On 11/26/2012 05:57 PM, Alan Stern wrote:
On Mon, 26 Nov 2012, Sebastian Andrzej Siewior wrote:
Step two is a folder where we setup individual configuration of a function.
/functions/ttyS0
/functions/ttyS1
/functions/storage
each one of the function folder has the file called "name". So for the
folder ttyS0 the file name contains "acm" because we use the ACM
function. The same is true for ttyS1 and for storage we write
"mass_storage" if that is what function will be called.
Alternative (my favorite): function_common-name
that would be
/functions/acm_ttyS0
/functions/acm_ttyS1
/functions/mass_storage_storage
So we avoid the _name_ file and it will be parsed from the directory
name. That is what target/tcm does. You have
rd_mcp_X for a ramdisk and fileio_X for a file as backend.
After name is written it will request the module and special
configuration related files pop up. For ACM it will be almost nothing
while for storage we get nofua or something like that and we will be
able to create the lunX folder.
Is there really any significant configuration to be done at this level?
I would expect functions to be written so that different settings can
be applied to different instances.
For example, if a config had two mass-storage interfaces, they should
be configurable separately. There should be very little, or nothing,
that would apply across all the instances.
I don't follow. acm_ttyS0 and acm_ttyS1 are different instances and can
have different settings. The first one comes up with GS0 the latter
with GS1 as its node name.
Same would apply to mass storage.
Is it possible that I missed your point?
Alan Stern
Sebastian
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
