Re: [PATCH] ARM: dts: socfpga: remove syscon compatible string for sysmgr node

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2/10/25 21:18, Rabara, Niravkumar L wrote:
Hi Dinh

-----Original Message-----
From: Dinh Nguyen <dinguyen@xxxxxxxxxx>
Sent: Tuesday, 11 February, 2025 8:07 AM
To: Rabara, Niravkumar L <niravkumar.l.rabara@xxxxxxxxx>; Rob Herring
<robh@xxxxxxxxxx>; Krzysztof Kozlowski <krzk+dt@xxxxxxxxxx>; Conor Dooley
<conor+dt@xxxxxxxxxx>; devicetree@xxxxxxxxxxxxxxx; linux-
kernel@xxxxxxxxxxxxxxx
Cc: lkp <lkp@xxxxxxxxx>
Subject: Re: [PATCH] ARM: dts: socfpga: remove syscon compatible string for
sysmgr node

On 2/10/25 16:29, Dinh Nguyen wrote:
On 1/17/25 09:42, niravkumar.l.rabara@xxxxxxxxx wrote:
From: Niravkumar L Rabara <niravkumar.l.rabara@xxxxxxxxx>

The SoCFPGA System Manager(sysmgr) dt bindings do not use the syscon
compitible, nor does the Linux system manager driver rely on it.
Remove "syscon" for Arria5, Arria10 and Cyclon5 sysmgr node and fixed
dtbs_check warnings like:

socfpga_arria5_socdk.dtb: sysmgr@ffd08000: compatible: 'oneOf'
conditional failed, one must be fixed:
     ['altr,sys-mgr', 'syscon'] is too long
     'altr,sys-mgr-s10' was expected
     'altr,sys-mgr' was expected
     from schema $id:
http://devicetree.org/schemas/soc/altera/altr,sys-mgr.yaml#

Reported-by: kernel test robot <lkp@xxxxxxxxx>
Closes:
https://lore.kernel.org/oe-kbuild-all/202501102323.Xnte2yhi-lkp@intel
.com/
Signed-off-by: Niravkumar L Rabara <niravkumar.l.rabara@xxxxxxxxx>
---
   arch/arm/boot/dts/intel/socfpga/socfpga.dtsi         | 2 +-
   arch/arm/boot/dts/intel/socfpga/socfpga_arria10.dtsi | 2 +-
   2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/arm/boot/dts/intel/socfpga/socfpga.dtsi
b/arch/arm/boot/dts/intel/socfpga/socfpga.dtsi
index 35be14150f41..f124fb72e260 100644
--- a/arch/arm/boot/dts/intel/socfpga/socfpga.dtsi
+++ b/arch/arm/boot/dts/intel/socfpga/socfpga.dtsi
@@ -853,7 +853,7 @@ spi1: spi@fff01000 {
           };
           sysmgr: sysmgr@ffd08000 {
-            compatible = "altr,sys-mgr", "syscon";
+            compatible = "altr,sys-mgr";
               reg = <0xffd08000 0x4000>;
           };
diff --git a/arch/arm/boot/dts/intel/socfpga/socfpga_arria10.dtsi
b/arch/arm/boot/dts/intel/socfpga/socfpga_arria10.dtsi
index 6b6e77596ffa..015120fb4b02 100644
--- a/arch/arm/boot/dts/intel/socfpga/socfpga_arria10.dtsi
+++ b/arch/arm/boot/dts/intel/socfpga/socfpga_arria10.dtsi
@@ -792,7 +792,7 @@ scu: snoop-control-unit@ffffc000 {
           };
           sysmgr: sysmgr@ffd06000 {
-            compatible = "altr,sys-mgr", "syscon";
+            compatible = "altr,sys-mgr";
               reg = <0xffd06000 0x300>;
               cpu1-start-addr = <0xffd06230>;
           };

Did you test this patch on actual hardware? Unless something has
changed in the system manager driver, this will probably cause the system
hang.


Actually, it will not fail to boot, but you will see SD/MMC fail if the bootloader did
not set the clk-phase correctly, or you booted from another source not SD/MMC.
The SD/MMC driver uses syscon to get access to the system manager to set it's
clk-phase.


Yes, I have tested this using NFS boot, however I didn't observe any issue with SD/MMC
driver.

=> fdt print /soc/mmc@ff808000
mmc@ff808000 {
	#address-cells = <0x00000001>;
	#size-cells = <0x00000000>;
	compatible = "altr,socfpga-dw-mshc";
	reg = <0xff808000 0x00001000>;
	interrupts = <0x00000000 0x00000062 0x00000004>;
	fifo-depth = <0x00000400>;
	clocks = <0x0000001a 0x00000024>;
	clock-names = "biu", "ciu";
	resets = <0x00000006 0x00000027>;
	altr,sysmgr-syscon = <0x0000001c 0x00000028 0x00000004>;
	status = "okay";
	cap-sd-highspeed;
	cap-mmc-highspeed;
	broken-cd;
	bus-width = <0x00000004>;
	clk-phase-sd-hs = <0x00000000 0x00000087>;
	phandle = <0x00000029>;
};
=> fdt print /soc/sysmgr@ffd06000
sysmgr@ffd06000 {
	compatible = "altr,sys-mgr";
	reg = <0xffd06000 0x00000300>;
	cpu1-start-addr = <0xffd06230>;
	phandle = <0x0000001c>;
};

.

[    1.095784] mmc_host mmc0: Bus speed (slot 0) = 50000000Hz (slot req 50000000Hz, actual 50000000HZ div = 0)
[    1.105692] mmc0: new high speed SDHC card at address 0001
[    1.108238] at24 0-0051: supply vcc not found, using dummy regulator
[    1.111817] mmcblk0: mmc0:0001 SD32G 29.1 GiB
[    1.118872] at24 0-0051: 4096 byte 24c32 EEPROM, writable, 32 bytes/write
[    1.129186]  mmcblk0: p1 p2 p3

.

root@arria10:~# ls /dev/mmcblk0*
/dev/mmcblk0    /dev/mmcblk0p1  /dev/mmcblk0p2  /dev/mmcblk0p3
root@arria10:~# mount /dev/mmcblk0p1 /tmp/
root@arria10:~# ls /tmp/
extlinux                         socfpga_arria10_socdk_sdmmc.dtb  zImage
fit_spl_fpga.itb                 u-boot.img



You didn't really test anything. There's a register in the System Manager that has set the SD/MMC clk-phase in U-Boot. So you won't see the failure unless you explicitly change the value in that register and then boot Linux, then you will see the failure. If you look at drivers/mmc/host/dw_mmc-pltfm.c and look at the function dw_mci_socfpga_priv_init(), you can see that work in action. If you remove the syscon property, then that portion of the driver will fail. Also the ethernet driver is using the system manager's to set the correct PHY mode through syscon. I think you need to test this patch more thoroughly.

DInh





[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux