@)

  ~ drm/i915_mem.c                        ~ drm/radeon_mem.c

  > Fix slight logic error in a previous commit. (oga@)

  ~ azalia.c                              

  > Remove the BUS_DMA_NOCACHE flag from all azalia dma mappings.
  > The fact that this has only just been implemented, and that this
  > hardware is only found on x86 shows that this flag should not be needed
  > for operation of this hardware. It's been here since the initial netbsd
  > commit the code came from.
  > "sounds good to me" jakemsr@, ok kettenis at . (oga@)

  ~ if_wpi.c                              

  > fix uninitialized var (in a code path that is not used yet).
  > found by chl@ using llvm/clang (damien@)

uvm

  ~ uvm_pmap.h                            

  > Add some (currently unused) MD pmap flags to be used for pushing in
  > parameters such as cacheability, which is too different per-arch to be
  > MI.
  > discussed with miod, kettenis and art. ok miod@, art at . (oga@)

== usr.sbin ========================================================== 03/03 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin

rtsold

  ~ rtsol.c                               

  > fix format string
  > ok naddy@ (chl@)

smtpd

  ~ ssl.c                                 

  > when calling ssl_session_destroy() from mta, our session will have a NULL
  > s->s_l field. this commit fixes a crash in mta caused by a NULL deref.
  > (gilles@)

===============================================================================