Hi Mike, On Tue, Nov 11, 2014 at 8:52 AM, Mike Frysinger <vapier@xxxxxxxxxx> wrote: > On 11 Nov 2014 08:14, Michael Kerrisk (man-pages) wrote: >> On 11/09/2014 04:50 PM, Heinrich Schuchardt wrote: >> > The list contains hex values of different constants. I just wonder for >> > which architecture (alpha, i386, mips, or sparc at that time). No >> > information is supplied. >> > >> > Current values depend on the architecture, e.g. >> > >> > On amd64 >> > 0x82307201 VFAT_IOCTL_READDIR_BOTH >> > 0x82307202 VFAT_IOCTL_READDIR_SHORT >> > 0x80047210 FAT_IOCTL_GET_ATTRIBUTES >> > 0x40047211 FAT_IOCTL_SET_ATTRIBUTES >> > 0x80047213 FAT_IOCTL_GET_VOLUME_ID >> > >> > On mips >> > 0x42187201 VFAT_IOCTL_READDIR_BOTH >> > 0x42187202 VFAT_IOCTL_READDIR_SHORT >> > 0x40047210 FAT_IOCTL_GET_ATTRIBUTES >> > 0x80047211 FAT_IOCTL_SET_ATTRIBUTES >> > 0x40047213 FAT_IOCTL_GET_VOLUME_ID >> > >> > Hence hex values should be removed. >> >> It sounds like you are right that the hex values should be >> removed. But, how did you determine those different >> hex values above? Grepping the sources, it's not >> obvious that amd64 and mips should be different. > > ioctl's can integrate the type size right ? so wouldn't be surprised if it > changed across arches. Yes. I'd actually noticed that detail,and should have said so. I suppose what I really meant was: is that the only source of the difference in the constants? > fwiw, strace carries a script to try and do all this magic automatically: > http://sourceforge.net/p/strace/code/ci/master/tree/linux/ioctlent.sh > of course it misses out on ioctls that aren't exported in headers and are > internal to specific drivers. i.e. it's not perfect at all, but that's because > ioctls are a goddamn mess :). Thanks for that pointer. > i agree though that the man page should drop hex values as they aren't useful. > if you're using a "standard" ioctl, you should get it from the relevant header > file. if you're using a "non-standard" ioctl, then you get the pieces. Agreed. >> > I further suggest to remove all documentation of structure details. >> >> Could you elaborate this point a little. Some examples, and >> why you think they should be removed. I'm not disagreeing, just >> looking to clarify what you mean. > > i think for the common/standard ioctls, it's useful to have manpages that delve > down into the details. but for the fringe ones, it's a waste of time for the > vast majority of people. > > maybe one way to determine whether it's worth documenting, see if strace has a > decoder for it :). Sounds like a reasonable idea to me. Cheers, Michael -- Michael Kerrisk Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/ Linux/UNIX System Programming Training: http://man7.org/training/ -- To unsubscribe from this list: send the line "unsubscribe linux-man" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
