On Thu, 2011-11-24 at 14:11 +0100, Sebastian Andrzej Siewior wrote: > This is probably a leftover from a rework. According to the wiki > | echo iblock_major=254,iblock_minor=2 > $TARGET/iblock_0/lvm_test0/control > > should be issued. However the code matches for udev_path and force. > Since we have already udev_path handling in target_core_configfs lets > use it. I remove the major/minor thingy since there is no evidence that > this is used. > > Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> > --- > drivers/target/target_core_iblock.c | 77 +++++------------------------------ > drivers/target/target_core_iblock.h | 3 - > 2 files changed, 10 insertions(+), 70 deletions(-) > Hi again Sebastian, I like this cleanup using se_device->se_dev_udev_path instead of internal struct member usage for iblock, but this will currently break userspace during iblock device setup when returning -EINVAL for existing iblock_set_configfs_dev_params() usage. So I'd like to avoid this for the moment as we still expect all backend devices to use some form of target/core/$HBA/$DEV/control input, and while making a special case for IBLOCK in userspace may be a future option, it's not worth the pain to userspace breakage w/ this patch atm. The other option here would be to return non exception status when using target/core/$HBA/$DEV/control for backend drivers (like IBLOCK) that would no longer provide se_subsystem_api->set_configfs_dev_params() from target_core_store_dev_control() usage. This would still expect to fail during target_core_store_dev_enable() -> check_configfs_dev_params() if udev_path has not been set via se_device->se_dev_udev_path, so returning non exception status here from legacy control attribute input should be enough to make existing userspace happy with iblock backend. --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
