RaSCSI version 22.08.01

RaSCSI version 22.10.01

PiSCSI version 22.12.1

PiSCSI Version 23.02.01

…#1135)

No need to configure the cache anymore, SonarCloud now has an automatic analysis caching. See https://docs.sonarcloud.io/advanced-setup/languages/c-c-objective-c/#analysis-cache.

…PiSCSI#1135)" (PiSCSI#1144)

This reverts commit 3ad668c.

PiSCSI version 23.04.01

PiSCSI version 23.11.01

tl;dr

Treat a computed length of 0 as `has_valid_page_code`.

Details:

The SRM console (aka 'BIOS') of DEC Alpha sends an empty
ModeSelect6 with the following data:
~~~
ModeSelect6, CDB $151000000c00
~~~

That makes 12 byte(s) as follows
~~~
  0  1  2  3   4   5  6  7   8   9 10 11
 00 00 00 08  00  00 00 00  00  00 02 00
~~~

decoding it (accoring to [1], Section 8.3.3, Table 94) gives us

Mode Data Length 0
Medium Type      0
Device-specific  0
Block desc len   8

Density Code     0
Number of blks   0
Reserved         0
Block length     512

`scsi_command_util::ModeSelect` computes
~~~
offset = 4 + buf[3];
~~~

giving 12 and

~~~
length -= offset;
~~~

giving 0.

Thus it never enters the `while` loop and `has_valid_page_code` stays
`false`, raising an error.

[1] [Small Computer System Interface - 2 rev 10L.pdf](https://dn790004.ca.archive.org/0/items/SCSISpecificationDocumentsSCSIDocuments/Small%20Computer%20System%20Interface%20-%202%20rev%2010L.pdf)

Signed-off-by: Klaus Kämpf <[email protected]>

OpenVMS Alpha (the operating system, not the SRM BIOS) uses
ModeSelect6 with a page code of 1.

The semantics are unknown, just accepting it works for me.

Signed-off-by: Klaus Kämpf <[email protected]>

tl;dr

The 'skip to next ModeSelect page' computation was off-by-one, either not
taking the page code itself into account or missing the fact that the
page length is given as `n - 1`.

Fix:

Add 1 to the computed length.

Details:

OpenVMS Alpha sends a ModeSelect6 as follows
~~~
command:

ModeSelect6, CDB $151000001900

payload:

 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
00 00 00 08 00 00 00 00 00 00 02 00 01 0a 24 00 00 00 00 00 00 00 00 00 00
~~~

This translates to (accoring to [1], Section 8.3.3)

~~~
Mode Data Length 0
Medium Type      0
Device-specific  0
Block desc len   8
~~~

with the following offset / length computation _before_ the `while` loop
~~~
offset = 12
length = 13
~~~

The first payload section is
~~~
 4  5  6  7  8  9 10 11
00 00 00 00 00 00 02 00
~~~

translating to

~~~
Density Code     0
Number of blks   0
Reserved         0
Block length   0x200 512
~~~

Then follows a pagecode 1 as
~~~
12 13 14 15 16 17 18 19 20 21 22 23 24
01 0a 24 00 00 00 00 00 00 00 00 00 00
~~~

translating to
~~~~
Page code        1
Page length -1  10
Mode parameters 24 00 00 00 00 00 00 00 00 00 00
~~~

computing (inside the `while` loop, as `// Advance to the next page`)

~~~
size =  10 + 2 = 12
~~~

followed by new `offset` and `length` values

~~~
offset = 25
length = 1
~~~

So it stays in the `while` loop (and has a larger-than-buffer `offset`
value)

Signed-off-by: Klaus Kämpf <[email protected]>

Signed-off-by: Klaus Kämpf <[email protected]>

Signed-off-by: Klaus Kämpf <[email protected]>

Signed-off-by: Klaus Kämpf <[email protected]>

Signed-off-by: Klaus Kämpf <[email protected]>