On 22/02/17 01:12, Russell King - ARM Linux wrote:

On Tue, Feb 21, 2017 at 07:10:11PM +0000, James Morse wrote:

quoted

Hi Tyler,

On 15/02/17 19:51, Tyler Baicar wrote:

quoted

+	} else {
+		const void *unknown_err;
+
+		unknown_err = acpi_hest_generic_data_payload(gdata);
+		printk("%ssection type: %pUl\n", newpfx, sec_type);
+		printk("%ssection length: %d\n", newpfx,

Nit: please use the "%s""section... that this file consistently uses. This means
this code will still work as expected when someone adds '%ss' support to printk!

No.  That is wrong:

"%s""section" is stored in memory as bytes containing:

'%' 's' 's' 'e' 'c' 't' 'i' 'o' 'n'

whereas "%ssection" is stored in memory as bytes containing:

'%' 's' 's' 'e' 'c' 't' 'i' 'o' 'n'

They're exactly the same, so when printk() comes to parse the string, it
sees exactly the same byte sequence.  So, the only thing that's happening
is code obfuscation for no good reason what so ever.

If you don't believe me, run some build tests and look at the resulting
strings... also look at the C standard.  "Adjacent string literal tokens
are concatenated."

Please get rid of this obfuscation.

Sure, I was always told not do this, clearly I didn't think about it for very long!

This file otherwise consistently uses the now-weird "%s""otherstring" pattern.



Thanks,

James