On 20/10/2019 17:55, Luc Van Oostenryck wrote:
On Wed, Sep 25, 2019 at 11:00:13AM +0100, Ben Dooks wrote:+ if (*ptr == 'x' || *ptr == 'X' || *ptr == 'u' || *ptr == 'o') { + ptr++; + type->test = printf_fmt_numtype; + switch (szmod) { + case -1: + type->data = &ushort_ctype; + break; + case 0: + type->data = &uint_ctype; + break; + case 1: + type->data = &ulong_ctype; + break; + case 2: + type->data = &ullong_ctype; + break; + default: + type->test = NULL; + } + } else if (*ptr == 'i' || *ptr == 'd') { + ptr++; + type->test = printf_fmt_numtype; + switch (szmod) { + case -1: + type->data = &short_ctype; + break; + case 0: + type->data = &int_ctype; + break; + case 1: + type->data = &long_ctype; + break; + case 2: + type->data = &llong_ctype; + break; + default: + type->test = NULL; + }When testing this on the kernel, I've a bunch of warning: incorrect type in argument . (different types) expected unsigned int got int This will quickly be quite annoying. I've also a bunch of: warning: incorrect type in argument . (different types) expected int got int but I can't investigate more for now.
I might spend some time fixing those up... there's plenty of warning fixes to go through. I've not yet run my test on a current kernel, but printf type seems to be one place where people have been a bit sloppy. Thanks! -- Ben Dooks http://www.codethink.co.uk/ Senior Engineer Codethink - Providing Genius https://www.codethink.co.uk/privacy.html
