On Fri, Feb 22, 2002 at 08:38:10PM -0800, Mark McClelland wrote: > Greg KH wrote: > > >What this means is you didn't send me all of the changesets from when > >you cloned my tree to the one that you sent me. > > > >Can you either send me all of the changesets, or just send me a patch? > > > > Here's a treediff between your tree and mine (which also doubles as a > patch). As you can see, my tree is OK, but its metadata must be corrupt.
No, it means that you didn't send me all of the changesets that you have added to my tree since you last pulled it. > Either that, or you accidentally tried to receive it on your 2.4 tree ;) Heh, no I would have gotten a different error then :) Two comments on the patch: > --- usb-2.5/drivers/usb/ov511_decomp.c Wed Dec 31 16:00:00 1969 > +++ usb-2.5-bad/drivers/usb/ov511_decomp.c Thu Feb 21 21:21:35 2002 > @@ -0,0 +1,1260 @@ > +/* OV511 Decompression Support Module (No-MMX version) > + * > + * Copyright (c) 1999-2002 Mark W. McClelland. All rights reserved. > + * Original decompression code Copyright 1998-2000 OmniVision Technologies > + * > + * Please see the file: linux/Documentation/usb/ov511.txt > + * and the web site at: http://alpha.dyndns.org/ov511 > + * for more info. > + * > + * This program is free software; you can redistribute it and/or modify it > + * under the terms of the GNU General Public License as published by the > + * Free Software Foundation; version 2 of the License. You may also, at your > + * option, follow the terms under the heading "BSD LICENSE", below. I don't see such a heading below :) And if this is a dual licensed module, please label it as such in the MODULE_LICENSE() macro. > + > +static void > +DecompLock(void) > +{ > + MOD_INC_USE_COUNT; > +} > + > +static void > +DecompUnlock(void) > +{ > + MOD_DEC_USE_COUNT; > +} > + > +static struct ov51x_decomp_ops decomp_ops = { > + decomp_400: Decompress400, > + decomp_420: Decompress420, > + decomp_lock: DecompLock, > + decomp_unlock: DecompUnlock > +}; I'd recommend removing the decomp_lock and decomp_unlock fields from your interface and just using a simple, owner: THIS_MODULE, type field, like the file_ops and usb core uses. Then when you would normally call decomp_lock, you would do something like: if (decomp_ops->owner) __MOD_INC_USE_COUNT(decomp_ops->owner); which will probably enable you to get rid of lock_kernel() in request_decompressor() and release_decompressor(). thanks, greg k-h _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
