linux-kernel - Re: [PATCH v5 5/9] clk: microchip: mpfs: use regmap for clocks

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20251024-vocalist-irritable-dcac9f7db3ce@spud>
Date: Fri, 24 Oct 2025 11:30:08 +0100
From: Conor Dooley <conor@...nel.org>
To: Claudiu Beznea <claudiu.beznea@...on.dev>
Cc: Conor Dooley <conor.dooley@...rochip.com>,
	Daire McNamara <daire.mcnamara@...rochip.com>,
	pierre-henry.moussay@...rochip.com,
	valentina.fernandezalanis@...rochip.com,
	Michael Turquette <mturquette@...libre.com>,
	Stephen Boyd <sboyd@...nel.org>, Rob Herring <robh@...nel.org>,
	Krzysztof Kozlowski <krzk+dt@...nel.org>,
	Philipp Zabel <p.zabel@...gutronix.de>,
	linux-riscv@...ts.infradead.org, linux-clk@...r.kernel.org,
	devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
	Brian Masney <bmasney@...hat.com>
Subject: Re: [PATCH v5 5/9] clk: microchip: mpfs: use regmap for clocks

On Fri, Oct 24, 2025 at 11:20:23AM +0100, Conor Dooley wrote:
> On Thu, Oct 23, 2025 at 07:06:01AM +0300, Claudiu Beznea wrote:
> > On 10/13/25 20:45, Conor Dooley wrote:
> > > From: Conor Dooley <conor.dooley@...rochip.com>
> > > +static int mpfs_cfg_clk_set_rate(struct clk_hw *hw, unsigned long rate, unsigned long prate)
> > > +{
> > > +	struct mpfs_cfg_hw_clock *cfg_hw = to_mpfs_cfg_clk(hw);
> > > +	struct mpfs_cfg_clock *cfg = &cfg_hw->cfg;
> > > +	unsigned long flags;
> > > +	u32 val;
> > > +	int divider_setting;
> > 
> > This could be moved near flags to keep the reverse christmas tree order as
> > in the rest of this patch.
> 
> The driver doesn't (intentionally) use reverse christmas tree. If it
> does, that's just a byproduct of putting bigger types before smaller
> ones.
> 
> > > +	divider_setting = divider_get_val(rate, prate, cfg->table, cfg->width, 0);
> > > +
> > > +	if (divider_setting < 0)
> > > +		return divider_setting;
> > > +
> > > +	spin_lock_irqsave(&mpfs_clk_lock, flags);
> > 
> > As spin locking is introduced in this file by this patch, you can go
> > directly w/ cleanup helpers for locking.

Actually, if using regmap_update_bits(), locking might be entirely
removable. Depends on if there's some rmw sequence that can't move to
that.

> > 
> > > +
> > > +	regmap_read(cfg->map, cfg->map_offset, &val);
> > > +	val &= ~(clk_div_mask(cfg->width) << cfg_hw->cfg.shift);
> > 
> > Why cfg_hw->cfg.shift here --------------------^ but cfg->shift on the next
> > line?
> > 
> > > +	val |= divider_setting << cfg->shift;
> > > +	regmap_write(cfg->map, cfg->map_offset, val);
> > 
> > Can't the regmap_read() + updated + regmap_write() be replaced by
> > regmap_update_bits() ?
> 
> Yeah, I suppose it could. Ultimately what's here is a revert of with
> readl()/writel() replaced by regmap operations directly, so the answer
> to the above three items is that that's how they were done before the
> patch I am reverting. That's probably the answer to 90% of the things
> you've said here, this is how they were done prior to the commit I am
> reverting.
> 



Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)