[U-Boot] [PATCH 6/6] linux/kernel.h: sync min, max, min3, max3 macros with Linux

[Masahiro Yamada <yamada.m@jp.panasonic.com>]

Hi Pavel,

Thanks for your close checking.


On Tue, 4 Nov 2014 20:50:13 +0100
Pavel Machek <pavel@denx.de> wrote:

> On Tue 2014-11-04 20:26:26, Masahiro Yamada wrote:
> > U-Boot has never cared about the type when we get max/min of two
> > values, but Linux Kernel does.  This commit gets min, max, min3, max3
> > macros synced with the kernel introduing type checks.
> 
> "introducing"


I will fix this.



> > Many of references of those macros must be fixed to suppress warnings.
> > We have two options:
> >  - Use min, max, min3, max3 only when the arguments have the same type
> >    (or add casts to the arguments)
> >  - Use min_t/max_t instead with the appropriate type for the first
> >    argument
> > 
> > Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
> 
> Acked-by: Pavel Machek <pavel@denx.de>
> 
> [I guess the conversion is okay, as is, still, there are places where
> code could be cleaned up afterwards...]

Fully agreed.
In some places, I think we should change the variable types as you suggested
but it is beyond the motivation of this series.





> > -		for (postdiv = 1; postdiv <= min(div, MAX_POSTDIV); postdiv++) {
> > +		for (postdiv = 1; postdiv <= min(div, (unsigned long)MAX_POSTDIV);
> > +								postdiv++) {
> 
> It might be cleaner to change MAX_POSTDIV definition to include UL?


Sounds good to me.



> 
> > @@ -1838,12 +1838,18 @@ static void program_tr(unsigned long *dimm_populated,
> >  			else
> >  				sdram_ddr1 = false;
> >  
> > -			t_rcd_ns = max(t_rcd_ns, spd_read(iic0_dimm_addr[dimm_num], 29) >> 2);
> > -			t_rrd_ns = max(t_rrd_ns, spd_read(iic0_dimm_addr[dimm_num], 28) >> 2);
> > -			t_rp_ns  = max(t_rp_ns,  spd_read(iic0_dimm_addr[dimm_num], 27) >> 2);
> > -			t_ras_ns = max(t_ras_ns, spd_read(iic0_dimm_addr[dimm_num], 30));
> > -			t_rc_ns  = max(t_rc_ns,  spd_read(iic0_dimm_addr[dimm_num], 41));
> > -			t_rfc_ns = max(t_rfc_ns, spd_read(iic0_dimm_addr[dimm_num], 42));
> > +			t_rcd_ns = max(t_rcd_ns,
> > +				       (unsigned long)spd_read(iic0_dimm_addr[dimm_num], 29) >> 2);
> > +			t_rrd_ns = max(t_rrd_ns,
> > +				       (unsigned long)spd_read(iic0_dimm_addr[dimm_num], 28) >> 2);
> > +			t_rp_ns  = max(t_rp_ns,
> > +				       (unsigned long)spd_read(iic0_dimm_addr[dimm_num], 27) >> 2);
> > +			t_ras_ns = max(t_ras_ns,
> > +				       (unsigned long)spd_read(iic0_dimm_addr[dimm_num], 30));
> > +			t_rc_ns  = max(t_rc_ns,
> > +				       (unsigned long)spd_read(iic0_dimm_addr[dimm_num], 41));
> > +			t_rfc_ns = max(t_rfc_ns,
> > +				       (unsigned long)spd_read(iic0_dimm_addr[dimm_num], 42));
> >  		}
> >  	}
> 
> Would it be feasible to make spd_read return unsigned long?


I am not familiar enough with this code
to judge this.




> > diff --git a/arch/sandbox/cpu/start.c b/arch/sandbox/cpu/start.c
> > index b3d7051..b7e12ab 100644
> > --- a/arch/sandbox/cpu/start.c
> > +++ b/arch/sandbox/cpu/start.c
> > @@ -38,7 +38,7 @@ int sandbox_early_getopt_check(void)
> >  
> >  	max_arg_len = 0;
> >  	for (i = 0; i < num_options; ++i)
> > -		max_arg_len = max(strlen(sb_opt[i]->flag), max_arg_len);
> > +		max_arg_len = max((int)strlen(sb_opt[i]->flag), max_arg_len);
> >  	max_noarg_len = max_arg_len + 7;
> >  
> >  	for (i = 0; i < num_options; ++i) {
> 
> make max_arg_len size_t?
> 
> > diff --git a/drivers/net/netconsole.c b/drivers/net/netconsole.c
> > index 623f749..677c89f 100644
> > --- a/drivers/net/netconsole.c
> > +++ b/drivers/net/netconsole.c
> > @@ -256,7 +256,7 @@ static void nc_puts(struct stdio_dev *dev, const char *s)
> >  
> >  	len = strlen(s);
> >  	while (len) {
> > -		int send_len = min(len, sizeof(input_buffer));
> > +		int send_len = min(len, (int)sizeof(input_buffer));
> >  		nc_send_packet(s, send_len);
> >  		len -= send_len;
> >  		s += send_len;
> 
> Looks like len/send_len wants to be size_t here.
> 
> Actually, I'd argue that anytime you need to explicitly cast one
> argument would be good time to use _t variant... but that would mean
> redoing rather big patch.
> 
> min((int) a, b) -> min_t(int, a, b).


Agreed but I think we should do this
after we understand the code well.



Best Regards
Masahiro Yamada