public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Manivannan Sadhasivam <mani@kernel.org>
To: Serge Semin <fancer.lancer@gmail.com>
Cc: "Manivannan Sadhasivam" <manivannan.sadhasivam@linaro.org>,
	"Jingoo Han" <jingoohan1@gmail.com>,
	"Gustavo Pimentel" <gustavo.pimentel@synopsys.com>,
	"Lorenzo Pieralisi" <lpieralisi@kernel.org>,
	"Krzysztof Wilczyński" <kw@linux.com>,
	"Rob Herring" <robh@kernel.org>,
	"Bjorn Helgaas" <bhelgaas@google.com>,
	linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-arm-msm@vger.kernel.org
Subject: Re: [PATCH 1/2] PCI: dwc: Add new accessors to enable/disable DBI CS2 while setting the BAR size
Date: Thu, 19 Oct 2023 10:58:35 +0530	[thread overview]
Message-ID: <20231019052835.GC5142@thinkpad> (raw)
In-Reply-To: <rsv5vgle2d36skx75ds4hqzmlqwldmj4g4ghrlyfuu3ideb3rh@74mnro7qnp4v>

On Wed, Oct 18, 2023 at 05:13:41PM +0300, Serge Semin wrote:
> On Tue, Oct 17, 2023 at 11:47:54AM +0530, Manivannan Sadhasivam wrote:
> > From: Manivannan Sadhasivam <mani@kernel.org>
> > 
> > As per the DWC databook v4.21a, section M.4.1, in order to write some read
> > only and shadow registers through application DBI, the device driver should
> > assert DBI Chip Select 2 (CS2) in addition to DBI Chip Select (CS).
> > 
> > This is a requirement at least on the Qcom platforms while programming the
> > BAR size, as the BAR mask registers are marked RO. So let's add two new
> > accessors dw_pcie_dbi_cs2_{en/dis} to enable/disable CS2 access in a vendor
> > specific way while programming the BAR size.
> 
> Emm, it's a known thing for all IP-core versions: dbi_cs2 must be
> asserted to access the shadow DW PCIe CSRs space for both RC and
> EP including the BARs mask and their enabling/disabling flag (there
> are many more shadow CSRs available on DW PCIe EPs, and a few in DW
> PCIe RCs). That's why the dw_pcie_ops->writel_dbi2 pointer has been
> defined in the first place (dbi2 suffix means dbi_cs2). You should use
> it to create the platform-specific dbi_cs2 write accessors like it's
> done in pci-keystone.c and pcie-bt1.c. For instance like this:
> 
> static void qcom_pcie_write_dbi2(struct dw_pcie *pci, u32 reg, size_t size, u32 val)
> {
> 	struct qcom_pcie_ep *pcie_ep = to_pcie_ep(pci);
> 	int ret;
> 
> 	writel(1, pcie_ep->elbi + ELBI_CS2_ENABLE);
> 
> 	ret = dw_pcie_write(pci->dbi_base2 + reg, size, val);
> 	if (ret)
> 		dev_err(pci->dev, "write DBI address failed\n");
> 
> 	writel(0, pcie_ep->elbi + ELBI_CS2_ENABLE);
> }
> 

Hmm, I was not aware that this write_dbi2() callback is supposed to enable the
CS2 access internally. But this approach doesn't look good to me.

We already have accessors for enabling write access to DBI RO registers:

dw_pcie_dbi_ro_wr_en()
dw_pcie_dbi_ro_wr_dis()

And IMO DBI_CS2 access should also be done this way instead of hiding it inside
the register write callback.

- Mani

> /* Common DWC controller ops */
> static const struct dw_pcie_ops pci_ops = {
> 	.link_up = qcom_pcie_dw_link_up,
> 	.start_link = qcom_pcie_dw_start_link,
> 	.stop_link = qcom_pcie_dw_stop_link,
> 	.write_dbi2 = qcom_pcie_write_dbi2,
> };
> 
> For that reason there is absolutely no need in adding the new
> callbacks.
> 
> -Serge(y)
> 
> > 
> > Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
> > ---
> >  drivers/pci/controller/dwc/pcie-designware-ep.c |  6 ++++++
> >  drivers/pci/controller/dwc/pcie-designware.h    | 13 +++++++++++++
> >  2 files changed, 19 insertions(+)
> > 
> > diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > index d34a5e87ad18..1874fb3d8df4 100644
> > --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> > +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > @@ -269,11 +269,17 @@ static int dw_pcie_ep_set_bar(struct pci_epc *epc, u8 func_no, u8 vfunc_no,
> >  
> >  	dw_pcie_dbi_ro_wr_en(pci);
> >  
> > +	dw_pcie_dbi_cs2_en(pci);
> >  	dw_pcie_writel_dbi2(pci, reg_dbi2, lower_32_bits(size - 1));
> > +	dw_pcie_dbi_cs2_dis(pci);
> > +
> >  	dw_pcie_writel_dbi(pci, reg, flags);
> >  
> >  	if (flags & PCI_BASE_ADDRESS_MEM_TYPE_64) {
> > +		dw_pcie_dbi_cs2_en(pci);
> >  		dw_pcie_writel_dbi2(pci, reg_dbi2 + 4, upper_32_bits(size - 1));
> > +		dw_pcie_dbi_cs2_dis(pci);
> > +
> >  		dw_pcie_writel_dbi(pci, reg + 4, 0);
> >  	}
> >  
> > diff --git a/drivers/pci/controller/dwc/pcie-designware.h b/drivers/pci/controller/dwc/pcie-designware.h
> > index 55ff76e3d384..3cba27b5bbe5 100644
> > --- a/drivers/pci/controller/dwc/pcie-designware.h
> > +++ b/drivers/pci/controller/dwc/pcie-designware.h
> > @@ -379,6 +379,7 @@ struct dw_pcie_ops {
> >  			     size_t size, u32 val);
> >  	void    (*write_dbi2)(struct dw_pcie *pcie, void __iomem *base, u32 reg,
> >  			      size_t size, u32 val);
> > +	void	(*dbi_cs2_access)(struct dw_pcie *pcie, bool enable);
> >  	int	(*link_up)(struct dw_pcie *pcie);
> >  	enum dw_pcie_ltssm (*get_ltssm)(struct dw_pcie *pcie);
> >  	int	(*start_link)(struct dw_pcie *pcie);
> > @@ -508,6 +509,18 @@ static inline void dw_pcie_dbi_ro_wr_dis(struct dw_pcie *pci)
> >  	dw_pcie_writel_dbi(pci, reg, val);
> >  }
> >  
> > +static inline void dw_pcie_dbi_cs2_en(struct dw_pcie *pci)
> > +{
> > +	if (pci->ops && pci->ops->dbi_cs2_access)
> > +		pci->ops->dbi_cs2_access(pci, true);
> > +}
> > +
> > +static inline void dw_pcie_dbi_cs2_dis(struct dw_pcie *pci)
> > +{
> > +	if (pci->ops && pci->ops->dbi_cs2_access)
> > +		pci->ops->dbi_cs2_access(pci, false);
> > +}
> > +
> >  static inline int dw_pcie_start_link(struct dw_pcie *pci)
> >  {
> >  	if (pci->ops && pci->ops->start_link)
> > 
> > -- 
> > 2.25.1
> > 

-- 
மணிவண்ணன் சதாசிவம்

  reply	other threads:[~2023-10-19  5:28 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-10-17  6:17 [PATCH 0/2] PCI: dwc: Fix the BAR size programming Manivannan Sadhasivam
2023-10-17  6:17 ` [PATCH 1/2] PCI: dwc: Add new accessors to enable/disable DBI CS2 while setting the BAR size Manivannan Sadhasivam
2023-10-18 14:13   ` Serge Semin
2023-10-19  5:28     ` Manivannan Sadhasivam [this message]
2023-10-19 14:37       ` Serge Semin
2023-10-19 16:50         ` Manivannan Sadhasivam
2023-10-17  6:17 ` [PATCH 2/2] PCI: qcom-ep: Implement dbi_cs2_access() function callback for DBI CS2 access Manivannan Sadhasivam
2023-10-17 14:24   ` Bjorn Andersson
2023-10-17 16:21     ` Manivannan Sadhasivam
2023-10-17 16:56       ` Bjorn Andersson
2023-10-17 17:41         ` Manivannan Sadhasivam
2023-10-17 22:18           ` Bjorn Andersson
2023-10-18 13:27             ` Manivannan Sadhasivam
2023-10-19  3:18               ` Bjorn Andersson
2023-10-19  5:19                 ` Manivannan Sadhasivam

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=20231019052835.GC5142@thinkpad \
    --to=mani@kernel.org \
    --cc=bhelgaas@google.com \
    --cc=fancer.lancer@gmail.com \
    --cc=gustavo.pimentel@synopsys.com \
    --cc=jingoohan1@gmail.com \
    --cc=kw@linux.com \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=lpieralisi@kernel.org \
    --cc=manivannan.sadhasivam@linaro.org \
    --cc=robh@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