From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Brownell Date: Tue, 23 Jan 2001 16:14:05 +0000 Subject: Re: Problem with hotplug-2001_01_16 + quick solution Message-Id: List-Id: References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-hotplug@vger.kernel.org Martin -- Thanks for that problem report for 2.2 hotplugging! > My USB mouse (Genius Netmouse Pro) has PRODUCTE8/2/0.0 ^^^ > (DEVICE=/proc/bus/usb/001/002, INTERFACE=3/1/2, TYPE not set) > in kernel 2.2.19pre6, so usb.agent aborts its execution > in usb_convert_vars function. In the kernel 2.4.0 it is OK. Greg, that looks like hotplug in 2.2.* needs to get a bugfix from a few months back -- changing the product version code to be straight hex, not a BCD-ized string. Can you get that fix to Alan? (A case of ripping out ugly to_bcd code ... :-) > With modutils 2.4.1 the file /lib/modules/`uname -r`/modules.usbmap > exists even in the 2.2 kernel, but contains only one line, so usb.distmap > is not used. Looks like modutils is not going to ensure that file only exists on kernels that support MODULE_DEVICE_TABLE sections in dynamically linked modules. I'll check in a fix (this AM) that will test the kernel version. (So Keith, that means you needn't bother tweaking modutils to avoid generating empty "modules.usbmap" files.) For the record, I hate testing software version stamps, because that's known to be a fragile scheme. Feature tests are reliable, but we don't seem to have a way to test "do modules on this system support MODULE_DEVICE_TABLE" -- so "is kernel version 2.4/2.5/..." is the best heuristic I can identify. - Dave _______________________________________________ Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net Linux-hotplug-devel@lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel