Comment 5 for bug 2049634

I have done some more tests on a different computer with the same OS, this time with the following Kernel version as reported by "uname -a":

Linux <hostname> 6.5.0-15-generic #15~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Jan 12 18:54:30 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

I generated a reproducible pseudo-random text file in this way:

( set -o pipefail && head --bytes=$(( 55 * 1000 )) /dev/zero | openssl enc -aes-128-ctr -nosalt -pass "pass:my-seed" -iter 1 | hexdump --no-squeezing --format '40/1 "%02x"' --format '"\n"' >"testdata.txt" )

The resulting text file is around 111 kbytes long.

I copied the file with 'cp' to the SMB 1.0 Windows server and back. I have attached both file versions to this bug report.

Comparing the files with 'meld' yields 5 holes with binary zeros, each one exactly 3900 bytes long. This is how I located the holes on the command line:

grep --perl-regexp --text --byte-offset --only-matching '\x00{3900}' testdata-back-from-server.txt | tr --delete '\000'

The resulting hole offsets are:

16580
37060
57540
78020
98500

That is, the holes repeat at a fixed offset interval of 20480 bytes, that is, exactly 20 KiB.

The other PC with Kernel version 6.2.0-39-generic is working fine, as it has in the past years.

It would be nice if someone else could reproduce this, as it might actually be the server side, however unlikely that is.