DietPi-Drive_Manager | NVMe support/tests

[Fourdee]

Hardware limited write speed on NanoPC-T4/RK (using 970 EVO)

6 partitions?

dd if=/dev/zero of=/dev/nvme0n1 bs=4K count=4000

#reboot

nvme0n1      259:0    0 232.9G  0 disk
├─nvme0n1p1  259:1    0     4M  0 part
├─nvme0n1p2  259:2    0     4M  0 part
├─nvme0n1p3  259:3    0    12M  0 part
├─nvme0n1p4  259:4    0    32M  0 part
├─nvme0n1p5  259:5    0    32M  0 part
└─nvme0n1p6  259:6    0 232.5G  0 part

Samsung/NanoPC-T4 only?

https://itpeernetwork.intel.com/finding-your-new-intel-ssd-for-pcie-think-nvme-not-scsi/
nvme0n1 259:0 0 372.6G 0 disk
└─nvme0n1p1 259:1 0 372.6G 0 part

[MichaIng]

Ah this confirms my question in the other topic. I will fix drive_manager accordingly, as currently it might try take /dev/nvme0 as drive instead of /dev/nvme0n1.

Strange about the partition though... If these other ones are somehow always created, we should ignore them in drive_manager. Never found that case when doing naming scheme research 🤔.

[Fourdee]

@MichaIng

Just finished, will send commit.

To do:

• Figure out if the 6th partition is limited to Samsung or NanoPC-T4 only: https://github.com/Fourdee/DietPi/issues/2102#issue-366001513

dd if=/dev/zero of=/dev/nvme0n1 bs=4K count=10000
reboot

Interesting, appears to match the same as MMC layout:

root@DietPi:~# lsblk /dev/nvme0n1
NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
nvme0n1     259:0    0 232.9G  0 disk
├─nvme0n1p1 259:1    0     4M  0 part
├─nvme0n1p2 259:2    0     4M  0 part
├─nvme0n1p3 259:3    0    12M  0 part
├─nvme0n1p4 259:4    0    32M  0 part
├─nvme0n1p5 259:5    0    32M  0 part
└─nvme0n1p6 259:6    0 232.5G  0 part /mnt/nvme
root@DietPi:~# lsblk /dev/mmcblk1
NAME        MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
mmcblk1     179:0    0 14.6G  0 disk
├─mmcblk1p1 179:1    0    4M  0 part
├─mmcblk1p2 179:2    0    4M  0 part
├─mmcblk1p3 179:3    0   12M  0 part
├─mmcblk1p4 179:4    0   32M  0 part
├─mmcblk1p5 179:5    0   32M  0 part
└─mmcblk1p6 179:6    0 14.2G  0 part /

Appears to be NanoPC-T4 specific? Sent email to FE, waiting for response.

[Fourdee]

@MichaIng

First integer might be controller number, second drive number and only third is partition number.

Yep, seems to make sense 👍

Zero all partitions, then disk

Seems NanoPC-T4 specific, i've already added a workaround for this in code.

root@DietPi:~# dd if=/dev/zero of=/dev/nvme0n1p6 bs=4K count=10000
10000+0 records in
10000+0 records out
40960000 bytes (41 MB, 39 MiB) copied, 0.233267 s, 176 MB/s
root@DietPi:~# dd if=/dev/zero of=/dev/nvme0n1p5 bs=4K count=10000
dd: error writing '/dev/nvme0n1p5': No space left on device
8193+0 records in
8192+0 records out
33554432 bytes (34 MB, 32 MiB) copied, 0.177884 s, 189 MB/s
root@DietPi:~# dd if=/dev/zero of=/dev/nvme0n1p4 bs=4K count=10000
dd: error writing '/dev/nvme0n1p4': No space left on device
8193+0 records in
8192+0 records out
33554432 bytes (34 MB, 32 MiB) copied, 0.171098 s, 196 MB/s
root@DietPi:~# dd if=/dev/zero of=/dev/nvme0n1p3 bs=4K count=10000
dd: error writing '/dev/nvme0n1p3': No space left on device
3073+0 records in
3072+0 records out
12582912 bytes (13 MB, 12 MiB) copied, 0.11826 s, 106 MB/s
root@DietPi:~# dd if=/dev/zero of=/dev/nvme0n1p2 bs=4K count=10000
dd: error writing '/dev/nvme0n1p2': No space left on device
1025+0 records in
1024+0 records out
4194304 bytes (4.2 MB, 4.0 MiB) copied, 0.0477202 s, 87.9 MB/s
root@DietPi:~# dd if=/dev/zero of=/dev/nvme0n1p1 bs=4K count=10000
dd: error writing '/dev/nvme0n1p1': No space left on device
1025+0 records in
1024+0 records out
4194304 bytes (4.2 MB, 4.0 MiB) copied, 0.0482256 s, 87.0 MB/s
root@DietPi:~# dd if=/dev/zero of=/dev/nvme0n1 bs=4K count=10000
10000+0 records in
10000+0 records out
40960000 bytes (41 MB, 39 MiB) copied, 0.236246 s, 173 MB/s

#Just to confirm non-disk
root@DietPi:~# dd if=/dev/zero of=/dev/nvme0 bs=4K count=10000
dd: error writing '/dev/nvme0': Invalid argument
1+0 records in
0+0 records out
0 bytes copied, 0.002604 s, 0.0 kB/s

Reboot

root@DietPi:~# lsblk
NAME         MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
mmcblk1      179:0    0  14.6G  0 disk
├─mmcblk1p1  179:1    0     4M  0 part
├─mmcblk1p2  179:2    0     4M  0 part
├─mmcblk1p3  179:3    0    12M  0 part
├─mmcblk1p4  179:4    0    32M  0 part
├─mmcblk1p5  179:5    0    32M  0 part
└─mmcblk1p6  179:6    0  14.2G  0 part /
mmcblk1boot0 179:32   0     4M  1 disk
mmcblk1boot1 179:64   0     4M  1 disk
mmcblk1rpmb  179:96   0     4M  0 disk
nvme0n1      259:0    0 232.9G  0 disk
├─nvme0n1p1  259:1    0     4M  0 part
├─nvme0n1p2  259:2    0     4M  0 part
├─nvme0n1p3  259:3    0    12M  0 part
├─nvme0n1p4  259:4    0    32M  0 part
├─nvme0n1p5  259:5    0    32M  0 part
└─nvme0n1p6  259:6    0 232.5G  0 part
root@DietPi:~# blkid
/dev/mmcblk1p6: UUID="2899cd3f-0af0-4395-934f-e72272e315e2" TYPE="ext4"

[Fourdee]

And we need to implement a Random R/W benchmark.
https://twitter.com/DietPi_/status/1047191607248936962

for i < 1000

dd if=/dev/zero of=/mnt/.bench_$i bs=4k count=1 & #probably finish before next loop... And, sequential result?

[james-b-nz]

Hi,
I just installed DietPi on a NanoPC T4 with an Intel NVME 600p SSD. I can confirm that the 6 partitions exist by default for this configuration as well:

root@DietPiNanoPCT4:~# lsblk /dev/nvme0n1
NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
nvme0n1     259:0    0 119.2G  0 disk
├─nvme0n1p1 259:1    0     4M  0 part
├─nvme0n1p2 259:2    0     4M  0 part
├─nvme0n1p3 259:6    0    12M  0 part
├─nvme0n1p4 259:8    0    32M  0 part
├─nvme0n1p5 259:9    0    32M  0 part
└─nvme0n1p6 259:10   0 118.9G  0 part /mnt/nvme

For anyone interested the results in the DietPi benchmark for the above drive is:

[Fourdee]

@james-b-nz

I just installed DietPi on a NanoPC T4 with an Intel NVME 600p SSD. I can confirm that the 6 partitions exist by default for this configuration as well:

Thank you for testing and confirming 👍

Ticket now resolved.

[Fourdee]

Friendly ARM have confirmed this is an issue with the kernel configuration. They have fixed it here:
https://github.com/friendlyarm/kernel-rockchip/commit/a3c93d1b469401211e66fdec965810c1a41087f9

@carlosedp

Any chance you'd be willing to update the NanoPC-T4 kernel for our images, to include the above fix?

Its not critical, as we have a workaround in DietPi code to compensate for block layout, however, would be nice to have it updated.

[carlosedp]

Sorry about the delay guys, I'll build the latest Kernel from Rockchip, 4.4.154. Just finishing up some stuff on my build VM.

[james-b-nz]

So the default kernel config has the Mali driver configured for Android only, I've rebuilt 4.4.143 with the correct options.
Installed the XFCE desktop from dietpi-software and logged in as below:

I included openvswitch and ip_tables for @carlosedp and similar setups although I was aiming to use it with a GUI. If anyone wants any other modules compiled in let me know and I can update it.

Files for testing with updated boot.img, resource.img, kernel.img and modules in a tar.gz to copy to rootfs:
https://wetransfer.com/downloads/a358f75410b3caeb9ddf26dce1e9d25820181012002709/d2cb9426e31c00c862f39ccadb62eb2720181012002709/c95aab

[carlosedp]

I've uploaded the Rockchip upstream Kernel version 4.4.154 on https://we.tl/t-241Gjd7EJN. Signaled @Fourdee to edit the install packages. Thanks @james-b-nz !

[Fourdee]

@carlosedp

I've uploaded the Rockchip upstream Kernel version 4.4.154

Legend, thank you 👍

Reopened to update images.

---

Note to self:

• NanoPC, SD install, update rootfs to 6.16. Read, package rootfs for RK.

[Fourdee]

Completed.