1. 01 6月, 2005 4 次提交
    • D
      [PATCH] USB: sl811-hcd fixes · 1e9a47b6
      David Brownell 提交于
      Various fixes to the sl811-hcd driver:
      
        * Fix small glitches that crept in during recent evolution of usbcore's hcd
          glue layer, coupling endpoint state records to usbcore and active urbs.
          (As noted by folk whose boards weren't stuck on 2.6.9 kernels...)
      
        * Cope with various system-specific issues:
            - Some configurations (e.g. a CF-card uses this chip) have iospace
              addresses for the two registers, rather than memory mapped ones.
            - Some configurations do interesting things with IRQs; maybe the
              line is shared, or it doesn't support level triggering.
            - Not all boards can drive the chip reset line in software.
      
        * Address a potential race during unlinking.
      
        * Tweak probe/remove section info to handle the case where this segment
          of a platform bus is hotpluggable (e.g. CF card).  (The basic problem
          is that CONFIG_HOTPLUG is global, which is wrong since not all busses
          can hotplug even on hotplug-friendly systems...)  Also export the
          driver, so that the CF driver can depend on it.
      
      Also removed some annoying end-of-line whitespace.
      Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
      Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
      1e9a47b6
    • L
    • T
      [IA64] Use "PER_CPU" form of EXPORT macro · fffcc150
      Tony Luck 提交于
      I was gently reminded that there are per-cpu forms of the EXPORT_SYMBOL macros.
      Signed-off-by: NTony Luck <tony.luck@intel.com>
      fffcc150
    • J
      [PATCH] Relax idecd dma alignment check · 4e7c6816
      Jens Axboe 提交于
      Only the address needs alignment of mask bits, length should work with
      a relaxed alignment check.
      Signed-off-by: NJens Axboe <axboe@suse.de>
      
      [ This is take 2: make the length check be for 16-byte alignment, not
        just word alignment.  That should hopefully keep everybody happy,
        while still allowing CD writing with DMA ]
      Signed-off-by: NLinus Torvalds <torvalds@osdl.org>
      4e7c6816
  2. 31 5月, 2005 1 次提交
  3. 30 5月, 2005 26 次提交
  4. 29 5月, 2005 9 次提交