Device Information
System Model or SKU
BIOS VERSION
03.06
DMI release date: 2026-04-23
DIY Edition information
Memory: [fill in manufacturer and SKU if known]
Internal storage:
- WD_BLACK SN850X 8 TB
- WD PC SN740 2 TB
Affected external storage:
- Zike USB4 NVMe drive/enclosure
- Installed SSD: WD_BLACK SN850X 8 TB
- SSD firmware: 638211WD
- Enclosure controller: ASMedia ASM2464PD
- Controller PCI ID: 1b21:2463
- Exact Zike model: [fill in if known]
Port/Peripheral information
Peripheral vendor and name:
- Zike USB4 NVMe drive/enclosure
- ASMedia ASM2464PD USB4 Device Controller 40G
- WD_BLACK SN850X 8 TB installed inside
Port configuration at the time of the failure:
- Left rear: USB-C expansion card
- Left middle: probably USB-C expansion card
- Left front: USB-C expansion card
- Right rear: USB-C expansion card; the Zike USB4 NVMe drive was connected here when the issue was observed
- Right middle: either USB-C expansion card or SD card reader
- Right front: USB-A expansion card with Bluetooth mouse dongle
The issue was observed/reproduced with the Zike drive connected through Slot 4, right rear.
This does not establish that the problem is specific to Slot 4. The issue has not yet been systematically compared across other USB4-capable ports.
At the time of the failure, Slot 5, right middle, contained either a USB-C expansion card or the SD card reader; the exact module is not remembered.
Standalone Operation (Laptop Only)
Describe the bug
A Zike USB4 NVMe drive intermittently disappears from the tunneled PCIe topology.
The Zike enclosure contains a WD_BLACK SN850X 8 TB SSD and uses an ASMedia ASM2464PD USB4 controller.
The Linux kernel logs PCIe hotplug events such as:
pcieport 0000:00:01.1: pciehp: Slot(0): Link Down
pcieport 0000:00:01.1: pciehp: Slot(0): Card not present
pcieport 0000:00:01.1: PME: Spurious native interrupt!
The device later reappears:
pcieport 0000:00:01.1: pciehp: Slot(0): Card present
pcieport 0000:00:01.1: pciehp: Slot(0): Link Up
The ASMedia ASM2464PD bridges and the NVMe SSD are then re-enumerated.
During sustained ZFS/zrepl backup writes, one of these disconnects caused block-layer WRITE I/O errors. ZFS subsequently suspended the single-device destination pool.
This appears to happen below ZFS and zrepl, but I cannot yet determine whether the cause is:
- Framework BIOS or AMD USB4 host behavior
- Linux USB4 or PCIe handling
- Zike enclosure firmware or hardware
- ASMedia ASM2464PD controller behavior
- the USB4 cable
- the selected Framework port
- or an interoperability problem between these components
I am reporting this as a possible Framework USB4/peripheral interoperability issue, not as a confirmed Framework defect.
Relevant PCIe topology:
00:01.1 AMD Strix/Strix Halo PCIe USB4 Bridge [1022:150a]
01:00.0 ASMedia ASM2464PD USB4 Device Controller 40G [1b21:2463]
02:00.0 ASMedia ASM2464PD USB4 Device Controller 40G [1b21:2463]
03:00.0 Sandisk SN8000S NVMe SSD [15b7:5049]
During enumeration or re-enumeration, the kernel has also logged:
pci 0000:02:00.0: broken device, retraining non-functional downstream link at 2.5GT/s
pci 0000:03:00.0: 2.000 Gb/s available PCIe bandwidth, limited by 2.5 GT/s PCIe x1 link at 0000:00:01.1
After a later successful enumeration, the ASM2464PD downstream link and NVMe endpoint showed PCIe 16GT/s x4.
Steps To Reproduce
The issue is intermittent.
The clearest observed failure occurred during sustained writes:
- Connect the Zike USB4 NVMe drive containing the WD_BLACK SN850X 8 TB SSD.
- Import the ZFS backup pool stored on the SSD.
- Start sustained replication writes using zrepl.
- The tunneled PCIe device may disappear.
- The kernel logs Link Down and Card not present.
- The NVMe block device reports WRITE I/O errors.
- ZFS suspends the destination pool.
- The Zike enclosure and NVMe device may later re-enumerate.
Later checks performed:
-
Kept zrepl stopped.
-
Reconnected/reseated the Zike drive and USB4 path.
-
Imported the pool read-only and without mounting datasets:
zpool import -N -o readonly=on -d /dev/disk/by-id ZfsBackup0
-
The pool reported ONLINE, zero READ/WRITE/CKSUM errors, and no known data errors.
-
Exported the pool successfully.
-
Performed a raw read-only test of 64 GiB from the NVMe partition.
-
The read completed at approximately 3.7 GB/s with exit code 0 and no new kernel error during that test.
This indicates that the Zike drive and installed SSD can function normally at times, but the USB4/PCIe path intermittently disconnects.
Expected behavior
The Zike USB4 NVMe drive should remain enumerated and stable during sustained I/O.
The USB4 PCIe tunnel should not disappear, and the storage device should not produce block I/O failures because of a USB4 or PCIe link drop.
Screenshots
Not applicable.
Detailed kernel logs, lspci -vv output, NVMe information, ZFS status, and the original failure capture are available.
Operating System
Additional software versions:
zfs-2.4.2-1
zfs-kmod-2.4.2-1
zrepl v0.7.0
Additional context
The original ZFS and zrepl symptoms appear to be consequences of the lower-level disappearance of the Zike USB4 NVMe device rather than the initiating failure:
- PCIe hotplug reports Link Down and Card not present
- the Zike enclosure and NVMe device disappear or re-enumerate
- block-layer WRITE I/O errors occur
- ZFS receives zio write errors and suspends the pool
- zrepl consequently stalls at 0 B/s
Questions:
- Is this a known Framework Laptop 16 BIOS 03.06 compatibility issue with Zike USB4 NVMe drives using the ASMedia ASM2464PD controller?
- Are there Framework, AMD USB4, retimer, or PCIe-tunneling diagnostics that would help distinguish the host side from the Zike enclosure side?
- Should this instead be reported to Zike, ASMedia, or upstream Linux?
- Which additional targeted test would be most useful?
I have preserved detailed kernel logs, lspci -vv output, NVMe information, ZFS status, and the original failure capture. I can provide a redacted diagnostic bundle publicly or complete logs privately.
Device Information
System Model or SKU
BIOS VERSION
03.06
DMI release date: 2026-04-23
DIY Edition information
Memory: [fill in manufacturer and SKU if known]
Internal storage:
Affected external storage:
Port/Peripheral information
Peripheral vendor and name:
Port configuration at the time of the failure:
The issue was observed/reproduced with the Zike drive connected through Slot 4, right rear.
This does not establish that the problem is specific to Slot 4. The issue has not yet been systematically compared across other USB4-capable ports.
At the time of the failure, Slot 5, right middle, contained either a USB-C expansion card or the SD card reader; the exact module is not remembered.
Standalone Operation (Laptop Only)
Describe the bug
A Zike USB4 NVMe drive intermittently disappears from the tunneled PCIe topology.
The Zike enclosure contains a WD_BLACK SN850X 8 TB SSD and uses an ASMedia ASM2464PD USB4 controller.
The Linux kernel logs PCIe hotplug events such as:
The device later reappears:
The ASMedia ASM2464PD bridges and the NVMe SSD are then re-enumerated.
During sustained ZFS/zrepl backup writes, one of these disconnects caused block-layer WRITE I/O errors. ZFS subsequently suspended the single-device destination pool.
This appears to happen below ZFS and zrepl, but I cannot yet determine whether the cause is:
I am reporting this as a possible Framework USB4/peripheral interoperability issue, not as a confirmed Framework defect.
Relevant PCIe topology:
During enumeration or re-enumeration, the kernel has also logged:
After a later successful enumeration, the ASM2464PD downstream link and NVMe endpoint showed PCIe 16GT/s x4.
Steps To Reproduce
The issue is intermittent.
The clearest observed failure occurred during sustained writes:
Later checks performed:
Kept zrepl stopped.
Reconnected/reseated the Zike drive and USB4 path.
Imported the pool read-only and without mounting datasets:
zpool import -N -o readonly=on -d /dev/disk/by-id ZfsBackup0
The pool reported ONLINE, zero READ/WRITE/CKSUM errors, and no known data errors.
Exported the pool successfully.
Performed a raw read-only test of 64 GiB from the NVMe partition.
The read completed at approximately 3.7 GB/s with exit code 0 and no new kernel error during that test.
This indicates that the Zike drive and installed SSD can function normally at times, but the USB4/PCIe path intermittently disconnects.
Expected behavior
The Zike USB4 NVMe drive should remain enumerated and stable during sustained I/O.
The USB4 PCIe tunnel should not disappear, and the storage device should not produce block I/O failures because of a USB4 or PCIe link drop.
Screenshots
Not applicable.
Detailed kernel logs, lspci -vv output, NVMe information, ZFS status, and the original failure capture are available.
Operating System
OS/Distribution: Arch Linux
Version: rolling release
Linux Kernel Version:
Linux vbm 7.0.10-zen1-1-zen The link in "Framework Laptop 13 Intel Core Ultra Series 1 BIOS 3.04 BETA Release" is wrong #1 ZEN SMP PREEMPT_DYNAMIC Sat, 23 May 2026 14:45:01 +0000 x86_64 GNU/Linux
Additional software versions:
Additional context
The original ZFS and zrepl symptoms appear to be consequences of the lower-level disappearance of the Zike USB4 NVMe device rather than the initiating failure:
Questions:
I have preserved detailed kernel logs, lspci -vv output, NVMe information, ZFS status, and the original failure capture. I can provide a redacted diagnostic bundle publicly or complete logs privately.