Jesper Juhl wrote:
On 12/06/06, Jim Cromie <jim.cromie@xxxxxxxxx> wrote:
Add a platform-device, then use it for dev_dbg(), once, just to prove it
works.
[snip]
+ /* should I export-gpl the pci-device available in scx200,
+ and use it instead of getting my own platform/isa device ?
+ */
I'd suggest finding that out and correct it before submitting for
inclusion instead of just leaving the question in a comment.
OK - Anyone ? Pls consider, restated as:
dev_dbg() needs a dev,
one is available from struct platform_device *pdev; pdev->dev,
another is in scx200.c:
- upon which scx200_gpio already depends.
could reach inside scx200.c for it,
- maybe too interconnected..
To complicate this further, theres patch 18/20,
which addresses the same problem, the same way, for the other driver.
The embedded issue is that the devp is needed by dev_dbg() in nsc_gpio.ko,
so that the common code can give useful user-support via its logging.
[snip]
+ // what about chrdev cleanup ?
Use C comments, please, not C++ comments.
old bad habit.
I now tend to use it partly to signal to reader that its not-for-inclusion.
Perhaps this new habit is also unhelpful, since it seems to draw attention,
.. you cant know that I know...
WRT the Q, is there anything to undo if init fails after cdev_init()s ?
LDD3 comment /* Fail gracefully if need be */
suggests not, and I didnt see anything that fits this shutdown-ish behavior,
but I could have missed it.
continued thanks,
Jim Cromie
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive: http://mail.nl.linux.org/kernelnewbies/
FAQ: http://kernelnewbies.org/faq/
