From: Bjorn Andersson <bjorn.andersson@linaro.org>
To: Ansuel Smith <ansuelsmth@gmail.com>
Cc: Andy Gross <agross@kernel.org>,
Michael Turquette <mturquette@baylibre.com>,
Stephen Boyd <sboyd@kernel.org>, Rob Herring <robh+dt@kernel.org>,
Philipp Zabel <p.zabel@pengutronix.de>,
Taniya Das <tdas@codeaurora.org>,
linux-arm-msm@vger.kernel.org, linux-clk@vger.kernel.org,
devicetree@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v4 05/16] clk: qcom: gcc-ipq806x: convert parent_names to parent_data
Date: Thu, 24 Feb 2022 08:08:43 -0800 [thread overview]
Message-ID: <YhetizewLY5SqcJl@ripper> (raw)
In-Reply-To: <Yhen5cLB32wGmhxu@Ansuel-xps.localdomain>
On Thu 24 Feb 07:45 PST 2022, Ansuel Smith wrote:
> On Wed, Feb 23, 2022 at 09:48:45PM -0600, Bjorn Andersson wrote:
> > On Thu 17 Feb 17:56 CST 2022, Ansuel Smith wrote:
> >
> > > Convert parent_names to parent_data to modernize the driver.
> > > Where possible use parent_hws directly.
> > >
> >
> > Really nice to see this kind of cleanup. Unfortunately I have two
> > comments below.
> >
> > > Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
> > > ---
> > > drivers/clk/qcom/gcc-ipq806x.c | 286 ++++++++++++++++++++-------------
> > > 1 file changed, 173 insertions(+), 113 deletions(-)
> > >
> > > diff --git a/drivers/clk/qcom/gcc-ipq806x.c b/drivers/clk/qcom/gcc-ipq806x.c
> > > index 34cddf461dba..828383c30322 100644
> > > --- a/drivers/clk/qcom/gcc-ipq806x.c
> > > +++ b/drivers/clk/qcom/gcc-ipq806x.c
> > > @@ -25,6 +25,10 @@
> > > #include "clk-hfpll.h"
> > > #include "reset.h"
> > >
> > > +static const struct clk_parent_data gcc_pxo[] = {
> > > + { .fw_name = "pxo" },
> >
> > I expect that this will break booting these boards with existing dtb,
> > because there's not yet a clocks <&pxo_board> in the gcc node.
> >
>
> Considering the lack of device using ipq806x in the kernel and the fact
> that we add the clocks to the global dtsi should we care? The breakage
> will be present on boards that use custom kernel anyway so in theory
> shouldn't be that hard to refresh the dtsi.
>
> > If you also add .name = "pxo" here that it should still fall back to map
> > to the board clock registered in gcc_ipq806x_probe() and once we have
> > passed 1-2 kernel releases we can clean out the old mapping.
> >
>
> Just to make sure, you are suggesting to put 2 entry (fw_name AND name)
> or replace the fw_name with the generic name variable?
>
What you have (.fw_name = "pxo") is perfect looking forward, but if
nothing else the clock and dts drivers are merged through different
paths up to Torvalds so merging the set as is in a single go might
actually break things for a while, even for you.
So if we go { .fw_name = "pxo", .name = "pxo" } we should handle both
the new and old dts. And I'm fine with saying that as soon as we see the
dts change landed in a release we drop the .name - if there aren't users
mixing and matching kernel and dtbs.
> Anyway thanks for the review!
>
Thank you for the nice work!
Regards,
Bjorn
> > > +};
> > > +
> > > static struct clk_pll pll0 = {
> > > .l_reg = 0x30c4,
> > > .m_reg = 0x30c8,
> > > @@ -35,7 +39,7 @@ static struct clk_pll pll0 = {
> > > .status_bit = 16,
> > > .clkr.hw.init = &(struct clk_init_data){
> > > .name = "pll0",
> > > - .parent_names = (const char *[]){ "pxo" },
> > > + .parent_data = gcc_pxo,
> > > .num_parents = 1,
> > > .ops = &clk_pll_ops,
> > > },
> > > @@ -46,7 +50,9 @@ static struct clk_regmap pll0_vote = {
> > > .enable_mask = BIT(0),
> > > .hw.init = &(struct clk_init_data){
> > > .name = "pll0_vote",
> > > - .parent_names = (const char *[]){ "pll0" },
> > > + .parent_hws = (const struct clk_hw*[]){
> > > + &pll0.clkr.hw,
> > > + },
> > > .num_parents = 1,
> > > .ops = &clk_pll_vote_ops,
> > > },
> > > @@ -62,7 +68,7 @@ static struct clk_pll pll3 = {
> > > .status_bit = 16,
> > > .clkr.hw.init = &(struct clk_init_data){
> > > .name = "pll3",
> > > - .parent_names = (const char *[]){ "pxo" },
> > > + .parent_data = gcc_pxo,
> > > .num_parents = 1,
> > > .ops = &clk_pll_ops,
> > > },
> > > @@ -89,7 +95,7 @@ static struct clk_pll pll8 = {
> > > .status_bit = 16,
> > > .clkr.hw.init = &(struct clk_init_data){
> > > .name = "pll8",
> > > - .parent_names = (const char *[]){ "pxo" },
> > > + .parent_data = gcc_pxo,
> > > .num_parents = 1,
> > > .ops = &clk_pll_ops,
> > > },
> > > @@ -100,7 +106,9 @@ static struct clk_regmap pll8_vote = {
> > > .enable_mask = BIT(8),
> > > .hw.init = &(struct clk_init_data){
> > > .name = "pll8_vote",
> > > - .parent_names = (const char *[]){ "pll8" },
> > > + .parent_hws = (const struct clk_hw*[]){
> > > + &pll8.clkr.hw,
> > > + },
> > > .num_parents = 1,
> > > .ops = &clk_pll_vote_ops,
> > > },
> > > @@ -123,7 +131,7 @@ static struct hfpll_data hfpll0_data = {
> > > static struct clk_hfpll hfpll0 = {
> > > .d = &hfpll0_data,
> > > .clkr.hw.init = &(struct clk_init_data){
> > > - .parent_names = (const char *[]){ "pxo" },
> > > + .parent_data = gcc_pxo,
> > > .num_parents = 1,
> > > .name = "hfpll0",
> > > .ops = &clk_ops_hfpll,
> > > @@ -149,7 +157,7 @@ static struct hfpll_data hfpll1_data = {
> > > static struct clk_hfpll hfpll1 = {
> > > .d = &hfpll1_data,
> > > .clkr.hw.init = &(struct clk_init_data){
> > > - .parent_names = (const char *[]){ "pxo" },
> > > + .parent_data = gcc_pxo,
> > > .num_parents = 1,
> > > .name = "hfpll1",
> > > .ops = &clk_ops_hfpll,
> > > @@ -175,7 +183,7 @@ static struct hfpll_data hfpll_l2_data = {
> > > static struct clk_hfpll hfpll_l2 = {
> > > .d = &hfpll_l2_data,
> > > .clkr.hw.init = &(struct clk_init_data){
> > > - .parent_names = (const char *[]){ "pxo" },
> > > + .parent_data = gcc_pxo,
> > > .num_parents = 1,
> > > .name = "hfpll_l2",
> > > .ops = &clk_ops_hfpll,
> > > @@ -194,7 +202,7 @@ static struct clk_pll pll14 = {
> > > .status_bit = 16,
> > > .clkr.hw.init = &(struct clk_init_data){
> > > .name = "pll14",
> > > - .parent_names = (const char *[]){ "pxo" },
> > > + .parent_data = gcc_pxo,
> > > .num_parents = 1,
> > > .ops = &clk_pll_ops,
> > > },
> > > @@ -205,7 +213,9 @@ static struct clk_regmap pll14_vote = {
> > > .enable_mask = BIT(14),
> > > .hw.init = &(struct clk_init_data){
> > > .name = "pll14_vote",
> > > - .parent_names = (const char *[]){ "pll14" },
> > > + .parent_hws = (const struct clk_hw*[]){
> > > + &pll14.clkr.hw,
> > > + },
> > > .num_parents = 1,
> > > .ops = &clk_pll_vote_ops,
> > > },
> > > @@ -238,7 +248,7 @@ static struct clk_pll pll18 = {
> > > .freq_tbl = pll18_freq_tbl,
> > > .clkr.hw.init = &(struct clk_init_data){
> > > .name = "pll18",
> > > - .parent_names = (const char *[]){ "pxo" },
> > > + .parent_data = gcc_pxo,
> > > .num_parents = 1,
> > > .ops = &clk_pll_ops,
> > > },
> > > @@ -259,9 +269,9 @@ static const struct parent_map gcc_pxo_pll8_map[] = {
> > > { P_PLL8, 3 }
> > > };
> > >
> > > -static const char * const gcc_pxo_pll8[] = {
> > > - "pxo",
> > > - "pll8_vote",
> > > +static const struct clk_parent_data gcc_pxo_pll8[] = {
> > > + { .fw_name = "pxo" },
> > > + { .hw = &pll8_vote.hw },
> > > };
> > >
> > > static const struct parent_map gcc_pxo_pll8_cxo_map[] = {
> > > @@ -270,10 +280,10 @@ static const struct parent_map gcc_pxo_pll8_cxo_map[] = {
> > > { P_CXO, 5 }
> > > };
> > >
> > > -static const char * const gcc_pxo_pll8_cxo[] = {
> > > - "pxo",
> > > - "pll8_vote",
> > > - "cxo",
> > > +static const struct clk_parent_data gcc_pxo_pll8_cxo[] = {
> > > + { .fw_name = "pxo" },
> > > + { .hw = &pll8_vote.hw },
> > > + { .fw_name = "cxo" },
> >
> > As with "pxo", I think you need a .name = "cxo" here as well.
> >
> > Regards,
> > Bjorn
>
> --
> Ansuel
next prev parent reply other threads:[~2022-02-24 16:08 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-02-17 23:56 [PATCH v4 00/16] Multiple addition and improvement to ipq8064 gcc Ansuel Smith
2022-02-17 23:56 ` [PATCH v4 01/16] dt-bindings: clock: split qcom,gcc.yaml to common and specific schema Ansuel Smith
2022-02-24 3:38 ` Bjorn Andersson
2022-02-17 23:56 ` [PATCH v4 02/16] dt-bindings: clock: simplify qcom,gcc-apq8064 Documentation Ansuel Smith
2022-02-24 3:41 ` Bjorn Andersson
2022-02-17 23:56 ` [PATCH v4 03/16] dt-bindings: clock: Document qcom,gcc-ipq8064 binding Ansuel Smith
2022-02-24 3:42 ` Bjorn Andersson
2022-02-17 23:56 ` [PATCH v4 04/16] clk: qcom: gcc-ipq806x: fix wrong naming for gcc_pxo_pll8_pll0 Ansuel Smith
2022-02-24 3:43 ` Bjorn Andersson
2022-02-17 23:56 ` [PATCH v4 05/16] clk: qcom: gcc-ipq806x: convert parent_names to parent_data Ansuel Smith
2022-02-24 3:48 ` Bjorn Andersson
2022-02-24 15:45 ` Ansuel Smith
2022-02-24 16:08 ` Bjorn Andersson [this message]
2022-02-17 23:56 ` [PATCH v4 06/16] clk: qcom: gcc-ipq806x: use ARRAY_SIZE for num_parents Ansuel Smith
2022-02-24 3:49 ` Bjorn Andersson
2022-02-17 23:56 ` [PATCH v4 07/16] clk: qcom: gcc-ipq806x: drop hardcoded pxo and cxo source clk Ansuel Smith
2022-02-24 3:50 ` Bjorn Andersson
2022-02-24 15:50 ` Ansuel Smith
2022-02-24 16:15 ` Bjorn Andersson
2022-02-17 23:56 ` [PATCH v4 08/16] clk: qcom: gcc-ipq806x: add additional freq nss cores Ansuel Smith
2022-02-24 3:55 ` Bjorn Andersson
2022-02-24 15:55 ` Ansuel Smith
2022-02-24 16:17 ` Bjorn Andersson
2022-02-17 23:56 ` [PATCH v4 09/16] clk: qcom: gcc-ipq806x: add unusued flag for critical clock Ansuel Smith
2022-02-17 23:56 ` [PATCH v4 10/16] clk: qcom: clk-rcg: add clk_rcg_floor_ops ops Ansuel Smith
2022-02-24 3:58 ` Bjorn Andersson
2022-02-17 23:56 ` [PATCH v4 11/16] clk: qcom: gcc-ipq806x: add additional freq for sdc table Ansuel Smith
2022-02-17 23:56 ` [PATCH v4 12/16] dt-bindings: clock: add ipq8064 ce5 clk define Ansuel Smith
2022-02-24 4:01 ` Bjorn Andersson
2022-02-24 16:01 ` Ansuel Smith
2022-02-24 16:18 ` Bjorn Andersson
2022-02-17 23:57 ` [PATCH v4 13/16] clk: qcom: gcc-ipq806x: add CryptoEngine clocks Ansuel Smith
2022-02-17 23:57 ` [PATCH v4 14/16] dt-bindings: reset: add ipq8064 ce5 resets Ansuel Smith
2022-02-24 4:01 ` Bjorn Andersson
2022-02-24 16:31 ` Philipp Zabel
2022-02-17 23:57 ` [PATCH v4 15/16] clk: qcom: gcc-ipq806x: add CryptoEngine resets Ansuel Smith
2022-02-17 23:57 ` [PATCH v4 16/16] ARM: dts: qcom: Add syscon and cxo/pxo clock to gcc node for ipq8064 Ansuel Smith
2022-02-24 4:08 ` Bjorn Andersson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YhetizewLY5SqcJl@ripper \
--to=bjorn.andersson@linaro.org \
--cc=agross@kernel.org \
--cc=ansuelsmth@gmail.com \
--cc=devicetree@vger.kernel.org \
--cc=linux-arm-msm@vger.kernel.org \
--cc=linux-clk@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mturquette@baylibre.com \
--cc=p.zabel@pengutronix.de \
--cc=robh+dt@kernel.org \
--cc=sboyd@kernel.org \
--cc=tdas@codeaurora.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).