[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAEQ9gE=ztgQ+pGJVxKgk5dVWDSSfOG7r=s1cDa_x0_Zsf2eyYA@mail.gmail.com>
Date: Sun, 16 Nov 2025 23:36:13 -0800
From: Roger Shimizu <rosh@...ian.org>
To: Jens Reidel <adrian@...nlining.org>,
Dmitry Baryshkov <dmitry.baryshkov@....qualcomm.com>
Cc: Hongyang Zhao <hongyang.zhao@...ndersoft.com>, Bjorn Andersson <andersson@...nel.org>,
Konrad Dybcio <konradybcio@...nel.org>, Rob Herring <robh@...nel.org>,
Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley <conor+dt@...nel.org>, linux-arm-msm@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] arm64: dts: qcom: rubikpi3: Add qcs6490-rubikpi3
board dts
Thanks Jens, and Dmitry for the review!
On Sat, Nov 15, 2025 at 9:25 AM Jens Reidel <adrian@...nlining.org> wrote:
>
> Hi,
>
> On 11/14/25 5:34 PM, Hongyang Zhao wrote:
> > Add DTS for Thundercomm qcs6490-rubikpi3 board which uses
> > QCS6490 SoC.
> >
> > Works:
> > - Bluetooth (AP6256)
> > - Wi-Fi (AP6256)
> > - Ethernet (AX88179B connected to UPD720201)
> > - FAN
> > - Two USB Type-A 3.0 ports (UPD720201 connected to PCIe0)
> > - M.2 M-Key 2280 PCIe 3.0
> > - RTC
> > - USB Type-C
> > - USB Type-A 2.0 port
> > - 40PIN: I2C x1, UART x1
> >
> > Signed-off-by: Hongyang Zhao <hongyang.zhao@...ndersoft.com>
> > Reviewed-by: Roger Shimizu <rosh@...ian.org>
> > ---
> > arch/arm64/boot/dts/qcom/Makefile | 1 +
> > .../boot/dts/qcom/qcs6490-thundercomm-rubikpi3.dts | 1415 ++++++++++++++++++++
> > 2 files changed, 1416 insertions(+)
> >
> > diff --git a/arch/arm64/boot/dts/qcom/Makefile b/arch/arm64/boot/dts/qcom/Makefile
> > index 6f34d5ed331c..2433b15754fe 100644
> > --- a/arch/arm64/boot/dts/qcom/Makefile
> > +++ b/arch/arm64/boot/dts/qcom/Makefile
> > @@ -138,6 +138,7 @@ qcs6490-rb3gen2-industrial-mezzanine-dtbs := qcs6490-rb3gen2.dtb qcs6490-rb3gen2
> >
> > dtb-$(CONFIG_ARCH_QCOM) += qcs6490-rb3gen2-industrial-mezzanine.dtb
> > dtb-$(CONFIG_ARCH_QCOM) += qcs6490-rb3gen2-vision-mezzanine.dtb
> > +dtb-$(CONFIG_ARCH_QCOM) += qcs6490-thundercomm-rubikpi3.dtb
> > dtb-$(CONFIG_ARCH_QCOM) += qcs8300-ride.dtb
> > dtb-$(CONFIG_ARCH_QCOM) += qcs8550-aim300-aiot.dtb
> > dtb-$(CONFIG_ARCH_QCOM) += qcs9100-ride.dtb
> > diff --git a/arch/arm64/boot/dts/qcom/qcs6490-thundercomm-rubikpi3.dts b/arch/arm64/boot/dts/qcom/qcs6490-thundercomm-rubikpi3.dts
> > new file mode 100644
> > index 000000000000..4c9016992de3
> > --- /dev/null
> > +++ b/arch/arm64/boot/dts/qcom/qcs6490-thundercomm-rubikpi3.dts
> > @@ -0,0 +1,1415 @@
>
> [snip]
>
> > +
> > +&pcie0 {
> > + perst-gpios = <&tlmm 87 GPIO_ACTIVE_LOW>;
> > + wake-gpios = <&tlmm 89 GPIO_ACTIVE_HIGH>;
> > +
> > + pinctrl-0 = <&pcie0_clkreq_n>,
> > + <&pcie0_reset_n>,
> > + <&pcie0_wake_n>;
> > + pinctrl-names = "default";
> > +
> > + status = "okay";
> > +};
> > +
> > +&pcie0_phy {
> > + vdda-phy-supply = <&vreg_l10c_0p88>;
> > + vdda-pll-supply = <&vreg_l6b_1p2>;
> > +
> > + status = "okay";
> > +};
> > +
> > +&pcie1 {
> > + /* Using traditional address mapping */
> > + reg = <0 0x01c08000 0 0x3000>,
> > + <0 0x40000000 0 0xf1d>,
> > + <0 0x40000f20 0 0xa8>,
> > + <0 0x40001000 0 0x1000>,
> > + <0 0x40100000 0 0x100000>;
> > +
> > + ranges = <0x01000000 0x0 0x00000000 0x0 0x40200000 0x0 0x100000>,
> > + <0x02000000 0x0 0x40300000 0x0 0x40300000 0x0 0x1fd00000>;
>
> Thanks for attempting to fix the PCIe issues. With your patch series
> applied on top of linux-next, I'm still seeing PCIe issues:
>
> [ 0.380693] Internal error: synchronous external abort:
> 0000000096000010 [#1] SMP
> [ 0.406491] Modules linked in:
> [ 0.406495] CPU: 5 UID: 0 PID: 106 Comm: kworker/u32:6 Tainted: G M
> 6.18.0-rc5-next-20251113 #13 NONE
> [ 0.406499] Tainted: [M]=MACHINE_CHECK
> [ 0.406500] Hardware name: thundercomm Thundercomm RUBIK Pi
> 3/Thundercomm RUBIK Pi 3, BIOS 2025.10-rc4 10/01/2025
> [ 0.406502] Workqueue: async async_run_entry_fn
> [ 0.406508] pstate: 40400005 (nZcv daif +PAN -UAO -TCO -DIT -SSBS
> BTYPE=--)
> [ 0.428362] pc : __pi_memset_generic+0x16c/0x188
> [ 0.428366] lr : dma_direct_alloc+0x19c/0x3d0
> [ 0.428370] sp : ffff8000810e3920
> [ 0.428371] x29: ffff8000810e3920 x28: ffff000080d0f810 x27:
> ffffba4c6196ec48
> [ 0.428373] x26: ffff000080d0f810 x25: ffffba4c607b31cc x24:
> 0000000000000000
> [ 0.428375] x23: ffff000080d0f810 x22: ffff000000c00000 x21:
> ffff000082858948
> [ 0.428376] x20: 0000000000001000 x19: fffffdffc0030000 x18:
> 000000000000000a
> [ 0.428378] x17: ffff0000823dae00 x16: 0000000000000000 x15:
> 0000000000000000
> [ 0.428380] x14: 00000000ffffffff x13: 0000000000000068 x12:
> 0000000000000100
> [ 0.449344] x11: 0000000000000000 x10: ffff0001fef99500 x9 :
> 0000000000000000
> [ 0.449345] x8 : ffff000000c00000 x7 : 0000000000000000 x6 :
> 000000000000003f
> [ 0.449347] x5 : 0000000000000040 x4 : 0000000000000000 x3 :
> 0000000000000004
> [ 0.449349] x2 : 0000000000000fc0 x1 : 0000000000000000 x0 :
> ffff000000c00000
> [ 0.449350] Call trace:
> [ 0.449351] __pi_memset_generic+0x16c/0x188 (P)
> [ 0.449354] dma_alloc_attrs+0x94/0x210
> [ 0.449357] dmam_alloc_attrs+0x74/0xc0
> [ 0.469967] dw_pcie_msi_host_init+0x100/0x300
> [ 0.469971] dw_pcie_host_init+0x5e4/0x6d8
> [ 0.491913] qcom_pcie_probe+0x5a8/0x838
> [ 0.491916] platform_probe+0x64/0xc0
> [ 0.491919] really_probe+0xc8/0x3f0
> [ 0.491921] __driver_probe_device+0x88/0x170
> [ 0.491922] driver_probe_device+0x48/0x130
> [ 0.491923] __device_attach_driver+0xc4/0x190
> [ 0.491925] bus_for_each_drv+0x90/0x100
> [ 0.491928] __device_attach_async_helper+0xb8/0x120
> [ 0.491929] async_run_entry_fn+0x3c/0x1e0
> [ 0.491931] process_one_work+0x150/0x3a0
> [ 0.491934] worker_thread+0x288/0x480
> [ 0.491936] kthread+0x118/0x1e0
> [ 0.491938] ret_from_fork+0x10/0x20
> [ 0.513092] Code: 91010108 54ffff4a 8b040108 cb050042 (d50b7428)
> [ 0.513094] ---[ end trace 0000000000000000 ]---
>
> I can only get the device to boot by disabling both pcie0 and pcie1.
I think there're some regressions in "next-20251114".
After some time to "git bisect", I found after running 2 revert
commands below, it can boot for both RUBIK Pi 3 and RB3 Gen2.
$ git revert b15ce3c0882c9cd2fbe4f87047874ad087b583ff -m 1
$ git revert 03e928442d469f7d8dafc549638730647202d9ce
> > +
> > +
> > + perst-gpios = <&tlmm 2 GPIO_ACTIVE_LOW>;
> > + wake-gpios = <&tlmm 3 GPIO_ACTIVE_LOW>;
> > +
> > + pinctrl-0 = <&pcie1_clkreq_n>,
> > + <&pcie1_reset_n>,
> > + <&pcie1_wake_n>;
> > + pinctrl-names = "default";
> > +
> > + status = "okay";
> > +};
> > +
> > +&pcie1_phy {
> > + vdda-phy-supply = <&vreg_l10c_0p88>;
> > + vdda-pll-supply = <&vreg_l6b_1p2>;
> > +
> > + status = "okay";
> > +};
> > +
>
> [snip]
>
> > +
> > +&remoteproc_adsp {
> > + firmware-name = "qcom/qcs6490/adsp.mbn";
> > +
> > + status = "okay";
> > +};
>
> I'm fairly sure that this is the wrong ADSP firmware. With the firmware
> in linux-firmware, I'm seeing charger pd crashes and the ADSP constantly
> restarting. Using the Radxa Dragon Q6A ADSP firmware which disables the
> charging feature in the firmware works way better and does not result in
> crashes.
I run the Ubuntu 24.04 base system:
* https://ubuntu.com/download/qualcomm-iot#rubikpi3
Currently it boots well with adsp fw from RB3 Gen2 (from deb pkg:
firmware-qcom-hlosfw) without crash.
But I heard from next release, adsp will be customized, so Hongyang
will make another patch to upstream the adsp for RUBIK Pi 3.
Cheers,
Roger
> > +
> > +&remoteproc_cdsp {
> > + firmware-name = "qcom/qcs6490/cdsp.mbn";
> > +
> > + status = "okay";
> > +};
> > +
>
> [snip]
>
> Thanks,
> Jens
Powered by blists - more mailing lists