Thread (1 message) 1 message, 1 author, 2009-12-09

Re: [PATCH v2] add the driver for Analog Devices Blackfin on-chip CAN controllers

From: Wolfgang Grandegger <hidden>
Date: 2009-12-09 19:50:47

Barry Song wrote:
On Wed, Dec 9, 2009 at 6:16 PM, Wolfgang Grandegger [off-list ref] wrote:
quoted
Barry Song wrote:
quoted
Signed-off-by: Barry Song <redacted>
Signed-off-by: H.J. Oertel <redacted>
---
      -v2: cleanup according to Wolfgang Grandegger's feedback
      1.delete some unnecessary debug print
      2.delete ndo_tx_timeout entry as it is not needed in can
      3.use alloc_can_skb, alloc_can_err_skb instead of netdev_alloc_skb
      4.add timeout while polling can status
      5.rename BFIN_CAN_READ/WRITE_MSG to bfin_can_read/write_data
      6.use kernel BIT instead of bit shift
      7.use void __iomem * for CAN memory base memory instead of u32
      8.delete "dev->last_rx = jiffies" since it is not needed now
      9.delete redundant "echo_skb" member in bfin can private data
      10.follow can convention to use "_" instead of "-" for file names
Did you consider using structs instead of these ugly and cumbersome
macro functions? This was my major criticisms with your previous version
of the patch.
Wolfgang,
I considered using the structures to describe registers, and I
completely agree with you about the advantages of structures. And I am
sure I will move to structures after some time and send a patch for
it. But for the moment, it is not too urgent, so I only fixed all your
other comments. Anyway, if you insist on using structures in just this
patch, I can fix and send the -v3 patch after finishing it.
Well, "it's not too urgent" is not a real argument for me, sorry. As
Marc pointed out, the current "#define's" are not acceptable, anyway.
Please invest time in moving to structs right a way.

Thanks,

Wolfgang.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help