rk1808节点启动失败

升级后的版本信息是什么？

升级后的版本信息是什么？确认一下是否升级成功。

bobby_jiang 发表于 2020-6-3 10:55
升级后的版本信息是什么？确认一下是否升级成功。

是通过uname -a看嘛？显示的信息如下
Linux rk1808.tb-rk1808s0 4.4.189 #1 SMP PREEMPT Tue Nov 26 11:41:47 CST 2019 aarch64 aarch64 aarch64 GNU/Linux

rpm -qa| grep rknn
rpm -qa| grep toybrick

bobby_jiang 发表于 2020-6-3 14:30
rpm -qa| grep rknn
rpm -qa| grep toybrick

[toybrick@rk1808 ~]$ rpm -qa| grep rknn
rknn-api-1.3.2-1.rk1808.fc28.aarch64
rknn-server-1.3.2-1.rk1808.fc28.aarch64
rknn-1.3.2-1.rk1808.fc28.aarch64
[toybrick@rk1808 ~]$ rpm -qa| grep toybrick
toybrick-deploy-server-1.1.0-1.rk1808.fc28.aarch64
toybrick_env-1.0.2-5.fc28.aarch64
toybrick-server-1.5.0-0.rk1808.fc28.aarch64
toybrick-deploy-1.1.0-1.rk1808.fc28.aarch64

你已经升级到最新的固件了

ssh登录计算棒后，dmesg |grep -ri galcore
把信息导给我

bobby_jiang 发表于 2020-6-4 10:57
ssh登录计算棒后，dmesg |grep -ri galcore
把信息导给我

应该是dmesg | grep -i galcore吧，显示信息如下，有一个地方显示failed
[    8.385908] galcore: loading out-of-tree module taints kernel.
[    8.390421] galcore: npu init.
[    8.393576] galcore: start npu probe.
[    8.400717] Galcore version 6.4.0.227915
[    8.400723] Galcore options:
[    8.409570] Galcore Info: ContiguousBase=0x32e00000 ContiguousSize=0x400000
[    8.414823] Galcore Info: MMU mapped external shared SRAM[0] CPU base=0xfec10000 GPU virtual address=0xfec10000 size=0x1f0000
[    8.414908] Galcore Info: MMU mapped core 0 SRAM[0] hardware virtual address=0x400000 size=0x80000
[    8.415113] Galcore Info: MMU mapped core 0 SRAM[1] hardware virtual address=0x480000 size=0x80000
[    8.419641] galcore ffbc0000.npu: Init npu devfreq
[    8.419853] galcore ffbc0000.npu: leakage=24
[    8.420099] galcore ffbc0000.npu: Failed to get pvtm
[    8.420587] galcore ffbc0000.npu: avs=0
[    8.449197] galcore ffbc0000.npu: l=0 h=2147483647 hyst=5000 l_limit=0 h_limit=0

那现在ssh登录计算棒，vi /usr/bin/start-rknn.sh,
增加
export VSI_NN_LOG_LEVEL=5
export VIV_VX_DEBUG_LEVEL=5
export RKNN_SERVER_LOGLEVEL=5

修改rknn_server > /var/log/rknn.log
保存退出。
然后tail -f /var/log/rknn.log
然后你在PC再跑一下程序，把计算棒上tail的log发给我

bobby_jiang 发表于 2020-6-4 15:08
那现在ssh登录计算棒，vi /usr/bin/start-rknn.sh,
增加
export VSI_NN_LOG_LEVEL=5

vi /usr/bin/start-rknn.sh后里面是空的，正常吗？目前已经把三行export添加进去了。

另外，“修改rknn_server > /var/log/rknn.log”是什么意思？
vi /var/log/rknn.log后内容如下：
start rknn server, version:1.3.2 (f113d64 build: 2020-04-23 09:35:26)
0 NNPluginManager loadPlugins(41): No plugin find!
I NPUTransfer: Starting NPU Transfer Server, Transfer version 2.0.0 (8f9ebbc@2020-04-03T09:12:43)