Patch from Steven Scholz:

should we rename the define into CONFIG_FEATURE_HDPARM_HDIO_DMA and
use it for "set" and "get" ???

Since although CONFIG_FEATURE_HDPARM_HDIO_GET_DMA is _not_ set I still
can use "hdparm -d 1 /dev/hda" to switch the DMA on.
Is this desireable or meant to be like that?

So how about the applied patch?


git-svn-id: svn://busybox.net/trunk/busybox@7173 69ca8d6d-28ef-0310-b511-8ec308f3f277

1 files changed, 8 insertions, 0 deletions

diff --git a/miscutils/Config.in b/miscutils/Config.in
index 0c56f0ee9..a994c5096 100644
--- a/miscutils/Config.in
+++ b/miscutils/Config.in
@@ -100,6 +100,14 @@ config CONFIG_FEATURE_HDPARM_HDIO_TRISTATE_HWIF
           and the '-b' option to get/set bus state.  This is dangerous
           stuff, so you should probably say N.
 
+config CONFIG_FEATURE_HDPARM_HDIO_GETSET_DMA
+        bool "  get/set using_dma flag (DANGEROUS)"
+        default n
+        depends on CONFIG_HDPARM
+        help
+          Enables the 'hdparm -d' option to get/set using_dma flag.
+          This is dangerous stuff, so you should probably say N.
+
 config CONFIG_MAKEDEVS
         bool "makedevs"
         default n