Andrea Arcangeli (firstname.lastname@example.org)
Fri, 27 Feb 1998 01:17:39 +0100 (CET)
On Thu, 26 Feb 1998, Tim Waugh wrote:
>On Thu, 26 Feb 1998, Andrea Arcangeli wrote:
>> I added the yield ioctls.
>Thanks. I think only the blocking version makes sense, though. Why would
>a user application ever be fussed about not blocking?
Yes, the not blocking is useful only if you need to yield in a interrupt
>> @@ -360,7 +382,6 @@
>> pp_table[minor].pdev = pdev;
>> - init_waitqueue (&pp_table[minor].irq_wait);
>> printk (KERN_DEBUG "%s: registered pardevice\n", name);
>> return 0;
Yup, I forget to repristin this line. I removed it because at first I
removed also the irq_wait field from pp_struct because I was going to use
the wait_q field as lp does, but then I understand that was wrong since a
buggy driver can run pp_poll() also if it hasn' t claimed the port. Using
a separate wait_queue assure that the buggy driver will stall showing the
bug. Please reject the line of course.
-- To unsubscribe, send mail to: email@example.com --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:17:30 EST