esp: ensure in-flight SCSI requests are always cancelled
commitde7e2cb15586a61d26fc3983ba2cbcbc5c234e15
authorMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Mon, 1 Nov 2021 18:35:15 +0000 (1 18:35 +0000)
committerPaolo Bonzini <pbonzini@redhat.com>
Tue, 2 Nov 2021 14:57:27 +0000 (2 15:57 +0100)
tree6f192f6b4cd7936ab89612747715804a303ad771
parentcabf9862e42ffeab9de9b1bfa12cddaf125c53e8
esp: ensure in-flight SCSI requests are always cancelled

There is currently a check in esp_select() to cancel any in-flight SCSI requests
to ensure that issuing multiple select commands without continuing through the
rest of the ESP state machine ignores all but the last SCSI request. This is
also enforced through the addition of assert()s in esp_transfer_data() and
scsi_read_data().

The get_cmd() function does not call esp_select() when TC == 0 which means it is
possible for a fuzzer to trigger these assert()s by sending a select command when
TC == 0 immediately after a valid SCSI CDB has been submitted.

Since esp_select() is only called from get_cmd(), hoist the check to cancel
in-flight SCSI requests from esp_select() into get_cmd() to ensure it is always
called when executing a select command to initiate a new SCSI request.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Closes: https://gitlab.com/qemu-project/qemu/-/issues/662
Closes: https://gitlab.com/qemu-project/qemu/-/issues/663
Message-Id: <20211101183516.8455-2-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
hw/scsi/esp.c