Revert "tests/qtest: use qos_printf instead of g_test_message"

Revert "tests/qtest: use qos_printf instead of g_test_message"

This reverts commit 30ea13e9d97dcbd4ea541ddf9e8857fa1d5cb30f.

Also rewrites qos_printf() calls added later.

"make check" prints many lines like

    stdout: 138: UNKNOWN:     # # qos_test running single test in subprocess
    stdout: 139: UNKNOWN:     # # set_protocol_features: 0x42
    stdout: 140: UNKNOWN:     # # set_owner: start of session
    stdout: 141: UNKNOWN:     # # vhost-user: un-handled message: 14
    stdout: 142: UNKNOWN:     # # vhost-user: un-handled message: 14
    stdout: 143: UNKNOWN:     # # set_vring(0)=enabled
    stdout: 144: UNKNOWN:     # # set_vring(1)=enabled
    stdout: 145: UNKNOWN:     # # set_vring(0)=enabled
    stdout: 146: UNKNOWN:     # # set_vring(1)=enabled
    stdout: 147: UNKNOWN:     # # set_vring(0)=enabled
    stdout: 148: UNKNOWN:     # # set_vring(1)=enabled
    stdout: 149: UNKNOWN:     # # set_vring(0)=enabled
    stdout: 150: UNKNOWN:     # # set_vring(1)=enabled
    stdout: 151: UNKNOWN:     # # set_vring(0)=enabled
    stdout: 152: UNKNOWN:     # # set_vring(1)=enabled
    stdout: 153: UNKNOWN:     # # set_vring_num: 0/256
    stdout: 154: UNKNOWN:     # # set_vring_addr: 0x7f9060000000/0x7f905ffff000/0x7f9060001000

Turns out this is qos-test, and the culprit is a commit meant to ease
debugging.  Revert it until a better solution is found.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-ID: <20250728145747.3165315-1-armbru@redhat.com>
[Commit message clarified]
(cherry picked from commit c9a1ea9c52e6462ad5c7814f3abd65baa69dc4ce)
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>