Andrea Arcangeli (andrea@e-mind.com)
Mon, 14 Dec 1998 21:29:05 +0100 (CET)
On Mon, 14 Dec 1998, Tim Waugh wrote:
>> This patch will implement a policy where the ieee_ops lowlevel function
>> has to be called with a callback not null. This will improve performance
>> because ~always there' s a valid callback set.
>
>Not sure about this one either. The saving isn't exactly colossal. ;-)
I think it' s cleaner to have not to check for null every time, once it' s
been proposed also for the VFS and if I remeber well other parts of the
kernel use this policy just now...
>> I just included patch-2.1.131-ac8-1284 + the patches above in my arca-tree
>> to handle more testing from people on more hardware...
>
>Okay. Are you planning to merge your arca tree into 2.2 though?
The objective of arca patches is to allow people to try the tree I am
running that can contains good/needed or experimental eventually broken
code. Once I think that something is pretty important (and I have a bit of
time) I ask for inclusion.
>> Tim and what about things like tunelp -w now?
>
>Um. What does tunelp -w do? ;-)
It's used to set LP_WAIT(minor).
With the ieee1284 code it should change the udelay(1):
unsigned char val = (PARPORT_STATUS_SELECT
| PARPORT_STATUS_ERROR
| PARPORT_STATUS_BUSY);
if (current->need_resched)
schedule();
/* Write the character to the data lines. */
parport_write_data (port, *byte++);
udelay (1);
^^^^^^^^ 1 -> LP_WAIT(minor)
/* Is the peripheral ready yet? */
if (!parport_wait_peripheral (port, mask, val))
/* Skip the loop */
goto ready;
Andrea Arcangeli
-- To unsubscribe, send mail to: linux-parport-request@torque.net --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:18:54 EST