linux-kernel - Re: [PATCH v8 6/8] PCI: cadence: Add support to configure virtual functions

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: <ce380e6f-d10a-4db2-9bde-82615cfbe4db@ti.com>
Date:   Wed, 18 Aug 2021 19:25:09 +0530
From:   Kishon Vijay Abraham I <kishon@...com>
To:     Bjorn Helgaas <helgaas@...nel.org>
CC:     Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Rob Herring <robh+dt@...nel.org>,
        Tom Joseph <tjoseph@...ence.com>,
        Jingoo Han <jingoohan1@...il.com>,
        Gustavo Pimentel <gustavo.pimentel@...opsys.com>,
        Marek Vasut <marek.vasut+renesas@...il.com>,
        Yoshihiro Shimoda <yoshihiro.shimoda.uh@...esas.com>,
        Shawn Lin <shawn.lin@...k-chips.com>,
        Heiko Stuebner <heiko@...ech.de>,
        Jonathan Corbet <corbet@....net>,
        Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        <linux-pci@...r.kernel.org>, <linux-doc@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>, <devicetree@...r.kernel.org>,
        <linux-renesas-soc@...r.kernel.org>,
        <linux-rockchip@...ts.infradead.org>,
        <linux-arm-kernel@...ts.infradead.org>,
        Lokesh Vutla <lokeshvutla@...com>
Subject: Re: [PATCH v8 6/8] PCI: cadence: Add support to configure virtual
 functions

Hi Bjorn,

On 17/08/21 9:08 pm, Bjorn Helgaas wrote:
> On Wed, Aug 11, 2021 at 12:16:54PM +0530, Kishon Vijay Abraham I wrote:
>> Now that support for SR-IOV is added in PCIe endpoint core, add support
>> to configure virtual functions in the Cadence PCIe EP driver.
>>
>> Signed-off-by: Kishon Vijay Abraham I <kishon@...com>
>> ---
>>  .../pci/controller/cadence/pcie-cadence-ep.c  | 136 +++++++++++++++---
>>  drivers/pci/controller/cadence/pcie-cadence.h |   9 ++
>>  2 files changed, 125 insertions(+), 20 deletions(-)
> 
>> @@ -92,21 +118,29 @@ static int cdns_pcie_ep_set_bar(struct pci_epc *epc, u8 fn, u8 vfn,
>>  
>>  	addr0 = lower_32_bits(bar_phys);
>>  	addr1 = upper_32_bits(bar_phys);
>> -	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar),
>> -			 addr0);
>> -	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar),
>> -			 addr1);
>>  
>>  	reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG(bar, fn);
>> +	if (vfn == 1)
>> +		reg = CDNS_PCIE_LM_EP_VFUNC_BAR_CFG(bar, fn);
> 
> Seems sort of weird to compute "reg", then sometimes overwrite it, as
> opposed to:
> 
>   if (vfn == 1)
>     reg = CDNS_PCIE_LM_EP_VFUNC_BAR_CFG(bar, fn);
>   else
>     reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG(bar, fn);

I tried to write it without "else". But I can change it back.
> 
> Also slightly weird that "vfn" is basically used as a boolean, but
> it's actually a u8 virtual function number.  I guess VF 1 is special
> and not like the other VFs?

VF1 is special in that it's enough for configuring the SR-IOV capability
but below the "vfn" is used for configuring inbound window.

Thanks
Kishon
> 
>>  	b = (bar < BAR_4) ? bar : bar - BAR_4;
>>  
>> -	cfg = cdns_pcie_readl(pcie, reg);
>> -	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>> -		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>> -	cfg |= (CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE(b, aperture) |
>> -		CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl));
>> -	cdns_pcie_writel(pcie, reg, cfg);
>> +	if (vfn == 0 || vfn == 1) {
>> +		cfg = cdns_pcie_readl(pcie, reg);
>> +		cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>> +			 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>> +		cfg |= (CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE(b, aperture) |
>> +			CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl));
>> +		cdns_pcie_writel(pcie, reg, cfg);
>> +	}
>>  
>> +	fn = cdns_pcie_get_fn_from_vfn(pcie, fn, vfn);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar),
>> +			 addr0);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar),
>> +			 addr1);
>> +
>> +	if (vfn > 0)
>> +		epf = &epf->epf[vfn - 1];
>>  	epf->epf_bar[bar] = epf_bar;
>>  
>>  	return 0;
>> @@ -122,18 +156,25 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn, u8 vfn,
>>  	u32 reg, cfg, b, ctrl;
>>  
>>  	reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG(bar, fn);
>> +	if (vfn == 1)
>> +		reg = CDNS_PCIE_LM_EP_VFUNC_BAR_CFG(bar, fn);
> 
> Similar recomputation of "reg".
> 
>>  	b = (bar < BAR_4) ? bar : bar - BAR_4;
>>  
>> -	ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_DISABLED;
>> -	cfg = cdns_pcie_readl(pcie, reg);
>> -	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>> -		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>> -	cfg |= CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl);
>> -	cdns_pcie_writel(pcie, reg, cfg);
>> +	if (vfn == 0 || vfn == 1) {
>> +		ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_DISABLED;
>> +		cfg = cdns_pcie_readl(pcie, reg);
>> +		cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>> +			 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>> +		cfg |= CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl);
>> +		cdns_pcie_writel(pcie, reg, cfg);
>> +	}