This is a good idea, and I have implemented it, but with a twist.

The configuration script will ask a question about whether or not you want to use type int for CPU flags. The default is 'y' for backward compatibility. If you answer 'n' you get unsigned long type and everything inside LiS (except inet) will compile with no warnings. However, you may get warnings when you compile your driver code. That is why the default is 'y'.

I will post a new LiS in a few minutes for y'all to play with over the weekend.

-- Dave

At 11:41 AM 5/27/2003 Tuesday, Howard Selover wrote:
Hello Dave,

It came to my attention today that the lis_spin_lock_irqsave() and the other variants supplied by LiS have the flags parameter prototyped as an int. The underlying Linux locks that are mapped by LiS prototype the flags parameter as an unsigned long. This is not a problem in the 32 bit Intel world but may become an issue if LiS is used on other architectures. I would recommend that we change the prototypes of the LiS functions to conform to those supplied by the OS. What do you think?

Howard

--
Howard Selover III
Principal Engineer
Chief Architect's Office
Ulticom, Inc.
1020 Briggs Road
Mount Laurel, NJ 08054

Direct: +1-856-787-2739
Mobile: +1-856-495-4181
Fax:    +1-856-866-2033
Email:  [EMAIL PROTECTED]
Web:    www.ulticom.com



_______________________________________________
Linux-streams mailing list
[EMAIL PROTECTED]
http://gsyc.escet.urjc.es/mailman/listinfo/linux-streams


_______________________________________________
Linux-streams mailing list
[EMAIL PROTECTED]
http://gsyc.escet.urjc.es/mailman/listinfo/linux-streams

Reply via email to