[linux-sunxi] Re: [RFC PATCH 2/5] crypto: sunxi-ss: fix min3() call to match types
On Tuesday 22 December 2015, Andre Przywara wrote: > The min3() macro expects all arguments to be of the same type (or > size at least). While two arguments are ints or u32s, one is size_t, > which does not match on 64-bit architectures. > Cast the size_t to u32 to make min3() happy. In this context here the > length should never exceed 32 bits anyway. > > Signed-off-by: Andre Przywara Looks correct, but a bit ugly. Could we avoid the casts by using temporary variables to keep the size_t based data? Arnd
[linux-sunxi] Re: [RFC PATCH 2/5] crypto: sunxi-ss: fix min3() call to match types
Hi Arnd, On 12/22/15 22:06, Arnd Bergmann wrote: > On Tuesday 22 December 2015, Andre Przywara wrote: >> The min3() macro expects all arguments to be of the same type (or >> size at least). While two arguments are ints or u32s, one is size_t, >> which does not match on 64-bit architectures. >> Cast the size_t to u32 to make min3() happy. In this context here the >> length should never exceed 32 bits anyway. >> >> Signed-off-by: Andre Przywara > > Looks correct, but a bit ugly. Could we avoid the casts by using > temporary variables to keep the size_t based data? I guess this gets even uglier, but I found a better solution by promoting the other involved variables to size_t in this function. This works nicely for most of the cases, I just need two size_t casts now. Will send an updated version soon. Cheers, Andre. -- You received this message because you are subscribed to the Google Groups "linux-sunxi" group. To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[linux-sunxi] Re: [RFC PATCH 2/5] crypto: sunxi-ss: fix min3() call to match types
On Tue, Dec 22, 2015 at 12:27:44PM +, Andre Przywara wrote: > The min3() macro expects all arguments to be of the same type (or > size at least). While two arguments are ints or u32s, one is size_t, > which does not match on 64-bit architectures. > Cast the size_t to u32 to make min3() happy. In this context here the > length should never exceed 32 bits anyway. > > Signed-off-by: Andre Przywara > --- > drivers/crypto/sunxi-ss/sun4i-ss-cipher.c | 12 ++-- > drivers/crypto/sunxi-ss/sun4i-ss-hash.c | 8 > 2 files changed, 10 insertions(+), 10 deletions(-) > > diff --git a/drivers/crypto/sunxi-ss/sun4i-ss-cipher.c > b/drivers/crypto/sunxi-ss/sun4i-ss-cipher.c > index a19ee12..b3bc7bd 100644 > --- a/drivers/crypto/sunxi-ss/sun4i-ss-cipher.c > +++ b/drivers/crypto/sunxi-ss/sun4i-ss-cipher.c > @@ -79,7 +79,7 @@ static int sun4i_ss_opti_poll(struct ablkcipher_request > *areq) > oi = 0; > oo = 0; > do { > - todo = min3(rx_cnt, ileft, (mi.length - oi) / 4); > + todo = min3(rx_cnt, ileft, (u32)(mi.length - oi) / 4); For this case the min function has a min_t variant to specify the argument. What about introducing min3_t? BTW, I don't understand why min3(x, y, z) isn't just defined as #define min3(x, y, z) min(min(x, y), z) but instead as: #define min3(x, y, z) min((typeof(x))min(x, y), z) . I thought min(x, y) has the same type as x anyhow? Best regards Uwe -- Pengutronix e.K. | Uwe Kleine-König| Industrial Linux Solutions | http://www.pengutronix.de/ | -- You received this message because you are subscribed to the Google Groups "linux-sunxi" group. To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.