From: Rob Herring <robh+dt@kernel.org>
To: Bjorn Helgaas <helgaas@kernel.org>
Cc: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>,
Bjorn Helgaas <bhelgaas@google.com>,
Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
PCI <linux-pci@vger.kernel.org>,
Punit Agrawal <punit1.agrawal@toshiba.co.jp>,
yuji2.ishikawa@toshiba.co.jp,
linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
Kishon Vijay Abraham I <kishon@ti.com>
Subject: Re: [PATCH v3 2/3] PCI: Visconti: Add Toshiba Visconti PCIe host controller driver
Date: Wed, 16 Jun 2021 10:32:05 -0600 [thread overview]
Message-ID: <CAL_JsqLMdxA_yVdz6_s7XP8SsCDhwctUxG+3+jAnJs5fwyk=MA@mail.gmail.com> (raw)
In-Reply-To: <20210524185839.GA1102116@bjorn-Precision-5520>
On Mon, May 24, 2021 at 12:58 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
>
> [+cc Kishon for cpu_addr_fixup() question]
>
> Please make the subject "PCI: visconti: Add ..." since the driver
> names are usually lower-case. When referring to the hardware itself,
> use "Visconti", of course.
>
> On Mon, May 24, 2021 at 03:30:03PM +0900, Nobuhiro Iwamatsu wrote:
> > Add support to PCIe RC controller on Toshiba Visconti ARM SoCs. PCIe
> > controller is based of Synopsys DesignWare PCIe core.
> >
> > This patch does not yet use the clock framework to control the clock.
> > This will be replaced in the future.
> >
> > v2 -> v3:
> > - Update subject.
> > - Wrap description in 75 columns.
> > - Change config name to PCIE_VISCONTI_HOST.
> > - Update Kconfig text.
> > - Drop blank lines.
> > - Adjusted to 80 columns.
> > - Drop inline from functions for register access.
> > - Changed function name from visconti_pcie_check_link_status to
> > visconti_pcie_link_up.
> > - Update to using dw_pcie_host_init().
> > - Reorder these in the order of use in visconti_pcie_establish_link.
> > - Rewrite visconti_pcie_host_init() without dw_pcie_setup_rc().
> > - Change function name from visconti_device_turnon() to
> > visconti_pcie_power_on().
> > - Unify formats such as dev_err().
> > - Drop error label in visconti_add_pcie_port().
> >
> > v1 -> v2:
> > - Fix typo in commit message.
> > - Drop "depends on OF && HAS_IOMEM" from Kconfig.
> > - Stop using the pointer of struct dw_pcie.
> > - Use _relaxed variant.
> > - Drop dw_pcie_wait_for_link.
> > - Drop dbi resource processing.
> > - Drop MSI IRQ initialization processing.
>
> Thanks for the changelog. Please move it after the "---" line for
> future versions. That way it won't appear in the commit log when this
> is merged. The notes about v1->v2, v2->v3, etc are useful during
> review, but not after this is merged.
>
> > Signed-off-by: Yuji Ishikawa <yuji2.ishikawa@toshiba.co.jp>
> > Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
> > ---
> > drivers/pci/controller/dwc/Kconfig | 9 +
> > drivers/pci/controller/dwc/Makefile | 1 +
> > drivers/pci/controller/dwc/pcie-visconti.c | 369 +++++++++++++++++++++
> > 3 files changed, 379 insertions(+)
> > create mode 100644 drivers/pci/controller/dwc/pcie-visconti.c
> > +static u64 visconti_pcie_cpu_addr_fixup(struct dw_pcie *pci, u64 pci_addr)
> > +{
> > + return pci_addr - PCIE_BUS_OFFSET;
>
> This is called from __dw_pcie_prog_outbound_atu() as:
>
> cpu_addr = pci->ops->cpu_addr_fixup(pci, cpu_addr);
>
> so I think the parameter here should be *cpu_addr*, not pci_addr.
>
> dra7xx and artpec6 also call it "pci_addr", which is at best
> confusing.
>
> I'm also confused about exactly what .cpu_addr_fixup() does. Is it
> applying an offset that cannot be deduced from the DT description? If
> so, *should* this offset be described in DT?
It could be perhaps, but it would be a custom property, not something
we can handle in 'ranges'. I'd rather it be implicit from the
compatible than a custom property.
AIUI, the issue is the cpu address gets masked (high bits discarded).
This can happen when the internal bus address decoding throws away
upper address bits.
For example:
0xa0000000 -> 0x20000000 -> 0x00000000
cpu addr -> DW local addr -> PCI bus addr
DT has the first and last addresses, but iATU needs the middle and last address.
This could be just a data value rather than an ops function. While a
subtract works here, that's fragile (the DT needs to match the
#define) and I think a mask would be more appropriate.
Rob
next prev parent reply other threads:[~2021-06-16 16:32 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-24 6:30 [PATCH v3 0/3] Visconti: Add Toshiba Visconti PCIe host controller driver Nobuhiro Iwamatsu
2021-05-24 6:30 ` [PATCH v3 1/3] dt-bindings: pci: Add DT binding for Toshiba Visconti PCIe controller Nobuhiro Iwamatsu
2021-05-26 16:33 ` Rob Herring
2021-05-27 0:22 ` Nobuhiro Iwamatsu
2021-05-24 6:30 ` [PATCH v3 2/3] PCI: Visconti: Add Toshiba Visconti PCIe host controller driver Nobuhiro Iwamatsu
2021-05-24 11:10 ` Krzysztof Wilczyński
2021-05-28 0:19 ` Nobuhiro Iwamatsu
2021-05-24 18:58 ` Bjorn Helgaas
2021-06-16 1:31 ` Nobuhiro Iwamatsu
2021-06-16 16:32 ` Rob Herring [this message]
2021-06-27 0:00 ` Nobuhiro Iwamatsu
2021-05-24 6:30 ` [PATCH v3 3/3] MAINTAINERS: Add entries for Toshiba Visconti PCIe controller Nobuhiro Iwamatsu
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='CAL_JsqLMdxA_yVdz6_s7XP8SsCDhwctUxG+3+jAnJs5fwyk=MA@mail.gmail.com' \
--to=robh+dt@kernel.org \
--cc=bhelgaas@google.com \
--cc=helgaas@kernel.org \
--cc=kishon@ti.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=lorenzo.pieralisi@arm.com \
--cc=nobuhiro1.iwamatsu@toshiba.co.jp \
--cc=punit1.agrawal@toshiba.co.jp \
--cc=yuji2.ishikawa@toshiba.co.jp \
/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).