USB MSC: Fix SCIS mode sense write protection bit

[mluessi]

The USB MSC / SCSI layer currently doesn't set the write protect bit in the SCSI ModeSense response. Due to this, the Emulated FAT file system used for the internal flash doesn't get mounted read-only. When mounting the drive on Linux, the host attempts to write to the drive (I assume for the access time) , which fails and then the mounting doesn't succeed.

This fix sets the relevant bit in the SCSI ModeSense responses. The host doesn't attempt to write to the drive and the mounting works as expected.

This is only for H7 targets. The fix also should be added to F7 etc.

Fixes #9160

Before:
[1191094.145018] scsi 3:0:0:0: Direct-Access INAV FC Onboard Flash PQ: 0 ANSI: 2
[1191094.145670] sd 3:0:0:0: [sdc] 527470 512-byte logical blocks: (270 MB/258 MiB)
[1191094.145815] sd 3:0:0:0: [sdc] Write Protect is off
[1191094.145817] sd 3:0:0:0: [sdc] Mode Sense: 00 00 00 00

With this fix:
[19418.047443] scsi 4:0:0:0: Direct-Access INAV FC Onboard Flash PQ: 0 ANSI: 2
[19418.048021] sd 4:0:0:0: [sdd] 527470 512-byte logical blocks: (270 MB/258 MiB)
[19418.048224] sd 4:0:0:0: [sdd] Write Protect is on
[19418.048229] sd 4:0:0:0: [sdd] Mode Sense: 00 00 80 00

[stronnag]

Is this only applied to Flash or generically to MSC?
If it's generic, it's not the correct solution as MSC for SD Card should allow delete (i.e. the device is not write protected).

btw. regardless, on F4, MSC flash devices are mounted (write protected, after write sense fails) without this PR.

[mluessi]

@stronnag it only sets the write protect flag if the IsWriteProtected function indicates it, so it works for both the internal flash (write protected) and SD Card (not write protected).

[stronnag]

Thanks for the clarification. This also needs to be applied to F4 / F7 in that case.

[stronnag]

@mluessi , can you please rebase this to release_7.1.0 as it's a useful bug fix. You might also make the same fix against F4 and F7. I tested F4 locally and it works perfectly.

[mluessi]

@stronnag sounds good, I will add F4 and F7 and change the target branch

[mluessi]

@stronnag done :)

[stronnag]

tested by the author on H7 and by me on F4.