Aw: Re: Re: [PATCH v2 0/7] rockchip: rk3568: Fix use of PCIe bifurcation
Frank Wunderlich
frank-w at public-files.de
Mon Oct 2 11:03:14 CEST 2023
Hi Jonas,
> Gesendet: Montag, 02. Oktober 2023 um 10:33 Uhr
> Von: "Jonas Karlman" <jonas at kwiboo.se>
> An: "Frank Wunderlich" <frank-w at public-files.de>, "Sebastian Reichel" <sebastian.reichel at collabora.com>
> Cc: "Kever Yang" <kever.yang at rock-chips.com>, "Simon Glass" <sjg at chromium.org>, "Philipp Tomsich" <philipp.tomsich at vrull.eu>, "Eugen Hristev" <eugen.hristev at collabora.com>, "Jon Lin" <jon.lin at rock-chips.com>, "FUKAUMI Naoki" <naoki at radxa.com>, "John Clark" <inindev at gmail.com>, u-boot at lists.denx.de
> Betreff: Re: Aw: Re: [PATCH v2 0/7] rockchip: rk3568: Fix use of PCIe bifurcation
>
> Hi Frank,
>
> On 2023-10-02 09:17, Frank Wunderlich wrote:
> > Hi,
> >
> >> Gesendet: Sonntag, 01. Oktober 2023 um 21:11 Uhr
> >> Von: "Sebastian Reichel" <sebastian.reichel at collabora.com>
> >> An: "Jonas Karlman" <jonas at kwiboo.se>
> >> Cc: "Kever Yang" <kever.yang at rock-chips.com>, "Simon Glass" <sjg at chromium.org>, "Philipp Tomsich" <philipp.tomsich at vrull.eu>, "Eugen Hristev" <eugen.hristev at collabora.com>, "Jon Lin" <jon.lin at rock-chips.com>, "FUKAUMI Naoki" <naoki at radxa.com>, "John Clark" <inindev at gmail.com>, u-boot at lists.denx.de
> >> Betreff: Re: [PATCH v2 0/7] rockchip: rk3568: Fix use of PCIe bifurcation
> >>
> >> Hi,
> >>
> >> On Wed, Sep 27, 2023 at 06:07:59PM +0000, Jonas Karlman wrote:
> >>> Above works with this series in U-Boot and also in mainline linux. Not
> >>> sure the RK3588 bifurcation part of driver have been verified on real HW,
> >>> rk3588 boards available for testing have typicality not used bifurcation.
> >>
> >> Upstream kernel bifurcation code looks ok to me. It's quite different
> >> from Rockchip vendor implementation, which might have confused Kever.
> >> I also missed this when having a quick look initially. As you said
> >> Rock 5A, Rock 5B and EVB1 do not use bifurcation, so no testing happened
> >> on our side.
> >
> > my Bananapi R2Pro (rk3568) uses bifurcation and it is working in mainline Linux, but not yet in uboot.
>
> The issue on your Bananapi R2Pro is most likely not related to the
> pcie bifurcation code added in this series.
>
> The coded added in this series works same/very similar as on mainline
> linux and fixes bifurcation use on devices such as Radxa E25 and NanoPi
> R5C/R5S.
>
> Please clarify if pcie work in mainline linux when booting with mainline
> u-boot, or does it only work when booting with vendor u-boot? (in case
> vendor u-boot set different/default pinconf).
>
> >
> > posted my test-results here:
> > https://patchwork.ozlabs.org/project/uboot/patch/20230918173624.31464-1-linux@fw-web.de/#3183519
>
> From your pci enum command I only see an expected pcie at fe270000 (ngff)
> link fail. Does it also print out link fail for pcie at fe280000 (minipcie)?
good catch, you're right, no message of mpcie, only ngff (no card)
#initial status (before enabling pcie-related regulators)
BPI-R2PRO> regulator status
Name Enabled uV mA Mode Status
vdd_logic enabled 900000 - - 0
vdd_gpu enabled 900000 - - 0
vcc_ddr enabled 500000 - - 0
vdd_npu disabled 500000 - - 0
vcc_1v8 enabled 1800000 - - 0
vdda0v9_image enabled 900000 - - 0
vdda_0v9 enabled 900000 - - 0
vdda0v9_pmu enabled 900000 - - 0
vccio_acodec enabled 3300000 - - 0
vccio_sd enabled 3300000 - - 0
vcc3v3_pmu enabled 3300000 - - 0
vcca_1v8 enabled 1800000 - - 0
vcca1v8_pmu enabled 1800000 - - 0
vcca1v8_image enabled 1800000 - - 0
vcc_3v3 enabled 0 - - 0
vcc3v3_sd enabled 0 - - 0
dc_12v enabled 12000000 - - 0
vcc3v3_sys enabled 3300000 - - 0
vcc5v0_sys enabled 5000000 - - 0
pcie30_avdd0v9 enabled 900000 - - 0
pcie30_avdd1v8 enabled 1800000 - - 0
vcc3v3_pcie disabled 3300000 - - 0
vcc3v3_minipcie disabled 3300000 - - 0
vcc3v3_ngff disabled 3300000 - - 0
vcc5v0_usb enabled 5000000 - - 0
vcc5v0_usb_host disabled 5000000 - - 0
vcc5v0_usb_otg disabled 5000000 - - 0
BPI-R2PRO> regulator dev vcc3v3_pcie
dev: vcc3v3_pcie @ vcc3v3-pi6c-05-regulator
BPI-R2PRO> regulator enable
BPI-R2PRO> regulator status
Regulator vcc3v3_pcie status:
* enable: 1 (true)
* value uV: 3300000
* current uA: No data available (err: -61)
* mode id: Function not implemented (err: -38)
BPI-R2PRO> regulator dev vcc3v3_minipcie
dev: vcc3v3_minipcie @ vcc3v3-minipcie-regulator
BPI-R2PRO> regulator status
Regulator vcc3v3_minipcie status:
* enable: 0 (false)
* value uV: 3300000
* current uA: No data available (err: -61)
* mode id: Function not implemented (err: -38)
BPI-R2PRO> regulator enable
BPI-R2PRO> regulator status
Regulator vcc3v3_minipcie status:
* enable: 1 (true)
* value uV: 3300000
* current uA: No data available (err: -61)
* mode id: Function not implemented (err: -38)
BPI-R2PRO> pci enum
pcie_dw_rockchip pcie at fe270000: PCIe-0 Link Fail
BPI-R2PRO> pci 1
Scanning PCI devices on bus 1
BusDevFun VendorId DeviceId Device Class Sub-Class
_____________________________________________________________
01.00.00 0x1d87 0x3566 Bridge device 0x04
BPI-R2PRO> pci 2
Scanning PCI devices on bus 2
BusDevFun VendorId DeviceId Device Class Sub-Class
_____________________________________________________________
02.00.00 0x14c3 0x7612 Network controller 0x80
so mpcie slot works so far as it detects the network-card (mt7612),
only need to find a way to enable the regulators (vcc3v3_pcie, vcc3v3_minipcie, vcc3v3_ngff).
btw. is there a way to show overall status (like the first regulator status) after enabling
the regulators (only shows current regulator)?
> If not, try running "pci 1" and "pci 2" after an initial "pci enum" to
> see if the root complex and your minipcie device is detected.
sorry for confusion, have not done much with pcie in uboot yet.
so you can add my tested by
Tested-by: Frank Wunderlich <frank-w at public-files.de>
Thank you very much for your work.
regards Frank
More information about the U-Boot
mailing list