Messages in this thread | | | Date | Wed, 22 Dec 1999 15:25:19 -0500 | From | Matthew Wilcox <> | Subject | Re: Linux 2.2.14pre15 |
| |
On Sun, Dec 19, 1999 at 09:38:34AM -0500, TenThumbs wrote: > Given the signed vs. unsigned issues there, I rebuilt pre 15 with "gcc_2.7.2.3 > -W" and turned up these. > > arch/i386/mm/init.c:256: warning: unsigned value >= 0 is always 1 > dirvers/misc/parport_pc.c:931: warning: unsigned value < 0 is always 0 > drivers/char/esp.c:2564: warning: unsigned value < 0 is always 0 > drivers/char/random.c:1443: warning: unsigned value < 0 is always 0 > drivers/char/random.c:1496: warning: unsigned value < 0 is always 0 > drivers/scsi/hosts.c:788: warning: unsigned value < 0 is always 0 > mm/memory.c:395: warning: unsigned value < 0 is always 0 > mm/slab.c:717: warning: unsigned value < 0 is always 0 > > 111 include/asm-i386/fixmap.h:95: warning: unsigned value >= 0 is always 1 > 300 include/linux/signal.h:104: warning: unsigned value < 0 is always 0 > 300 include/linux/signal.h:107: warning: unsigned value < 0 is always 0 > 300 include/linux/signal.h:110: warning: unsigned value < 0 is always 0 > 300 include/linux/signal.h:136: warning: unsigned value < 0 is always 0 > > I checked the first group and gcc definitely did not do what was intended. An > expert need to look at these. The second group is surprising but what is wrong > depends on exactly how teh include file is used.
In fact, nothing is wrong with the signal.h file, the value which is less than 0 is constant and leads to the eliminatin of the loop which is precisely what we want. I have a patch that would remove this warning and thus make -W a usable option, I just need to tidy it up and submit it.
-- "If it shine or if it rain, Little will I care or know. Days, like drops upon a pane, Slip, and join, and go."
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |