LTFS-backends make errors ubuntu 18.04

[p4xel]

Greetings.
I am trying to compile the backend driver.
I ran autogen and configure successfully without errors, but unfortunately make fails with the following:

root@cds-Precision-T7610:~/ltfs-backends# make
make  all-recursive
make[1]: Entering directory '/root/ltfs-backends'
Making all in messages
make[2]: Entering directory '/root/ltfs-backends/messages'
Processing driver_ltotape
genrb number of files: 3
ar: creating ./libdriver_ltotape.a
make[2]: Leaving directory '/root/ltfs-backends/messages'
Making all in src/tape_drivers/linux/ltotape
make[2]: Entering directory '/root/ltfs-backends/src/tape_drivers/linux/ltotape'
/bin/bash ../../../../libtool  --tag=CC --tag=disable-static  --mode=compile gcc -DHAVE_CONFIG_H   -D_GNU_SOURCE -I/usr/local/include/ltfs -D_FILE_OFFSET_BITS=64 -I/usr/include/fuse -DIBM_LTFS_BUILD -DGENERIC_OEM_BUILD -I/usr/include  -I../../../../src   -g -O2 -MT libdriver_ltotape_la-ltotape.lo -MD -MP -MF .deps/libdriver_ltotape_la-ltotape.Tpo -c -o libdriver_ltotape_la-ltotape.lo `test -f 'ltotape.c' || echo './'`ltotape.c
libtool: compile:  gcc -DHAVE_CONFIG_H -D_GNU_SOURCE -I/usr/local/include/ltfs -D_FILE_OFFSET_BITS=64 -I/usr/include/fuse -DIBM_LTFS_BUILD -DGENERIC_OEM_BUILD -I/usr/include -I../../../../src -g -O2 -MT libdriver_ltotape_la-ltotape.lo -MD -MP -MF .deps/libdriver_ltotape_la-ltotape.Tpo -c ltotape.c  -fPIC -DPIC -o .libs/libdriver_ltotape_la-ltotape.o
ltotape.c: In function ‘ltotape_remaining_capacity’:
ltotape.c:1456:62: warning: passing argument 3 of ‘ltotape_logsense’ makes integer from pointer without a cast [-Wint-conversion]
   status = ltotape_logsense (device, LOG_PAGE_TAPE_CAPACITY, logdata, LOGSENSEPAGE);
                                                              ^~~~~~~
ltotape.c:1381:5: note: expected ‘uint8_t {aka const unsigned char}’ but argument is of type ‘unsigned char *’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
In file included from ../ibmtape/IBM_tape.h:1:0,
                 from ltotape_diag.h:71,
                 from ltotape.c:43:
/usr/local/include/ltfs/tape_drivers/ibm_tape.h:169:22: warning: passing argument 4 of ‘ltotape_logsense’ makes pointer from integer without a cast [-Wint-conversion]
 #define LOGSENSEPAGE 1024     /* max data xfer for log sense page ioctl */
                      ^
ltotape.c:1456:71: note: in expansion of macro ‘LOGSENSEPAGE’
   status = ltotape_logsense (device, LOG_PAGE_TAPE_CAPACITY, logdata, LOGSENSEPAGE);
                                                                       ^~~~~~~~~~~~
ltotape.c:1381:5: note: expected ‘unsigned char *’ but argument is of type ‘int’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
ltotape.c:1456:12: error: too few arguments to function ‘ltotape_logsense’
   status = ltotape_logsense (device, LOG_PAGE_TAPE_CAPACITY, logdata, LOGSENSEPAGE);
            ^~~~~~~~~~~~~~~~
ltotape.c:1381:5: note: declared here
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
ltotape.c: In function ‘ltotape_get_cartridge_health’:
ltotape.c:2932:56: warning: passing argument 3 of ‘ltotape_logsense’ makes integer from pointer without a cast [-Wint-conversion]
   rc = ltotape_logsense (device, LOG_PAGE_VOLUMESTATS, logdata, LOGSENSEPAGE);
                                                        ^~~~~~~
ltotape.c:1381:5: note: expected ‘uint8_t {aka const unsigned char}’ but argument is of type ‘unsigned char *’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
In file included from ../ibmtape/IBM_tape.h:1:0,
                 from ltotape_diag.h:71,
                 from ltotape.c:43:
/usr/local/include/ltfs/tape_drivers/ibm_tape.h:169:22: warning: passing argument 4 of ‘ltotape_logsense’ makes pointer from integer without a cast [-Wint-conversion]
 #define LOGSENSEPAGE 1024     /* max data xfer for log sense page ioctl */
                      ^
ltotape.c:2932:65: note: in expansion of macro ‘LOGSENSEPAGE’
   rc = ltotape_logsense (device, LOG_PAGE_VOLUMESTATS, logdata, LOGSENSEPAGE);
                                                                 ^~~~~~~~~~~~
ltotape.c:1381:5: note: expected ‘unsigned char *’ but argument is of type ‘int’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
ltotape.c:2932:8: error: too few arguments to function ‘ltotape_logsense’
   rc = ltotape_logsense (device, LOG_PAGE_VOLUMESTATS, logdata, LOGSENSEPAGE);
        ^~~~~~~~~~~~~~~~
ltotape.c:1381:5: note: declared here
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
ltotape.c: In function ‘ltotape_get_tape_alert’:
ltotape.c:3002:55: warning: passing argument 3 of ‘ltotape_logsense’ makes integer from pointer without a cast [-Wint-conversion]
   rc = ltotape_logsense (device, LOG_PAGE_TAPE_ALERT, logdata, LOGSENSEPAGE);
                                                       ^~~~~~~
ltotape.c:1381:5: note: expected ‘uint8_t {aka const unsigned char}’ but argument is of type ‘unsigned char *’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
In file included from ../ibmtape/IBM_tape.h:1:0,
                 from ltotape_diag.h:71,
                 from ltotape.c:43:
/usr/local/include/ltfs/tape_drivers/ibm_tape.h:169:22: warning: passing argument 4 of ‘ltotape_logsense’ makes pointer from integer without a cast [-Wint-conversion]
 #define LOGSENSEPAGE 1024     /* max data xfer for log sense page ioctl */
                      ^
ltotape.c:3002:64: note: in expansion of macro ‘LOGSENSEPAGE’
   rc = ltotape_logsense (device, LOG_PAGE_TAPE_ALERT, logdata, LOGSENSEPAGE);
                                                                ^~~~~~~~~~~~
ltotape.c:1381:5: note: expected ‘unsigned char *’ but argument is of type ‘int’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
ltotape.c:3002:8: error: too few arguments to function ‘ltotape_logsense’
   rc = ltotape_logsense (device, LOG_PAGE_TAPE_ALERT, logdata, LOGSENSEPAGE);
        ^~~~~~~~~~~~~~~~
ltotape.c:1381:5: note: declared here
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
ltotape.c: In function ‘ltotape_get_eod_status’:
ltotape.c:3091:56: warning: passing argument 3 of ‘ltotape_logsense’ makes integer from pointer without a cast [-Wint-conversion]
   rc = ltotape_logsense (device, LOG_PAGE_VOLUMESTATS, logdata, LOGSENSEPAGE);
                                                        ^~~~~~~
ltotape.c:1381:5: note: expected ‘uint8_t {aka const unsigned char}’ but argument is of type ‘unsigned char *’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
In file included from ../ibmtape/IBM_tape.h:1:0,
                 from ltotape_diag.h:71,
                 from ltotape.c:43:
/usr/local/include/ltfs/tape_drivers/ibm_tape.h:169:22: warning: passing argument 4 of ‘ltotape_logsense’ makes pointer from integer without a cast [-Wint-conversion]
 #define LOGSENSEPAGE 1024     /* max data xfer for log sense page ioctl */
                      ^
ltotape.c:3091:65: note: in expansion of macro ‘LOGSENSEPAGE’
   rc = ltotape_logsense (device, LOG_PAGE_VOLUMESTATS, logdata, LOGSENSEPAGE);
                                                                 ^~~~~~~~~~~~
ltotape.c:1381:5: note: expected ‘unsigned char *’ but argument is of type ‘int’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
ltotape.c:3091:8: error: too few arguments to function ‘ltotape_logsense’
   rc = ltotape_logsense (device, LOG_PAGE_VOLUMESTATS, logdata, LOGSENSEPAGE);
        ^~~~~~~~~~~~~~~~
ltotape.c:1381:5: note: declared here
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
In file included from ltotape.c:3259:0:
ltotape_compat.c: In function ‘ltotape_wrapper_logsense’:
ltotape_compat.c:373:47: warning: passing argument 3 of ‘ltotape_logsense’ makes integer from pointer without a cast [-Wint-conversion]
         status = ltotape_logsense(&sio, page, buf, size);
                                               ^~~
ltotape.c:1381:5: note: expected ‘uint8_t {aka const unsigned char}’ but argument is of type ‘unsigned char *’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
In file included from ltotape.c:3259:0:
ltotape_compat.c:373:52: warning: passing argument 4 of ‘ltotape_logsense’ makes pointer from integer without a cast [-Wint-conversion]
         status = ltotape_logsense(&sio, page, buf, size);
                                                    ^~~~
ltotape.c:1381:5: note: expected ‘unsigned char *’ but argument is of type ‘size_t {aka const long unsigned int}’
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
In file included from ltotape.c:3259:0:
ltotape_compat.c:373:18: error: too few arguments to function ‘ltotape_logsense’
         status = ltotape_logsense(&sio, page, buf, size);
                  ^~~~~~~~~~~~~~~~
ltotape.c:1381:5: note: declared here
 int ltotape_logsense (void *device, const uint8_t page, const uint8_t subpage,
     ^~~~~~~~~~~~~~~~
In file included from ltotape.c:3259:0:
ltotape_compat.c: At top level:
ltotape_compat.c:701:35: warning: initialization from incompatible pointer type [-Wincompatible-pointer-types]
         .logsense               = ltotape_wrapper_logsense,
                                   ^~~~~~~~~~~~~~~~~~~~~~~~
ltotape_compat.c:701:35: note: (near initialization for ‘ltotape_drive_handler.logsense’)
ltotape_compat.c:732:3: error: ‘struct tape_ops’ has no member named ‘get_block_in_buffer’; did you mean ‘get_next_block_to_xfer’?
  .get_block_in_buffer    = ltotape_get_block_in_buffer,
   ^~~~~~~~~~~~~~~~~~~
   get_next_block_to_xfer
ltotape_compat.c:732:28: warning: initialization from incompatible pointer type [-Wincompatible-pointer-types]
  .get_block_in_buffer    = ltotape_get_block_in_buffer,
                            ^~~~~~~~~~~~~~~~~~~~~~~~~~~
ltotape_compat.c:732:28: note: (near initialization for ‘ltotape_drive_handler.get_next_block_to_xfer’)
Makefile:479: recipe for target 'libdriver_ltotape_la-ltotape.lo' failed
make[2]: *** [libdriver_ltotape_la-ltotape.lo] Error 1
make[2]: Leaving directory '/root/ltfs-backends/src/tape_drivers/linux/ltotape'
Makefile:475: recipe for target 'all-recursive' failed
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory '/root/ltfs-backends'
Makefile:383: recipe for target 'all' failed
make: *** [all] Error 2

[piste-jp]

Oops, sorry this backend correction is no more maintained.

I'm not sure what you want to expect on this. But the main tree https://github.com/LinearTapeFileSystem/ltfs supports multiple vendors already.

I strongly recommend the code on the main project.

[p4xel]

I'm having problems with LTO9 tape when trying to get data off. System had a power cut while writing.
when using ltfs or ltfsck (from main repo) it returns errors:

root@Precision-T7610:~# ltfsck -z /dev/sg2
LTFS16000I Starting ltfsck, LTFS version 2.5.0.0 (Prelim), log level 2.
LTFS16088I Launched by "ltfsck -z /dev/sg2".
LTFS16089I This binary is built for Linux (x86_64).
LTFS16090I GCC version is 7.5.0.
LTFS17087I Kernel version: Linux version 5.4.0-150-generic (buildd@bos03-amd64-012) (gcc version 7.5.0 (Ubuntu 7.5.0-3ubuntu1~18.04)) #167~18.04.1-Ubuntu SMP Wed May 24 00:51:42 UTC 2023 i386.
LTFS17089I Distribution: NAME="Ubuntu".
LTFS17089I Distribution: DISTRIB_ID=Ubuntu.
LTFS17085I Plugin: Loading "sg" tape backend.
LTFS30209I Opening a device through sg-ibmtape driver (/dev/sg2).
LTFS30250I Opened the SCSI tape device 10.0.0.0 (/dev/sg2).
LTFS30207I Vendor ID is IBM     .
LTFS30208I Product ID is ULTRIUM-HH9     .
LTFS30214I Firmware revision is PA61.
LTFS30215I Drive serial is XXXXXXXXXX.
LTFS30285I The reserved buffer size of /dev/sg2 is 1048576.
LTFS30294I Setting up timeout values from RSOC.
LTFS17160I Maximum device block size is 1048576.
LTFS11330I Loading cartridge.
LTFS30252I Logical block protection is disabled.
LTFS11332I Load successful.
LTFS17157I Changing the drive setting to write-anywhere mode.
LTFS16014I Checking LTFS file system on '/dev/sg2'.
LTFS30252I Logical block protection is disabled.
LTFS30216W Length mismatch is detected. (Act = 4096, resid = 0, resid_sense = -61440).
LTFS12049E Cannot read: backend call failed (-21716).
LTFS11174E Cannot read ANSI label: read failed (-21716).
LTFS11170E Failed to read label (-21716) from partition 0.
LTFS11009E Cannot read volume: failed to read partition labels.
LTFS16080E Cannot check volume (8).
LTFS30252I Logical block protection is disabled.

I thought maybe using a different backend would make a difference.

[piste-jp]

I thought maybe using a different backend would make a difference.

No, the backend code here is supporting HP's driver. So I believe it doesn't work at all even if you can build the code here.

[piste-jp]

For the problem now you are facing might be a FAQ. Please search the LTFS project with LTFS30216W. I believe you can find LinearTapeFileSystem/ltfs#383.

It would be an uncomfortable HBA for LTFS. I recommend a HBA that is in the support list provided from IBM.

Finally, I'm little bit curious because you are trying to recover the tape in spite of failing to read teh first record in tape. I believe no tape can be mounted in this system... Is this tape written in another system?

[p4xel]

This tape was created on Windows system by our customer.
I have LSI SAS contoller in my system, but after recompiling the driver with --enable-buggy-ifs did not help
LSI Logic / Symbios Logic SAS2116 PCI-Express Fusion-MPT SAS-2 [Meteor] (rev 02)

I dumped label and index using dd

<?xml version="1.0" encoding="UTF-8"?>
<ltfslabel version="2.4.0">
    <creator>IBM LTFS 2.4.4.0 (10470) - Windows - mkltfs</creator>
    <formattime>2023-10-23T12:24:29.851000000Z</formattime>
    <volumeuuid>25eaa1c8-517d-41fe-ab0b-ae70754d1903</volumeuuid>
    <location>
        <partition>b</partition>
    </location>
    <partitions>
        <index>a</index>
        <data>b</data>
    </partitions>
    <blocksize>524288</blocksize>
    <compression>true</compression>
</ltfslabel>

and also index

<?xml version="1.0" encoding="UTF-8"?>
<ltfsindex version="2.4.0">
<creator>IBM LTFS 2.4.4.0 (10470) - Windows - mkltfs - Format</creator>
<volumeuuid>25eaa1c8-517d-41fe-ab0b-ae70754d1903</volumeuuid>
<generationnumber>1</generationnumber>
<updatetime>2023-10-23T13:04:22.910000000Z</updatetime>
<location>
<partition>b</partition>
<startblock>5</startblock>
</location>
<allowpolicyupdate>true</allowpolicyupdate>
<highestfileuid>1</highestfileuid>
<volumelockstate>unlocked</volumelockstate>
<directory>
<name>PBL007</name>
<readonly>false</readonly>
<creationtime>2023-10-23T12:24:29.851000000Z</creationtime>
<changetime>2023-10-23T12:24:29.851000000Z</changetime>
<modifytime>2023-10-23T12:24:29.851000000Z</modifytime>
<accesstime>2023-10-23T12:24:29.851000000Z</accesstime>
<backuptime>2023-10-23T12:24:29.851000000Z</backuptime>
<fileuid>1</fileuid>
<contents/>
</directory>
</ltfsindex>

so I am guessing i can read the tape.

I also ran ltfsindextool -d /dev/sg2 which produced two .dmp files, but i don't really know what to do with them.

[p4xel]

Could this error be related? I have an IBM drive ULTRIUM-HH9 and LSI controller

LinearTapeFileSystem/ltfs#382

Specifically LinearTapeFileSystem/ltfs#382 (comment)

but I cal see that ULTRIUM-HH9 should already be supported
ibm_tape.c

[piste-jp]

I'm not sure because no log is provided from you about #10 (comment).

But, from the first log you provided shows length mismatch against a READ command. It means this HBA cannot return correct residual length. The LTFS issued an read command with 64KiB, actual data returned from the drive is 4iK, the drive reported 60KiB underun but HBA never provide any underun length. So LTFS reported an error.

LTFS30252I Logical block protection is disabled.
LTFS30216W Length mismatch is detected. (Act = 4096, resid = 0, resid_sense = -61440).
LTFS12049E Cannot read: backend call failed (-21716).
LTFS11174E Cannot read ANSI label: read failed (-21716).

Here, there is another problem. Now LTFS tries to read an ANSI label at the beginning of teh tape. But it's length shall be 80 bytes not 4KiB. So it might be a problem to stop the recovery.

Could this error be related? I have an IBM drive ULTRIUM-HH9 and LSI controller

LinearTapeFileSystem/ltfs#382

Specifically LinearTapeFileSystem/ltfs#382 (comment)

I think the same thing in the comment highly likely happens on your machine. I guess you are using SAS HBA only for HDD/SSD. 4KiB is one of the typical sector size of HDD.

[p4xel]

It seemed to be that the controller could be an issue so I also tried on the following SAS controllers:

02:00.0 Serial Attached SCSI controller: LSI Logic / Symbios Logic SAS3216 PCI-Express Fusion-MPT SAS-3 (rev 01)
03:00.0 SCSI storage controller: Areca Technology Corp. ARC-1300ix-16 16-Port PCI-Express to SAS Non-RAID Host Adapter (rev 02)

The drive wouldn't even be detected on LSI SAS3216, and on Areca it behaves the same as on LSI SAS2116
Would you have any info what SAS controller should be working correctly?
I don't have any others to try.

I also started just dumping data from this tape using dd and got interesting results.

dd if=/dev/nst0 of=/mnt/usb/fileXXX bs=512K status=progress

and each time it resulted in an index XML followed by data blob. Is there any way I can extract data from those pairs?

Just a side note - original backups were made in Windows, could that be an issue?

[p4xel]

I tried another SAS HBA - Adaptec ASC-1405 - with exactly the same symptoms
I then tried on a good LTO9 tape and got nearly the same errors.

This leads me to believe the problem is with using SAS HBAs regardless of vendor.
Is there anything I can do to fix this issue?

[piste-jp]

I also started just dumping data from this tape using dd and got interesting results.

dd if=/dev/nst0 of=/mnt/usb/fileXXX bs=512K status=progress

and each time it resulted in an index XML followed by data blob. Is there any way I can extract data from those pairs?

That is what I'm saying, your HBA returns wrong size information to device driver and application. So you need to use a HBA which support tape drive, the SSC command set in SCSI spec, correctly.

You can find a HBA from the link below. Because you are using IBM LTO9 HH.
https://www.ibm.com/systems/support/storage/ssic/interoperability

Just a side note - original backups were made in Windows, could that be an issue?

No. Not related at all.

This leads me to believe the problem is with using SAS HBAs regardless of vendor.
Is there anything I can do to fix this issue?

Use the HBA which IBM lists. Please contact to IBM if you still have the same problem even if you are using a IBM listed HBA.

[p4xel]

Hi. I have checked the interoperability form and found that LSI (Broadcom) 9300 chipset is listed, but just to rule this out I have ordered an ATTO H1280GT. After installing it I am again getting the same output from ltfs driver:

root@ubuntu1804:~# ltfsck /dev/st0
LTFS16000I Starting ltfsck, LTFS version 2.5.0.0 (Prelim), log level 2.
LTFS16088I Launched by "ltfsck /dev/st0".
LTFS16089I This binary is built for Linux (x86_64).
LTFS16090I GCC version is 7.5.0.
LTFS17087I Kernel version: Linux version 5.4.0-150-generic (buildd@bos03-amd64-012) (gcc version 7.5.0 (Ubuntu 7.5.0-3ubuntu1~18.04)) #167~18.04.1-Ubuntu SMP Wed May 24 00:51:42 UTC 2023 i386.
LTFS17089I Distribution: NAME="Ubuntu".
LTFS17089I Distribution: DISTRIB_ID=Ubuntu.
LTFS17085I Plugin: Loading "sg" tape backend.
LTFS30209I Opening a device through sg-ibmtape driver (/dev/st0).
LTFS30250I Opened the SCSI tape device 0.0.0.-1 (/dev/st0).
LTFS30207I Vendor ID is IBM     .
LTFS30208I Product ID is ULTRIUM-HH9     .
LTFS30214I Firmware revision is PA61.
LTFS30215I Drive serial is 90WT800001.
LTFS30285I The reserved buffer size of /dev/st0 is 1048576.
LTFS30294I Setting up timeout values from RSOC.
LTFS17160I Maximum device block size is 1048576.
LTFS11330I Loading cartridge.
LTFS30252I Logical block protection is disabled.
LTFS11332I Load successful.
LTFS17157I Changing the drive setting to write-anywhere mode.
LTFS16014I Checking LTFS file system on '/dev/st0'.
LTFS30252I Logical block protection is disabled.
LTFS30216W Length mismatch is detected. (Act = 4096, resid = 0, resid_sense = -61440).
LTFS12049E Cannot read: backend call failed (-21716).
LTFS11174E Cannot read ANSI label: read failed (-21716).
LTFS11170E Failed to read label (-21716) from partition 0.
LTFS11009E Cannot read volume: failed to read partition labels.
LTFS16080E Cannot check volume (8).
LTFS30252I Logical block protection is disabled.

It seems there is a problem with the way the ltfs driver is handling SAS connected drives.
Can this issue be moved to ltfs repository?

[piste-jp]

Hmm, it is really hard to respond because multiple things are not correct...

1. I don't understand why you are specifying the device as /dev/st0. Need to specify a sg device like /dev/sg4
2. I checked the SSIC but ATTO H1280GT is not listed into the Ubuntu 18.X LTS
3. I chekced the SSCI but LTO9 is not supported on Ubuntu 18.X LTS at all

[piste-jp]

Can this issue be moved to ltfs repository?

I don't know, you can do it if you want to do. But I recommend you open another one to point what is the problem. (Because title of this issue is build breakage.)

Another aspect is this kind of HBA interoperability related issue is not a scope of the LTFS project. Because it just use the sg driver through a sg device. In this case, the sg driver (and lower component) cannot return a correct response to LTFS. I believe LTFS doesn't have any solution about that.

[manu0401]

On Mon, Jun 10, 2024 at 08:29:35PM -0700, Atsushi Abe wrote: 1. I don't understand why you are specifying the device as `/dev/st0`. Need to specify a sg device like `/dev/sg4`

Not sure about the discussed setup, but /dev/st0 is not that odd. This is what I use on NetBSD. -- Emmanuel Dreyfus ***@***.***