lkml.org 
[lkml]   [2019]   [May]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [PATCH 02/16] treewide: rename match_string() -> __match_string()
Date
Am Mittwoch, 8. Mai 2019, 19:00:06 CEST schrieb Stephen Boyd:
> (Trimming the lists but keeping lkml)
>
> Quoting Alexandru Ardelean (2019-05-08 04:28:28)
> > This change does a rename of match_string() -> __match_string().
> >
> > There are a few parts to the intention here (with this change):
> > 1. Align with sysfs_match_string()/__sysfs_match_string()
> > 2. This helps to group users of `match_string()` into simple users:
> > a. those that use ARRAY_SIZE(_a) to specify the number of elements
> > b. those that use -1 to pass a NULL terminated array of strings
> > c. special users, which (after eliminating 1 & 2) are not that many
> > 3. The final intent is to fix match_string()/__match_string() which is
> > slightly broken, in the sense that passing -1 or a positive value does
> > not make any difference: the iteration will stop at the first NULL
> > element.
> >
> > Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
> > ---
> [...]
> > diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
> > index 96053a96fe2f..0b6c3d300411 100644
> > --- a/drivers/clk/clk.c
> > +++ b/drivers/clk/clk.c
> > @@ -2305,8 +2305,8 @@ bool clk_has_parent(struct clk *clk, struct clk *parent)
> > if (core->parent == parent_core)
> > return true;
> >
> > - return match_string(core->parent_names, core->num_parents,
> > - parent_core->name) >= 0;
> > + return __match_string(core->parent_names, core->num_parents,
> > + parent_core->name) >= 0;
>
> This is essentially ARRAY_SIZE(core->parent_names) so it should be fine
> to put this back to match_string() later in the series.
>
> > }
> > EXPORT_SYMBOL_GPL(clk_has_parent);
> >
> > diff --git a/drivers/clk/rockchip/clk.c b/drivers/clk/rockchip/clk.c
> > index c3ad92965823..373f13e9cd83 100644
> > --- a/drivers/clk/rockchip/clk.c
> > +++ b/drivers/clk/rockchip/clk.c
> > @@ -276,8 +276,8 @@ static struct clk *rockchip_clk_register_frac_branch(
> > struct clk *mux_clk;
> > int ret;
> >
> > - frac->mux_frac_idx = match_string(child->parent_names,
> > - child->num_parents, name);
> > + frac->mux_frac_idx = __match_string(child->parent_names,
> > + child->num_parents, name);
>
> I suspect this is the same as above, but Heiko can ack/confirm.

Right now, clock-driver do carry the parent_names in static arrays,
but that isn't necessarily guaranteed. So only in the driver declaring
the array do we use ARRAY_SIZE to determine the number of elements
(from inside the helper-macros in drivers/clk/rockchip/clk.h), but the
rest of the code does explicitly carry around the real number of elements
so later clocks could in theory also do that dynamically.


Heiko


\
 
 \ /
  Last update: 2019-05-09 11:03    [W:0.057 / U:2.524 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site