| commit | 026699961337ea21e05f08e1043c4a8411a7f92b | |
| author | Josef 'Jeff' Sipek <jeffpc@josefsipek.net> | |
| Tue, 15 Nov 2011 21:53:10 +0000 (15 16:53 -0500) | ||
| committer | Josef 'Jeff' Sipek <jeffpc@josefsipek.net> | |
| Tue, 15 Nov 2011 21:53:10 +0000 (15 16:53 -0500) | ||
| tree | 2a536b070d5e5af772ce71f98d3d3166a72ea26b | treesnapshot (tar.gz zip) |
| parent | 95e8e4164bcc30f2506813966feb7233b8e6a3a1 | commitdiff |
cp: replace crusty console line buffering mechanism with a spoolfile
The spool file layer needs to exist anyways so let's use it. It does mean
that a console I/O from a guest will end up being copied 3 times (guest
buffer -> spooldev_cons buffer -> spoolfile -> I/O buffer). The is no way
to avoid at least one copy. At this point, it's pointless to try to
optimize that and so we'll live this behavior for now.
Signed-off-by: Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
The spool file layer needs to exist anyways so let's use it. It does mean
that a console I/O from a guest will end up being copied 3 times (guest
buffer -> spooldev_cons buffer -> spoolfile -> I/O buffer). The is no way
to avoid at least one copy. At this point, it's pointless to try to
optimize that and so we'll live this behavior for now.
Signed-off-by: Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
| cp/drivers/console.c | diffblobblamehistory | |
| cp/include/console.h | diffblobblamehistory |