On Sun, 6 Oct 2013, Hans de Goede wrote: > >> I was wondering about this myself when I wrote this. What do you suggest > >> as an alternative, kmalloc an array with > >> intf->cur_altsetting->desc.bNumEndpoints > >> pointers ? > > > > Or simply do 30 pointers every time. 240 bytes isn't all that much for > > kmalloc. > > Right, the allocating intf->cur_altsetting->desc.bNumEndpoints versus 30 > pointers is just an optimization. I was wondering more if it would be > ok here to do a kmalloc(..., GPF_KERNEL), or if this code path is > ever called under interrupt, or some such ?
GFP_KERNEL will be okay. This code always runs in process context with interrupts enabled. > Also do we allow C-99 style variable length arrays in the kernel ? > Then we could just put what ever we need on the stack (which almost always > will be way less then 240 bytes). I think people try to avoid variable-length arrays. Certainly such arrays in structures are frowned upon. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
