OS_FSControl 55 (xosfscontrol_free_space64) failed

Hello,

Can anybody please explain why xosfscontrol_free_space64() function may failed with error “(92670) Bad command”?

My code:

int64_t free_space = 0, total_space = 0;
int max_obj_size = 0;
os_error *err;
err = xosfscontrol_free_space64(_current_root, 
 (bits*)&free_space, 
 ((int*)&free_space+1), 
 &max_obj_size, 
 (bits *)&total_space, 
 ((int*)&total_space + 1));

I get err with number 92670 and message “Bad command”

Seems to work here (consistent with the Free option on HostFS) from BASIC, so there may be a bug in xosfscontrol_free_space64. Try _kernel_swi

I suspect that’s because it’s not implemented by the filing system. Filing system number &69 is AppleFS according to the list of FS numbers I’ve got here, but you stated that your error was for HostFS::H.$ – which is HostFS. It’s also not usually the name of the HostFS filesystem is it ? It’s usually HostFS::HostFS.$. But HostFS declares itself as filing system number &99 (http://www.home.marutan.net/hg/rpcemu/annotate/7e56f1c02af8/riscos-progs/HostFS/hostfs.s#l41), so that’s weird too.

If I do this on HostFS::H.$ with RPCEmu, I get ‘Disc not found" and an error number of 104916 which is &199d4 – ie the HostFS filesystem. Which implies that my assumption that you’re using RPCEmu is wrong.

So I can’t explain it, but the parameters you describe and the response are all odd.

What is the evaluation order of “(int*)&xxx_space + 1”?
cast (&xxx_space + 1) as (int*), which points you at xxx_space + 64 bits
or
(cast &xxx_space) as (int*)) + 1, which points you at xxx_space + 32 bits

If it is the former, you should use “((int*)&xxx_space) + 1” instead to point at xxx_space + 32 bits.

Or use a union of one uint64_t and two uint32_t. Simples.

I get err with number 92670 and message “Bad command”

Possibly unrelated, but whilst developing Partition Manager I have been plagued with random issues caused by either OS_FSControl 49 or OS_FSControl 55 against 3rd party filing systems.

I can’t make out if the issue I’m seeing is triggered by the filesystem Module, FileCore or CLib. To date I’ve not investigated, just commented out the call under the assumption it’s an issue with the underlying filesystem. I was half expecting someone using Partition Manager to report the issue, but nothing to date which also reaffirmed my suspicion around the underlying filesystem.

Partition Manager link doesn’t work.

Partition Manager link doesn’t work.

Looks like it’s supposed to be https://forums.jaspp.org.uk/forum/viewforum.php?f=29