[Attempting to send to the right FreeBSD list this time.]

On Apr 5, 2025, at 09:02, Mark Millard <mark...@yahoo.com> wrote:

Steve Kargl <kargls_at_comcast.net> wrote on
Date: Sat, 05 Apr 2025 01:45:53 UTC :

> Anyone using a Samsung T7 external SSD with FreeBSD current?
> 
> If I plug the drive into a USB 2.0 port, I see
> 
> usb_msc_auto_quirk: UQ_MSC_NO_GETMAXLUN set for USB mass storage device 
> Samsung PSSD T7 Shield (0x04e8:0x61fb)
> ugen0.2: <Samsung PSSD T7 Shield> at usbus0
> umass0 on uhub1
> umass0: <Samsung PSSD T7 Shield, class 0/0, rev 2.10/1.00, addr 47> on 
> usbus0
> umass0: SCSI over Bulk-Only; quirks = 0x0100
> umass0:5:0: Attached to scbus5
> da0 at umass-sim0 bus 0 scbus5 target 0 lun 0
> da0: <Samsung PSSD T7 Shield 0> Fixed Direct Access SPC-4 SCSI device
> da0: Serial Number S6NPNS0Y201077Y
> da0: 40.000MB/s transfers

USB2 port connections always show <= 40 MiByte/sec.

Maximums:

USB 2.0 High Speed 480   Mbps
USB 1.1 Full speed  12   Mbps
USB 1.0 Low  speed   1.5 Mbps

It does not matter if the USB device supports any
of the faster USB3+ speeds:

USB 3.0 SuperSpeed          5 Gbps (FreeBSD reports 400 MiByte/sec)
USB 3.1 SuperSpeed/Gen1     5 Gbps (FreeBSD reports 400 MiByte/sec)
USB 3.1 SuperSpeed+/Gen2   10 Gbps
USB 3.2 SuperSpeed+/Gen2   10 Gbps
USB 3.2 SuperSpeed+/Gen2x2 20 Gbps (requires USB-C)
USB4 10Gbps/Gen2x2         10 Gbps (requires USB-C)?
USB4 20Gbps/Gen2x2         20 Gbps (requires USB-C)
USB4 40Gbps/Gen3x2         40 Gbps (requires USB-C)
USB4 80Gbps/Gen3x2         80 Gbps (requires USB-C)

Note: FreeBSD never reports more than 400 MB/sec in
the "da*: * MB/s transfers" messages. The figure shown
is not based on measurement.

I do sometimes have connections that normally
report 400 MB/sec (so: 5 Gbps mode use)
instead report 40 MiByte/sec (so: USB 2.0
480 Mbps mode used instead).

So far as I know, FreeBSD has no explicit support for
enabling use of any 10+ Gbps modes.

FreeBSD support for USB4 varies based on if the likes
of UEFI/ACPI handles various things when the kernel
does not: FreeBSD does not have all the kernel software
that would be required. So it does not disable the
UEFI/ACPI handling. Even when UEFI/ACPI or the like
handles various things, and so enabling some USB4 use,
there may be issues with hot-unplugging and/or
hot-plugging USB4 connections.

> da0: 1907729MB (3907029168 512 byte sectors)
> da0: quirks=0x2<NO_6_BYTE>
> 
> However, the SSD is supposedly a USB 3.2 gen 2 device with a ~1000 MBps
> read/write speed.

https://semiconductor.samsung.com/us/consumer-storage/portable-ssd/t7-touch/

reports in the "*" footnote:

"To reach maximum read/write speeds of up to 1,050/1,000 MB/s,
respectively,the host device and connection cables must
support USB 3.2 Gen 2 and the UASP mode must be enabled."

FreeBSD does not support UASP mode use: just "SCSI over
Bulk-Only". So do not expect 1,050/1,000 MB/s even if
everything else about the hardware environment could
support it.

(Such consequences for lack of UASP is not unique to
the T7 Touch.)

> When plugged into a USB 3.x port, I typically see
> 
> ugen0.2: <vendor 0x0507 product 0x0204> at usbus0
> 
> and the device is not listed with usbconfig. Repeatedly, unplugging the
> ssd and re-plugging it into the USB 3.x port, I eventually get the above 
> dmesg output.

Seems like a possible cable / signal integrity problem.

In my use of T7 Touches some hosts are picky about which
type of cable I use. For 2 of them, the pre-supplied
cable does not work reliably. But the cables that work
for those, do not generally work well elsewhere.

". . . the internal resistance and maximum allowable
current can vary depending on which type of cable you are
using."

> Do I need a quirk for this SSD to get recognized?

I've never had to adjust the quirk status.

But on some RPi*'s I've had to have U-Boot use
usb_pgood_delay adjustments to reliably boot. (An issue
that happens before even the FreeBSD loader is involved.)

> Also,
> shouldn't it connect with faster transfer rate than 'da0: 40.0MB/s'?

See earlier notes above.

===
Mark Millard
marklmi at yahoo.com


===
Mark Millard
marklmi at yahoo.com


Reply via email to