On Tue, Jul 14, 2009 at 20:46, Marcel Holtmann<marcel@xxxxxxxxxxxx> wrote:
>> > What the action you want to take at removal? You can set
>> > ENV{REMOVE_CMD}= and it will be passed to RUN+= for that exact device
>> > when it goes away.
>> >
>> > You can also just store any other custom property in the udev
>> > database, and it will be there on remove.
>> >
>> Thanks for the recommendation on ENV{REMOVE_CMD}. With some minor
>> modifications to my old patch, I've got it working using ENV{REMOVE_CMD}
>> now reliably post suspend.
>>
>> Although rewalking the child devices to find the right one to prod might
>> seem tedious, it should be more future-proof in case later hardware ends
>> up changing.
>>
>> Would you mind reviewing the attached patch for submission to udev then?
>
> + for (bus = usb_get_busses(); bus; bus = bus->next)
> + for (dev = bus->devices; dev; dev = dev->next)
> + if (dev->descriptor.idVendor == base_vendor &&
> + dev->descriptor.idProduct == base_product)
> + for (i = 0; i < dev->num_children; i++)
> + for (j = 0; j < dev->children[i]->descriptor.bNumConfigurations; j++)
> + for (k = 0; k < dev->children[i]->config[j].bNumInterfaces; k++)
> + for (l = 0; l < dev->children[i]->config[j].interface[k].num_altsetting; l++)
> + if (dev->children[i]->config[j].interface[k].altsetting[l].bInterfaceProtocol == bInterfaceProtocol)
> + return dev->children[i];
> + return NULL;
> +}
>
> welcome to the land of fully unreadable code. You need to fix this and
> split it up if needed. Nobody has ever a chance to understand what this
> code does if it goes wrong.
That looks crazy. What are you trying to solve here?
We also don't want these weird multiline rules with on match per line,
and the use of properties merged by earlier rules. I complained too
many times about this now.
That whole thing get pretty annoying, it does not seem to be ready to
be shipped with udev. Please properly fix _and_ test the stuff now.
Your earlier version already broke random keyboards and hubs, because
you obviously did not test it properly. If you don't provide a sane
solution now, I'm going to remove it from udev, and you need to create
your own package.
Thanks,
Kay
--
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
