svn commit: r283014 - head/sys/arm/arm

[Warner Losh]

Author: imp
Date: Sat May 16 21:24:32 2015
New Revision: 283014
URL: https://svnweb.freebsd.org/changeset/base/283014

Log:
  Don't allow unmapped I/O. The pmap isn't quite up to the task. Add a
  comment to this effect and switch the default. My old AT91SAM9G20
  now boots, fsck's the SD card and runs w/o an issue for the first
  time since a 9.1-ish stable build I did a few years ago.
  
  Problems with unmapped I/O:
  o un-page-aligned I/O requests to devices fail (notably fsck
    and newfs).
  o write-back caching was totally broken. write-through caching
    needed to be enabled.
  o Even page-aligned I/O requests sometimes failed for reasons
    not thoroughly investigated.
  
  Suggested by: ian@
  MFC after: 2 days

Modified:
  head/sys/arm/arm/pmap.c

Modified: head/sys/arm/arm/pmap.c
==============================================================================
--- head/sys/arm/arm/pmap.c	Sat May 16 21:08:33 2015	(r283013)
+++ head/sys/arm/arm/pmap.c	Sat May 16 21:24:32 2015	(r283014)
@@ -4310,7 +4310,13 @@ pmap_copy_page(vm_page_t src, vm_page_t 
 	pmap_copy_page_func(VM_PAGE_TO_PHYS(src), VM_PAGE_TO_PHYS(dst));
 }
 
-int unmapped_buf_allowed = 1;
+/*
+ * We have code to do unmapped I/O. However, it isn't quite right
+ * an causes un-page-aligned I/O to devices to fail (most notably
+ * newfs or fsck). We give up a little performance to do this, but
+ * we trade that for rock-solid stability so it is a good trade.
+ */
+int unmapped_buf_allowed = 0;
 
 void
 pmap_copy_pages(vm_page_t ma[], vm_offset_t a_offset, vm_page_t mb[],