[MCMAN/MCSERV]: `mcGetInfo` always returns -2 when accessing an unknown format card

[israpps]

I have found this issue while making a program to format PS2 memory cards as soulcalibur2 conquest cards

based on the reading i'm doing here of the comments, mcSync should return -2 when detecting a new unformatted memory card. then 0 when calling the function on a card it was called before.

To rephrase:

What happens:

• plug unknown format card
• call mcGetInfo
• mcSync returns -2
• call mcGetInfo
• mcSync returns -2

What I expected to happen based on the comments

• plug unknown format card
• call mcGetInfo
• mcSync returns -2
• call mcGetInfo
• mcSync returns 0

ps2sdk/ee/rpc/memorycard/include/libmc.h

Lines 210 to 223 in 566ed82

	/** get memcard state
	* mcSync result: 0 = same card as last getInfo call
	* -1 = formatted card inserted since last getInfo call
	* -2 = unformatted card inserted since last getInfo call
	* < -2 = memcard access error (could be due to accessing psx memcard)
	*
	* @param port port number
	* @param slot slot number
	* @param type pointer to get memcard type
	* @param free pointer to get number of free clusters
	* @param format pointer to get whether or not the card is formatted (Note: Originally, sceMcGetInfo didn't have a 5th argument for returning the format status. As this is emulated based on the return value of sceMcSync() when rom0:MCSERV is used, please keep track of the return value from sceMcSync instead!)
	* @return 0 = successful; < 0 = error
	*/
	extern int mcGetInfo(int port, int slot, int* type, int* free, int* format);

I will check the behaviour of rom0:MCMAN and report back if it's either an issue on the comment or a bug on MCSERV/MCMAN

[israpps]

tested with the SDK libmc app and XMCMAN module from 1.70romver SCPH-50001 console

it behaves as comments say, the issue is on homebrew MCMAN/MCSERV

[rickgaiser]

Nice investigation, can you fix it and make a PR?

[israpps]

I have found the problem and implemented a solution... However.... I'm not sure if this will have an impact on any other MCMAN behaviors

Basically, I checked how mcdetecard2 is doing it's job, compared XMCMAN from romver 1.70 with our mcman and with DONGLEMAN. found a diff.

maybe move the discussion to the PR that's being opened now

[sp193]

Let me guess... it's the correct behaviour for whatever version of XMCMAN our mcman was cloned from? I recall that the X-modules aren't actually the same across all PS2s. They're technically the modules for the OSD, which really depend on the SDK version which Sony used at the time for the ROM. It's different from the basic MCMAN, which was from something like release 1.34 and they kept it the same in all ROMs.

[israpps]

Let me guess... it's the correct behaviour for whatever version of XMCMAN our mcman was cloned from? I recall that the X-modules aren't actually the same across all PS2s. They're technically the modules for the OSD, which really depend on the SDK version which Sony used at the time for the ROM. It's different from the basic MCMAN, which was from something like release 1.34 and they kept it the same in all ROMs.

I checked several MCMANs. Not sure where our MCMAN was based from
Edit:

ps2sdk/iop/memorycard/mcman/src/main.c

Lines 15 to 19 in b2bfd09

	// TODO: last sdk 3.1.0 has MCMAN module version 0x2,0x30 with significant changes.
	// Check what was changed, and maybe port changes.
	// CRC32: EC5F33F9
	// Note: currently is based on the last XMCMAN from BOOTROM:
	// 0x02,0x0a (looks like mistake, cause last XMCMAN is 0x02,0x09)