Rockchip RK3576 UFS 完整开发指南：DTS、LUN、Vendor Storage 与调试

本文完整整理 RK-KF-YF-C15 V1.5.0 中英文技术正文：中文第 4–15 页，英文第 4 页下半部分至第 17 页。原稿范围为 RK3576、kernel 6.1；须核对实际 ArmSoM 板卡是否引出 UFS、PHY 电源和参考时钟、SDK/BSP、U-Boot 与颗粒支持列表。本文为平台参考，未经 ArmSoM 硬件验证。四语代码以中文原稿为共同基准，中文使用中文 PDF 原图，EN/RU/DE 使用英文 PDF 对应原图。性能日志为原厂 512GB UFS 3.1 示例，不是 ArmSoM 实测。

写入 /dev/block/sda 的 FIO 测试会覆盖设备数据，不能在含有需要保留数据的盘上直接执行。LUN 配置、烧录及修改供电探测同样须按目标硬件独立验证。原文 status = <okay>; 写法保留并列为 DTS 语法疑点；HS-GEAR4 表值不代表 RK3576 支持 G4（本篇资源表只列 G1–G3）。表中脚注 (2)/(3) 没有附注正文，不推断其含义。

ArmSoM Docs · Rockchip Linux kernel

1. 芯片资源介绍

RK3576

资源	兼容UFS版本	PHY支持速率	PHY供电	备注
UFS	2.1,2.2,3.0和3.1	HS G1-G3 和 PWM G1-G3	0.85v和1.8v

注意事项:

PHY没有供电的情况下不能去初始化UFS,系统会挂死。
没有使用UFS的设备,DTS里面把UFS节点配置为“disabled”,不然会影响开机速度。

2. DTS 配置

RK3576

资源	参考配置	控制器节点	PHY节点
UFS	rk3576.dtsi	ufs: ufs@2a2d0000	–

compatible = “rockchip,rk3576-ufs”;必须配置项:默认配置,对应驱动:drivers/ufs/host/ufs-rockchip.c。
assigned-clock-parents = <&cru CLK_REF_MPHY_26M>必须配置项:默认选择PPLL输出的26MHz参考时钟。如果有使用外部独立26Mhz晶振,可以配置为: assigned-clock-parents = <&clk_gpio_mphy_i>;同时dts还需要额外指定io输入的时钟和频率:

clk_gpio_mphy_i: clk_gpio_mphy_i{
            compatible = "fixed-clock";
            #clock-cells = <0>;
            clock-frequency = <26000000>;
            clock-output-names = "clk_gpio_mphy_i";
};

status = ;必须配置项:启用UFS,如果项目没有使用UFS,建议配置为"disabled",不然会影响开机启动速度。
pinctrl-0 = <&ufs_refclk>;必须配置项:配置UFS_REFCLK为功能IO,输出参考时钟给UFS颗粒。
reset-gpios = <&gpio4 RK_PD0 GPIO_ACTIVE_HIGH>;

必须配置项:配置UFS_RSTN为GPIO,在驱动里面控制UFS颗粒的复位。

vcc-supply = <&vcc_ufs_s0>;可选配置项:配置UFS颗粒主电源。
vccq-supply = <&vcc1v2_ufs_vccq_s0>;可选配置项:配置UFS VCCQ电源。
vccq2-supply = <&vcc1v8_ufs_vccq2_s0>;可选配置项:配置UFS VCCQ2电源。
freq-table-hz = <50000000 250000000>, <0 0>, <0 0>, <0 0>;可选配置项:配置UFS dvfs调频时钟区间。每个数组元素定义了最小时钟与最大时钟,次序与dtsi中UFS节点所引用的clocks一一对应。当无IO传输时,各路时钟频率将下降到最低,开启传输前将上升到最高。如果某一路不需要动态调整,请设置其最大与最小值为0即可。

3. menuconfig 配置

需要确保如下配置打开。

CONFIG_SCSI_UFSHCD=y
CONFIG_SCSI_UFS_BSG=y
CONFIG_SCSI_UFS_HWMON=y
CONFIG_SCSI_UFSHCD_PLATFORM=y
CONFIG_SCSI_UFS_ROCKCHIP=y

4. U-boot支持UFS

默认SDK代码已经支持UFS,直接编译就可以支持UFS。

4.1 DTS修改

DTS节点的配置和kernel一样,u-boot下需要在rk3576-u-boot.dtsi里面配置启用。

&ufs {
    u-boot,dm-spl;
    status = "okay";
};

4.2 config配置

UFS需要SCSI协议支持,需要配置上SCSI协议。

CONFIG_SPL_UFS_SUPPORT=y
CONFIG_CMD_UFS=y
CONFIG_UFS=y
CONFIG_ROCKCHIP_UFS=y
CONFIG_SCSI=y
CONFIG_DM_SCSI=y

4.3 UFS MPHY供电探测

UFS MPHY在没有供电的情况下去访问寄存器会造成系统卡住,usbplug和uboot会在ufs初始化前会先探测一下UFS MPHY是否供电。但是探测代码没法兼容个别老型号的UFS颗粒,uboot代码提供关闭探测的配置选项,参考下面补丁:

diff --git a/configs/rk3576-usbplug.config b/configs/rk3576-usbplug.config
index 822967e7cf5..26cfa5956fe 100644
--- a/configs/rk3576-usbplug.config
+++ b/configs/rk3576-usbplug.config
@@ -101,3 +101,4 @@ CONFIG_SCSI=y
 CONFIG_CMD_UFS=y
 CONFIG_ROCKCHIP_UFS=y
 CONFIG_CMD_SCSI=y
+CONFIG_ROCKCHIP_UFS_DISABLED_LINKUP_TEST=y
diff --git a/configs/rk3576_defconfig b/configs/rk3576_defconfig
index 3ec1fdefe00..15d0451ddc9 100644
--- a/configs/rk3576_defconfig
+++ b/configs/rk3576_defconfig
@@ -223,3 +223,4 @@ CONFIG_RK_AVB_LIBAVB_USER=y
 CONFIG_OPTEE_CLIENT=y
 CONFIG_OPTEE_V2=y
 CONFIG_OPTEE_ALWAYS_USE_SECURITY_PARTITION=y
+CONFIG_ROCKCHIP_UFS_DISABLED_LINKUP_TEST=y

修改后需要重新编译usbplug和uboot,编译方法如下:

#echo 编译usbplug
./make.sh rk3576-usbplug
#echo 拷⻉usbplug.bin到rkbin替换对应⽂件，下⾯⽂件名只是参考
cp usbplug.bin ../rkbin/bin/rk35/rk3576_usbplug_v1.02.bin
#重新编译uboot和loader
./make.sh rk3576

5. UFS LUN定制

UFS需要先配置LUN后才可以使用,默认SDK是配置4个LUN,详细⻅下表:

LUN ID	名称	大小	FLASH 模式	是否支持 BOOT	用途
0	user(sda)	总容量 - 48MB	XLC	否	存放系统固件和用戶数据
1	boot0 (sdb)	4MB	SLC	是	存放loader
2	boot1 (sdc)	4MB	SLC	是	存放备份loader
3	data(sdd)	8MB	SLC	否	用戶可以自定义使用

如果默认配置不能满足需求,可以自己修改LUN配置,具体代码在uboot工程的函数ufs_lu_configuration中。修改后需要重新编译usbplug,编译方法如下:

#echo 编译usbplug
./make.sh rk3576-usbplug
#echo 拷⻉usbplug.bin到rkbin替换对应⽂件，下⾯⽂件名只是参考
cp usbplug.bin ../rkbin/bin/rk35/rk3576_usbplug_v1.02.bin
#重新编译uboot和loader
./make.sh rk3576

6. Vendor Storage支持

6.1 Uboot

Uboot默认支持Vendor Storage,没有限制。

6.2 Kernel

Kernel 下UFS驱动不支持Vendor Storage,但是部分驱动在kernel阶段有读需求,可以在dts里面定义“ram-vendor-storage”节点和保留内存,uboot会把读取到的数据传递给kernel驱动读使用。

diff --git a/arch/arm64/boot/dts/rockchip/rk3576-android.dtsi b/arch/arm64/boot/dts/rockchip/rk3576-android.dtsi
index 4e6574156dc0..b4ec015c2638 100644
--- a/arch/arm64/boot/dts/rockchip/rk3576-android.dtsi
+++ b/arch/arm64/boot/dts/rockchip/rk3576-android.dtsi
@@ -55,6 +55,17 @@ ramoops: ramoops@40110000 {
                        ftrace-size = <0x00000>;
                        record-size = <0x14000>;
                };
+
+               vendor_storage_rm: vendor-storage-rm@00000000 {
+                       compatible = "rockchip,vendor-storage-rm";
+                       reg = <0x0 0x0 0x0 0x0>;
+               };
+       };
+
+       vendor_storage: vendor-storage {
+               compatible = "rockchip,ram-vendor-storage";
+               memory-region = <&vendor_storage_rm>;
+               status = "okay";
        };
 };

应用层写vendor storage,可以参考android下代码“hardware/rockchip/libvendor_storage/vendor_storage_test.c”进行读写。kernel写vendor storage,需要将要写的数据通过应用层代理来实现,比较麻烦,kernel提供的写接口是写内存,不会真正保存到UFS存储里面。

7. UFS接口性能

使用FIO直接对sda设备进行读写,没有额外文件系统和加解密开销,测试UFS为512GB 3.1版本。

7.1 顺序写性能

测试命令:

/data/fio -filename=/dev/block/sda -direct=1 -iodepth 32 -thread -rw=write  -bs=1024k -size=1G -numjobs=8 -runtime=180 -group_reporting -name=seq_100write_1024k

测试结果:

seq_100write_1024k: (g=0): rw=write, bs=(R) 1024KiB-1024KiB, (W) 1024KiB-1024KiB,
(T) 1024KiB-1024KiB, ioengine=psync, iodepth=32
...
fio-2.20
Starting 8 threads
Jobs: 7 (f=1): [W(1),f(3),_(1),f(3)][100.0%][r=0KiB/s,w=966MiB/s][r=0,w=965 IOPS]
[eta 00m:00s]
seq_100read_1024k: (groupid=0, jobs=8): err= 0: pid=2389: Wed Jun 26 16:34:04
2024
  write: IOPS=946, BW=947MiB/s (993MB/s)(8192MiB/8652msec)
    clat (msec): min=3, max=32, avg= 8.25, stdev= 2.53
     lat (msec): min=3, max=32, avg= 8.42, stdev= 2.53
    clat percentiles (usec):
     |  1.00th=[ 7072],  5.00th=[ 7136], 10.00th=[ 7200], 20.00th=[ 7264],
     | 30.00th=[ 7264], 40.00th=[ 7264], 50.00th=[ 7264], 60.00th=[ 7328],
     | 70.00th=[ 7392], 80.00th=[ 9536], 90.00th=[ 9920], 95.00th=[14400],
     | 99.00th=[14912], 99.50th=[25728], 99.90th=[29824], 99.95th=[32384],
     | 99.99th=[32384]
   bw (  KiB/s): min=89495, max=126959, per=0.01%, avg=121531.16, stdev=7781.67
    lat (msec) : 4=0.02%, 10=90.32%, 20=9.12%, 50=0.54%
  cpu          : usr=2.04%, sys=5.14%, ctx=16467, majf=0, minf=0
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwt: total=0,8192,0, short=0,0,0, dropped=0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32

Run status group 0 (all jobs):
  WRITE: bw=947MiB/s (993MB/s), 947MiB/s-947MiB/s (993MB/s-993MB/s), io=8192MiB
(8590MB), run=8652-8652msec

Disk stats (read/write):
  sda: ios=0/16382, merge=0/0, ticks=0/126485, in_queue=126485, util=98.29%

7.2 顺序读性能

测试命令:

/data/fio -filename=/dev/block/sda -direct=1 -iodepth 32 -thread -rw=read  -bs=1024k -size=1G -numjobs=8 -runtime=180 -group_reporting -name=seq_100read_1024k

测试结果:

seq_100read_1024k: (g=0): rw=read, bs=(R) 1024KiB-1024KiB, (W) 1024KiB-1024KiB,
(T) 1024KiB-1024KiB, ioengine=psync, iodepth=32
...
fio-2.20
Starting 8 threads
Jobs: 8 (f=8): [R(8)][100.0%][r=1020MiB/s,w=0KiB/s][r=1020,w=0 IOPS][eta 00m:00s]
seq_100read_1024k: (groupid=0, jobs=8): err= 0: pid=2368: Wed Jun 26 16:31:59
2024
   read: IOPS=1001, BW=1002MiB/s (1050MB/s)(8192MiB/8177msec)
    clat (msec): min=5, max=32, avg= 7.96, stdev= 1.69
     lat (msec): min=5, max=33, avg= 7.96, stdev= 1.69
    clat percentiles (usec):
     |  1.00th=[ 7584],  5.00th=[ 7712], 10.00th=[ 7712], 20.00th=[ 7776],
     | 30.00th=[ 7840], 40.00th=[ 7840], 50.00th=[ 7840], 60.00th=[ 7840],
     | 70.00th=[ 7840], 80.00th=[ 7904], 90.00th=[ 7968], 95.00th=[ 7968],
     | 99.00th=[ 8160], 99.50th=[20096], 99.90th=[33024], 99.95th=[33024],
     | 99.99th=[33024]
   bw (  KiB/s): min=96256, max=132395, per=0.01%, avg=128637.07, stdev=8204.70
    lat (msec) : 10=99.44%, 20=0.06%, 50=0.50%
  cpu          : usr=0.38%, sys=3.85%, ctx=24282, majf=0, minf=2048
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwt: total=8192,0,0, short=0,0,0, dropped=0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32

Run status group 0 (all jobs):
   READ: bw=1002MiB/s (1050MB/s), 1002MiB/s-1002MiB/s (1050MB/s-1050MB/s),
io=8192MiB (8590MB), run=8177-8177msec

Disk stats (read/write):
  sda: ios=16384/0, merge=0/0, ticks=119146/0, in_queue=119145, util=98.96%

7.3 随机写性能

测试命令:

/data/fio -filename=/dev/block/sda -direct=1 -iodepth 32 -thread -rw=randwrite  -bs=4k -size=1G -numjobs=8 -runtime=180 -group_reporting -name=rand_100write_4k

测试结果:

rand_100write_4k: (g=0): rw=randwrite, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T)
4096B-4096B, ioengine=psync, iodepth=32
...
fio-2.20
Starting 8 threads
Jobs: 7 (f=7): [_(1),w(7)][100.0%][r=0KiB/s,w=270MiB/s][r=0,w=69.0k IOPS][eta
00m:00s]
seq_100read_1024k: (groupid=0, jobs=8): err= 0: pid=2402: Wed Jun 26 16:35:55
2024
  write: IOPS=55.2k, BW=215MiB/s (226MB/s)(8192MiB/38022msec)
    clat (usec): min=35, max=18933, avg=137.53, stdev=421.10
     lat (usec): min=36, max=18935, avg=138.68, stdev=421.21
    clat percentiles (usec):
     |  1.00th=[   49],  5.00th=[   56], 10.00th=[   61], 20.00th=[   68],
     | 30.00th=[   75], 40.00th=[   82], 50.00th=[   88], 60.00th=[   96],
     | 70.00th=[  108], 80.00th=[  127], 90.00th=[  183], 95.00th=[  278],
     | 99.00th=[  532], 99.50th=[  796], 99.90th=[ 6688], 99.95th=[ 6944],
     | 99.99th=[11584]
   bw (  KiB/s): min=22220, max=39863, per=0.01%, avg=27702.29, stdev=2097.44
    lat (usec) : 50=1.25%, 100=62.00%, 250=30.71%, 500=4.86%, 750=0.66%
    lat (usec) : 1000=0.07%
    lat (msec) : 2=0.03%, 4=0.01%, 10=0.40%, 20=0.02%
  cpu          : usr=3.73%, sys=13.05%, ctx=3190681, majf=0, minf=0
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwt: total=0,2097152,0, short=0,0,0, dropped=0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32

Run status group 0 (all jobs):
  WRITE: bw=215MiB/s (226MB/s), 215MiB/s-215MiB/s (226MB/s-226MB/s), io=8192MiB
(8590MB), run=38022-38022msec

Disk stats (read/write):
  sda: ios=0/2096660, merge=0/13, ticks=0/180540, in_queue=180540, util=99.93%

7.4 随机读性能

测试命令:

/data/fio -filename=/dev/block/sda -direct=1 -iodepth 32 -thread -rw=randread  -bs=4k -size=1G -numjobs=8 -runtime=180 -group_reporting -name=rand_100read_4k

测试结果:

rand_100read_4k: (g=0): rw=randread, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T)
4096B-4096B, ioengine=psync, iodepth=32
...
fio-2.20
Starting 8 threads
[ 1393.016346][  T542] healthd: battery l=50 v=3300 t=2.6 h=2 st=3 c=-1600 fc=100
chg=au
Jobs: 8 (f=8): [r(8)][100.0%][r=223MiB/s,w=0KiB/s][r=56.0k,w=0 IOPS][eta 00m:00s]
seq_100read_1024k: (groupid=0, jobs=8): err= 0: pid=2412: Wed Jun 26 16:37:30
2024
   read: IOPS=57.5k, BW=224MiB/s (235MB/s)(8192MiB/36499msec)
    clat (usec): min=60, max=2949, avg=133.28, stdev=34.67
     lat (usec): min=61, max=2949, avg=133.69, stdev=34.78
    clat percentiles (usec):
     |  1.00th=[   95],  5.00th=[   99], 10.00th=[  103], 20.00th=[  108],
     | 30.00th=[  113], 40.00th=[  118], 50.00th=[  124], 60.00th=[  131],
     | 70.00th=[  141], 80.00th=[  153], 90.00th=[  173], 95.00th=[  197],
     | 99.00th=[  262], 99.50th=[  294], 99.90th=[  374], 99.95th=[  418],
     | 99.99th=[  580]
   bw (  KiB/s): min=27040, max=29907, per=0.01%, avg=28834.42, stdev=390.23
    lat (usec) : 100=5.06%, 250=93.56%, 500=1.36%, 750=0.02%, 1000=0.01%
    lat (msec) : 2=0.01%, 4=0.01%
  cpu          : usr=3.61%, sys=13.20%, ctx=3825319, majf=0, minf=8
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwt: total=2097152,0,0, short=0,0,0, dropped=0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32

Run status group 0 (all jobs):
   READ: bw=224MiB/s (235MB/s), 224MiB/s-224MiB/s (235MB/s-235MB/s), io=8192MiB
(8590MB), run=36499-36499msec

Disk stats (read/write):
  sda: ios=2088272/0, merge=16/0, ticks=222685/0, in_queue=222685, util=99.91%

8. UFS 设备信息查询与操作

在Linux内核的平台,可以采用ufs-utils工具对UFS设备的相关信息进行查询、操作。常用命令如下请参考:

查询颗粒健康状态

ufs-utils  desc -t 9 -p /dev/bsg/ufs-bsg0  #Linux平台
ufs-utils  desc -t 9 -p /dev/ufs-bsg0      #Android平台
Device Health Descriptor: [Byte offset 0x0]: bLength = 0x25
Device Health Descriptor: [Byte offset 0x1]: bDescriptorType = 0x9
Device Health Descriptor: [Byte offset 0x2]: bPreEOLInfo = 0x1
Device Health Descriptor: [Byte offset 0x3]: bDeviceLifeTimeEstA = 0x1
Device Health Descriptor: [Byte offset 0x4]: bDeviceLifeTimeEstB = 0x1

bPreEOLInfo：
Pre End of Life Information
00h: Not defined
01h: Normal
02h: Warning. Consumed 80% of reserved blocks.
03h: Critical. Consumed 90% of reserved blocks.
Others: Reserved

bDeviceLifeTimeEstA
bDeviceLifeTimeEstB：
provides an indication of the device life time  based on the amount of performed
program/erase cycles
00h: Information not available
01h: 0% - 10% device life time used
02h: 10% - 20% device life time used
03h: 20% - 30% device life time used
04h: 30% - 40% device life time used
05h: 40% - 50% device life time used
06h: 50% - 60% device life time used
07h: 60% - 70% device life time used
08h: 70% - 80% device life time used
09h: 80% - 90% device life time used
0Ah: 90% - 100% device life time used
0Bh: Exceeded its maximum estimated device life time
Others: Reserved

查询当前颗粒运行的协议版本,速度,lane数量

#查询当前双⽅运⾏的协议版本：
ufs-utils  desc -t 9 -p /dev/bsg/ufs-bsg0  #Linux平台
ufs-utils  desc -t 9 -p /dev/ufs-bsg0      #Android平台

#查询链接状态
ufs-utils uic -t 1 -a -p /dev/bsg/ufs-bsg0 #Linux平台
ufs-utils uic -t 1 -a -p /dev/ufs-bsg0     #Android平台

1. 速度(HS Gear接⼝速度请查看附录)：
Gear: 1：Gear1  2:Gear2  3:Gear3  4:Gear4
HSSeries: 1: rate A  2: rate B
[0x1568]PA_TxGear                : local = 0x00000001, peer = 0x00000001
[0x1583]PA_RxGear                : local = 0x00000001, peer = 0x00000001
[0x156a]PA_HSSeries              : local = 0x00000002, peer = 0x00000002

2. Lane数量：
双⽅可以提供的物理层lane数量
[0x1520]PA_AvailTxDataLanes      : local = 0x00000002, peer = 0x00000002
[0x1540]PA_AvailRxDataLanes      : local = 0x00000002, peer = 0x00000002
双⽅实际物理层链接上的lane数量
[0x1561]PA_ConnectedTxDataLanes  : local = 0x00000002, peer = 0x00000002
[0x1581]PA_ConnectedRxDataLanes  : local = 0x00000002, peer = 0x00000002
双⽅实际通信使⽤到的lane数量
[0x1560]PA_ActiveTxDataLanes     : local = 0x00000001, peer = 0x00000001
[0x1580]PA_ActiveRxDataLanes     : local = 0x00000001, peer = 0x00000001

如果双⽅物理层链接上lane数量⼩于双⽅可以提供的物理层lane数量，则说明信号不佳导致有lane训练失
败，或者实际硬件只接了⼀个lane。
如果双⽅实际通信使⽤的lane数量⼩于双⽅实际物理层链接上的lane数量，说明配置信息被修改过，可以是
ufs驱动修改pwr_info或者通过ufs-utils修改。

上述查询中 Gear 的 1/2/3/4 对应 Gear1/Gear2/Gear3/Gear4，HSSeries 的 1/2 对应 rate A/rate B。PA_AvailTxDataLanes/PA_AvailRxDataLanes 表示双方可提供的物理层 lane 数量，PA_ConnectedTxDataLanes/PA_ConnectedRxDataLanes 表示实际连接数量，PA_ActiveTxDataLanes/PA_ActiveRxDataLanes 表示实际通信使用数量。连接数量少于可提供数量说明信号不佳导致训练失败或硬件仅接一个 lane；使用数量少于连接数量说明配置被修改，可通过 ufs 驱动的 pwr_info 或 ufs-utils 修改。

9. 常⻅问题

9.1 升级固件后不能启动,GPT报错

UFS的块大小是4K,GPT需要按4K的块大小生成,工具需要更新到下面对应版本或者更新的版本:

RKDevTool_v3.28_for_window

FactoryTool_v1.88

upgrade_tool_v2.30_for_linux

9.2 升级固件后不能启动,串口没有任何打印

“BOOT MODE CONFIG”配置不对,可以参考电路原理图里面的“BOOT MODE CONFIG”表格配置为UFS启动。使用了不支持的UFS颗粒,可以换个全新的其他型号UFS颗粒试试。

9.3 在Kernel老化过程中,需要统计异常信息可以查看如下节点

bash-5.2# cat /sys/kernel/debug/ufshcd/2a2d0000.ufs/stats
PHY Adapter Layer errors (except LINERESET): 0  #UECPA类型
Data Link Layer errors: 0                       #UECDL类型
Network Layer errors: 0                         #UECN类型
Transport Layer errors: 0                       #UECT类型
Generic DME errors: 0                           #UECDME类型
Auto-hibernate errors: 0                        #UHES和UHXS类型
IS Fatal errors (CEFES, SBFES, HCFES, DFES): 0
DME Link Startup errors: 0                      #DME link startup错误
PM Resume errors: 0                             #唤醒失败
PM Suspend errors : 0                           #待机失败
Logical Unit Resets: 1                          #复位设备次数，开机、休眠唤醒、出错处理
均会增加
Host Resets: 0                                  #复位控制器重新初始化次数，出错会增加
SCSI command aborts: 0                          #SCSI命令通信出错次数

9.4 Kernel启动后无法挂载分区

[    3.220417]  sda: sda1 sda2 sda3 sda4 sda5 sda6 sda7 sda8 sda9
[    3.221108] sd 0:0:0:0: [sda] Attached SCSI disk
[    3.467971] EXT4-fs (sda7): bad block size 1024
[    3.468341] EXT4-fs (sda7): bad block size 1024
[    3.468528] EXT4-fs (sda7): bad block size 1024
[    3.468733] FAT-fs (sda7): utf8 is not a recommended IO charset for FAT
filesystems, filesystem will be case sensitive!
[    3.468903] ISOFS: unsupported/invalid hardware sector size 4096
[    3.470737] EXT4-fs (sda7): bad block size 1024
[    3.471013] EXT4-fs (sda7): bad block size 1024
[    3.471260] EXT4-fs (sda7): bad block size 1024
[    3.471494] FAT-fs (sda7): utf8 is not a recommended IO charset for FAT
filesystems, filesystem will be case sensitive!
[    3.471659] ISOFS: unsupported/invalid hardware sector size 4096
[    3.472991] List of all partitions:
[    3.473009] 0100            4096 ram0
[    3.473014]  (driver?)
[    3.473035] 0800       124936192 sda
[    3.473039]  driver: sd
[    3.473050]   0801            8192 sda1 17b18c43-1d24-4484-8883-d0c618e42411
[    3.473054]
[    3.473065]   0802            8192 sda2 97c64638-8824-46d3-d83e-582a6a48d4fb
[    3.473070]
[    3.473080]   0803            4096 sda3 e6edc612-f94f-44b7-b9f1-1561474fafbb
[    3.473084]
[    3.473095]   0804           65536 sda4 d04a9010-9d09-4d47-96ba-3fa854d3c5c5
[    3.473099]
[    3.473110]   0805          131072 sda5 f2d9eb7f-2700-4ba8-9422-0e502db7c35c
[    3.473114]
[    3.473124]   0806           32768 sda6 1382357d-a655-48e7-f1da-99e05054ea34
[    3.473129]
[    3.473139]   0807        14680064 sda7 614e0000-0000-4b53-8000-1d28000054a9
[    3.473143]
[    3.473153]   0808          131072 sda8 d5f63744-ff57-4447-c9ee-b98e63e9b69b
[    3.473158]
[    3.473168]   0809       109871084 sda9 ec8c0d6e-d426-462e-bfc0-366c6e064804
[    3.473172]
[    3.473185] 0810            4096 sdb
[    3.473189]  driver: sd
[    3.473201] 0820            4096 sdc
[    3.473205]  driver: sd
[    3.473217] 0830            8192 sdd

发生故障的原因是分区镜像制作时候,制作工具指定了错误的文件系统的block size。如上述log中,ext4镜像是按1KB的block size来制作的,而UFS需要支持4KB的block size。

9.5 U-Boot探测UFS失败

当系统支持双存储启动,如eMMC+UFS, 且UFS不是作为主存储器件的情况下,若探测失败将会导致U-Boot关闭DTB中对于UFS的支持。使得进入内核之后,也将看不到UFS控制器初始化的打印。这种情况一般是UFS硬件出问题,先确保颗粒是我们AVL列表中支持的,并且在maskrom升级模式下开过卡,再按照参考原理图排查颗粒供电,外围贴件等。

Timedout waiting for UIC response
Host controller enable failed
ufshcd_pltfrm_init() failed -5
FDT: UFS was not detected, disabling UFS.

10. 附录

10.1 ufs-utils获取地址

取得redmine权限后可以直接访问https://redmine.rock-chips.com/documents/108获取ufs-utils程序。

10.2 HS Gear信息

HS Gear：中文 PDF 原图

HS Gear：中文 PDF 原图

HS-GEAR	Rate A 系列 (f_ref 19.2/26/38.4 MHz)	Rate B 系列 (f_ref 19.2/38.4 MHz)	Rate B 系列 (f_ref 26 MHz)	Rate A 系列 (来自 MIPI-M-PHY, f_ref 19.2/26/38.4 MHz)	Rate B 系列 (3) (来自 MIPI-M-PHY, f_ref 19.2/26/38.4 MHz)	单位
HS-GEAR1	1248 (2)	1459.2	1456.0	1248	1457.6	Mbps
HS-GEAR2	2496	2918.4	2912.0	2496	2915.2	Mbps
HS-GEAR3	4992	5836.8	5824.0	4992	5830.4	Mbps
HS-GEAR4	9984	11673.6	11648.0	9984	11660.8	Mbps

Rockchip RK3576 UFS 完整开发指南：DTS、LUN、Vendor Storage 与调试

1. 芯片资源介绍

2. DTS 配置

3. menuconfig 配置

4. U-boot支持UFS

4.1 DTS修改

4.2 config配置

4.3 UFS MPHY供电探测

5. UFS LUN定制

6. Vendor Storage支持

6.1 Uboot

6.2 Kernel

7. UFS接口性能

7.1 顺序写性能

7.2 顺序读性能

7.3 随机写性能

7.4 随机读性能

8. UFS 设备信息查询与操作

9. 常⻅问题

9.1 升级固件后不能启动,GPT报错

9.2 升级固件后不能启动,串口没有任何打印

9.3 在Kernel老化过程中,需要统计异常信息可以查看如下节点

9.4 Kernel启动后无法挂载分区

9.5 U-Boot探测UFS失败

10. 附录

10.1 ufs-utils获取地址

10.2 HS Gear信息

相关文章

Rockchip 分区完整说明：Parameter、GPT、CMDLINE与烧录地址

Rockchip RT-Thread SPI Flash 开发指南

Rockchip Linux SATA 完整开发指南：DTS、U-Boot、速率限制与 FBS