Re: [Xen-devel] [PATCH] pygrub: make it work

[Frank van der Linden <Frank.Vanderlinden@xxxxxxx>]

Ian Jackson wrote:

> Surely this dependence on uname is wrong.  Why would set the termios
> flags on the pty master ?  They should be set on the slave
> unconditionally.
It's been a while since I originally did this patch, so my memory is a 
little rusty, but as far as I remember, I simply wanted to avoid 
breaking Linux after an earlier version of the patch did so. I tested 
the final version of the patch on Linux, and it worked fine. It could be 
that the uname tests aren't needed anymore.
> And keeping the slave fd around for a while is harmless too.  So I
> guess I'm asking why these changes need to be conditional.
>
> > +    # filedescriptors:
> > +    #   r - input from the bootloader (bootstring output)
> > +    #   m1 - input/output from/to xenconsole
> > +    #   m2 - input/output from/to pty that controls the bootloader
> Can you explain what the purpose of this new code is or relevantly how
> it differs from the old (ie, what was wrong with the old code) ?
The comments I added explain it:

+    # The filedescriptors are NDELAY, so it's ok to try to read
+    # bigger chunks than may be available, to keep e.g. curses
+    # screen redraws in the bootloader efficient. m1 is the side that
+    # gets xenconsole input, which will be keystrokes, so a small number
+    # is sufficient. m2 is pygrub output, which will be curses screen
+    # updates, so a larger number (1024) is appropriate there.
+    #
+    # For writeable descriptors, only include them in the set for select
+    # if there is actual data to write, otherwise this would loop too fast,
+    # eating up CPU time.


- Frank

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel