Thread (20 messages) 20 messages, 8 authors, 2006-10-16

Re: [Sm5xx-devel] SM501 framebuffer driver

flat view

From: Bill Gatliff <hidden>
Date: 2006-10-16 19:47:54

Robert Whaley wrote:
Bill Gatliff wrote:
quoted
Paul:

Paul Mundt wrote:
quoted
Hi Bill, fancy seeing you in this thread ;-)
 
Heh, you need to talk to the guy at the door.  He'll let anyone in, 
I'm proof!  :)
quoted
We want the 8051 blob in userspace if anything, there's little 
reason to
tie this in to the module specifically, and it's handy to expose the
interface for people that are specifically interested in loading in
alternate code on the 8051.

As far as including source for the 8051 code itself, that would be nice
to have, but I doubt it's something the majority of users are going to
care about. I've certainly never seen any.
 
I totally agree on all points.  But I hate blobs.  If the source code 
is around, I'd love to have it recorded somewhere even though we all 
know that Kbuild and most users won't know what to do with it.  Just 
In Case.

Do we even know where this blob came from originally?
quoted
On that note, do you have a pointer to some of the various ALSA
implementations that are floating around? I've only seen the OSS driver
we had in CVS, and that provides a rather dysmal starting point..
 
Whaley sent me one just last week (hi Robert!).  Given that, I don't 
suppose he'd mind my forwarding the patch on.

Not at all.  BTW: we use this (moderately buggy) compiler to compile 
the 8051 code:

http://sourceforge.net/project/showfiles.php?group_id=599

Between the bugs in the compiler and the absence of documentation in 
the SM501 manual about 8051 interrupts, etc.   It was pretty difficult 
making this work.

Don't try anything remotely like this with this compiler:

u8 a, b;
u32 c;

   c |= a << b;
WTF?!    :)

And here I was thinking that SDCC was probably pretty good...

quoted hunk ↗ jump to hunk
+
+/* There are 2 copies of shared because the shared SRAM is not safe if
+ * the 8051 is reading/writing to a location while the CPU is
+ * writing/reading the same location (according to the SM501 manual
+ * and verified by tests).  So 2 redundant copies are maintained and
+ * for the contents to be considered valid the same data must be
+ * present in both copies.
+ */
 
Classic SM501.  "Here's a feature that works.... but it doesn't!"  :)


b.g.

-- 
Bill Gatliff
bgat@billgatliff.com


-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help