From: Varadarajan Narayanan <quic_varada@quicinc.com>
To: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Cc: "Alex G." <mr.nuke.me@gmail.com>,
"Bjorn Andersson" <andersson@kernel.org>,
"Konrad Dybcio" <konrad.dybcio@linaro.org>,
"Lorenzo Pieralisi" <lpieralisi@kernel.org>,
"Krzysztof Wilczyński" <kw@linux.com>,
"Rob Herring" <robh@kernel.org>,
"Bjorn Helgaas" <bhelgaas@google.com>,
"Krzysztof Kozlowski" <krzk+dt@kernel.org>,
"Conor Dooley" <conor+dt@kernel.org>,
"Vinod Koul" <vkoul@kernel.org>,
"Kishon Vijay Abraham I" <kishon@kernel.org>,
"Michael Turquette" <mturquette@baylibre.com>,
"Stephen Boyd" <sboyd@kernel.org>,
"Manivannan Sadhasivam" <manivannan.sadhasivam@linaro.org>,
linux-arm-msm@vger.kernel.org, linux-pci@vger.kernel.org,
devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-phy@lists.infradead.org, linux-clk@vger.kernel.org
Subject: Re: [PATCH v3 6/7] phy: qcom-qmp-pcie: add support for ipq9574 gen3x2 PHY
Date: Mon, 29 Apr 2024 11:50:13 +0530 [thread overview]
Message-ID: <Zi88Hdx6UgBo/gti@hu-varada-blr.qualcomm.com> (raw)
In-Reply-To: <CAA8EJppGW=qyk2P6Z_S=dp0njsCjqZaXjw8qU4MY1HOZR-N=4A@mail.gmail.com>
On Wed, Apr 17, 2024 at 12:50:49AM +0300, Dmitry Baryshkov wrote:
> On Wed, 17 Apr 2024 at 00:25, Alex G. <mr.nuke.me@gmail.com> wrote:
> >
> > Hi Dmitry,
> >
> > On 4/15/24 16:25, mr.nuke.me@gmail.com wrote:
> > >
> > >
> > > On 4/15/24 15:10, Dmitry Baryshkov wrote:
> > >> On Mon, 15 Apr 2024 at 21:23, Alexandru Gagniuc <mr.nuke.me@gmail.com>
> > >> wrote:
> > >>>
> > >>> Add support for the gen3x2 PCIe PHY on IPQ9574, ported form downstream
> > >>> 5.4 kernel. Only the serdes and pcs_misc tables are new, the others
> > >>> being reused from IPQ8074 and IPQ6018 PHYs.
> > >>>
> > >>> Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
> > >>> ---
> > >>> drivers/phy/qualcomm/phy-qcom-qmp-pcie.c | 136 +++++++++++++++++-
> > >>> .../phy/qualcomm/phy-qcom-qmp-pcs-pcie-v5.h | 14 ++
> > >>> 2 files changed, 149 insertions(+), 1 deletion(-)
> > >>>
> > >>
> > >> [skipped]
> > >>
> > >>> @@ -2448,7 +2542,7 @@ static inline void qphy_clrbits(void __iomem
> > >>> *base, u32 offset, u32 val)
> > >>>
> > >>> /* list of clocks required by phy */
> > >>> static const char * const qmp_pciephy_clk_l[] = {
> > >>> - "aux", "cfg_ahb", "ref", "refgen", "rchng", "phy_aux",
> > >>> + "aux", "cfg_ahb", "ref", "refgen", "rchng", "phy_aux",
> > >>> "anoc", "snoc"
> > >>
> > >> Are the NoC clocks really necessary to drive the PHY? I think they are
> > >> usually connected to the controllers, not the PHYs.
> > >
> > > The system will hang if these clocks are not enabled. They are also
> > > attached to the PHY in the QCA 5.4 downstream kernel.
>
> Interesting.
> I see that Varadarajan is converting these clocks into interconnects.
> Maybe it's better to wait for those patches to land and use
> interconnects instead. I think it would better suit the
> infrastructure.
>
> Varadarajan, could you please comment, are these interconnects
> connected to the PHY too or just to the PCIe controller?
Sorry for the late response. Missed this e-mail.
These 2 clks are related to AXI port clk on Aggnoc/SNOC, not
directly connected to PCIE wrapper, but it should be enabled to
generate pcie traffic.
Thanks
Varada
> > They are named "anoc_lane", and "snoc_lane" in the downstream kernel.
> > Would you like me to use these names instead?
>
> I'm fine either way.
>
> > e>>> };
> > >>>
> > >>> /* list of regulators */
> > >>> @@ -2499,6 +2593,16 @@ static const struct qmp_pcie_offsets
> > >>> qmp_pcie_offsets_v4x1 = {
> > >>> .rx = 0x0400,
> > >>> };
> > >>>
> > >>> +static const struct qmp_pcie_offsets qmp_pcie_offsets_ipq9574 = {
> > >>> + .serdes = 0,
> > >>> + .pcs = 0x1000,
> > >>> + .pcs_misc = 0x1400,
> > >>> + .tx = 0x0200,
> > >>> + .rx = 0x0400,
> > >>> + .tx2 = 0x0600,
> > >>> + .rx2 = 0x0800,
> > >>> +};
> > >>> +
> > >>> static const struct qmp_pcie_offsets qmp_pcie_offsets_v4x2 = {
> > >>> .serdes = 0,
> > >>> .pcs = 0x0a00,
> > >>> @@ -2728,6 +2832,33 @@ static const struct qmp_phy_cfg
> > >>> sm8250_qmp_gen3x1_pciephy_cfg = {
> > >>> .phy_status = PHYSTATUS,
> > >>> };
> > >>>
> > >>> +static const struct qmp_phy_cfg ipq9574_pciephy_gen3x2_cfg = {
> > >>> + .lanes = 2,
> > >>> +
> > >>> + .offsets = &qmp_pcie_offsets_ipq9574,
> > >>> +
> > >>> + .tbls = {
> > >>> + .serdes = ipq9574_gen3x2_pcie_serdes_tbl,
> > >>> + .serdes_num =
> > >>> ARRAY_SIZE(ipq9574_gen3x2_pcie_serdes_tbl),
> > >>> + .tx = ipq8074_pcie_gen3_tx_tbl,
> > >>> + .tx_num = ARRAY_SIZE(ipq8074_pcie_gen3_tx_tbl),
> > >>> + .rx = ipq6018_pcie_rx_tbl,
> > >>> + .rx_num = ARRAY_SIZE(ipq6018_pcie_rx_tbl),
> > >>> + .pcs = ipq6018_pcie_pcs_tbl,
> > >>> + .pcs_num = ARRAY_SIZE(ipq6018_pcie_pcs_tbl),
> > >>> + .pcs_misc = ipq9574_gen3x2_pcie_pcs_misc_tbl,
> > >>> + .pcs_misc_num =
> > >>> ARRAY_SIZE(ipq9574_gen3x2_pcie_pcs_misc_tbl),
> > >>> + },
> > >>> + .reset_list = ipq8074_pciephy_reset_l,
> > >>> + .num_resets = ARRAY_SIZE(ipq8074_pciephy_reset_l),
> > >>> + .vreg_list = NULL,
> > >>> + .num_vregs = 0,
> > >>> + .regs = pciephy_v4_regs_layout,
> > >>
> > >> So, is it v4 or v5?
> > >
> > > Please give me a day or so to go over my notes and give you a more
> > > coherent explanation of why this versioning was chosen. I am only
> > > working from the QCA 5.4 downstream sources. I don't have any
> > > documentation for the silicon
> >
> > The downstream QCA kernel uses the same table for ipq6018, ipq8074-gen3,
> > and ipq9574. It is named "ipq_pciephy_gen3_regs_layout". Thus, it made
> > sense to use the same upstream table for ipq9574, "pciephy_v4_regs_layout".
> >
> > As far as the register tables go, the pcs/pcs_misc are squashed into the
> > same table in the downstream 5.4 kernel. I was able to separate the two
> > tables because the pcs_misc registers were defined with an offset of
> > 0x400. For example:
> >
> > /* QMP V2 PHY for PCIE gen3 2 Lane ports - PCS Misc registers */
> > #define PCS_PCIE_X2_POWER_STATE_CONFIG2 0x40c
> > #define PCS_PCIE_X2_POWER_STATE_CONFIG4 0x414
> > #define PCS_PCIE_X2_ENDPOINT_REFCLK_DRIVE 0x420
> > #define PCS_PCIE_X2_L1P1_WAKEUP_DLY_TIME_AUXCLK_L 0x444
> > #define PCS_PCIE_X2_L1P1_WAKEUP_DLY_TIME_AUXCLK_H 0x448
> > #define PCS_PCIE_X2_L1P2_WAKEUP_DLY_TIME_AUXCLK_L 0x44c
> > #define PCS_PCIE_X2_L1P2_WAKEUP_DLY_TIME_AUXCLK_H 0x450
> > ...
> >
> > Here, QPHY_V4_PCS_PCIE_POWER_STATE_CONFIG2 = 0xc would be correct,
> > assuming a pcs_misc offset of 0x400. However, starting with
> > ENDPOINT_REFCLK_DRIVE, the register would be
> > QPHY_V4_PCS_PCIE_ENDPOINT_REFCLK_DRIVE = 0x1c. Our offsets are off-by 0x4.
> >
> > The existing V5 offsets, on the other hand, were all correct. For this
> > reason, I considered that V5 is the most likely place to add the missing
> > PCS misc definitions.
>
> Ok, sounds sane. Please use _v5 for the regs layout.
>
> >
> > Is this explanation sufficiently convincing? Where does the v4/v5 scheme
> > in upstream kernel originate?
>
> Sometimes it's vendor kernels, sometimes it's a feedback from devs
> that have access to actual specs.
>
>
> --
> With best wishes
> Dmitry
next prev parent reply other threads:[~2024-04-29 6:20 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-15 18:20 [PATCH v3 0/7] ipq9574: Enable PCI-Express support Alexandru Gagniuc
2024-04-15 18:20 ` [PATCH v3 1/7] dt-bindings: clock: Add PCIe pipe related clocks for IPQ9574 Alexandru Gagniuc
2024-04-15 18:20 ` [PATCH v3 2/7] clk: qcom: gcc-ipq9574: Add PCIe pipe clocks Alexandru Gagniuc
2024-04-15 20:04 ` Dmitry Baryshkov
2024-04-19 22:22 ` Stephen Boyd
2024-04-15 18:20 ` [PATCH v3 3/7] dt-bindings: PCI: qcom: Add IPQ9574 PCIe controller Alexandru Gagniuc
2024-04-18 12:59 ` Rob Herring
2024-04-15 18:20 ` [PATCH v3 4/7] PCI: qcom: Add support for IPQ9574 Alexandru Gagniuc
2024-04-15 20:04 ` Dmitry Baryshkov
2024-04-15 20:07 ` mr.nuke.me
2024-04-17 7:06 ` Manivannan Sadhasivam
2024-04-19 19:44 ` mr.nuke.me
2024-04-22 7:11 ` Manivannan Sadhasivam
2024-04-17 7:14 ` Manivannan Sadhasivam
2024-04-15 18:20 ` [PATCH v3 5/7] dt-bindings: phy: qcom,ipq8074-qmp-pcie: add ipq9574 gen3x2 PHY Alexandru Gagniuc
2024-04-18 13:00 ` Rob Herring
2024-04-15 18:20 ` [PATCH v3 6/7] phy: qcom-qmp-pcie: add support for " Alexandru Gagniuc
2024-04-15 20:10 ` Dmitry Baryshkov
2024-04-15 21:25 ` mr.nuke.me
2024-04-16 21:25 ` Alex G.
2024-04-16 21:50 ` Dmitry Baryshkov
2024-04-29 6:20 ` Varadarajan Narayanan [this message]
2024-04-29 10:55 ` Dmitry Baryshkov
2024-04-30 6:31 ` Varadarajan Narayanan
2024-04-30 21:51 ` mr.nuke.me
2024-04-30 21:59 ` Dmitry Baryshkov
2024-04-15 18:20 ` [PATCH v3 7/7] arm64: dts: qcom: ipq9574: add PCIe2 nodes Alexandru Gagniuc
2024-04-17 7:34 ` Manivannan Sadhasivam
2024-04-18 15:33 ` mr.nuke.me
2024-04-22 7:10 ` Manivannan Sadhasivam
2024-04-22 14:29 ` Konrad Dybcio
2024-04-19 14:28 ` [PATCH v3 0/7] ipq9574: Enable PCI-Express support Kathiravan Thirumoorthy
2024-04-19 19:47 ` mr.nuke.me
2024-04-23 6:11 ` Kathiravan Thirumoorthy
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=Zi88Hdx6UgBo/gti@hu-varada-blr.qualcomm.com \
--to=quic_varada@quicinc.com \
--cc=andersson@kernel.org \
--cc=bhelgaas@google.com \
--cc=conor+dt@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=dmitry.baryshkov@linaro.org \
--cc=kishon@kernel.org \
--cc=konrad.dybcio@linaro.org \
--cc=krzk+dt@kernel.org \
--cc=kw@linux.com \
--cc=linux-arm-msm@vger.kernel.org \
--cc=linux-clk@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=linux-phy@lists.infradead.org \
--cc=lpieralisi@kernel.org \
--cc=manivannan.sadhasivam@linaro.org \
--cc=mr.nuke.me@gmail.com \
--cc=mturquette@baylibre.com \
--cc=robh@kernel.org \
--cc=sboyd@kernel.org \
--cc=vkoul@kernel.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).