On Tuesday, July 29, 2014 at 03:22:29 AM, Scott Wood wrote:
On Tue, 2014-07-29 at 03:11 +0200, Marek Vasut wrote:
On Tuesday, July 29, 2014 at 02:12:47 AM, Scott Wood wrote:
On Tue, 2014-07-29 at 02:09 +0200, Marek Vasut wrote:
On Tuesday, July 29, 2014 at 01:35:57 AM, Scott Wood wrote:
On Tue, 2014-07-22 at 02:34 +0200, Marek Vasut wrote:
Printing u32 with %02x is just a bad idea, fix it.
Why is it "just a bad idea" if the values aren't expected to exceed 0xff?
NAND_CMD_DEPLETE1 is 0x100 for example. I doubt anyone will use AND with this controller, but I'd be much happier to see the print properly matching the variable we're printing.
It will match it. %02x doesn't restrict the output to two characters; it just makes sure there are at least two characters.
The output with %02x in this case is 0xffffffXY , so there is something really wrong going on. With %08x, the output is as expected (because that does match ths size of the variable).
Is that for NAND_CMD_NONE which is -1? Or something else?
OK, now it makes sense.
Also regarding the changelog, neither of these values are declared as u32.
They're int and unsigned int, which on ARMv7 and lower (this is armv4) is 32bit with GCC.
Best regards, Marek Vasut