Mail Archives: djgpp-workers/2000/12/01/01:02:09
At 10:08 PM 11/30/00 -0500, Mark E. wrote:
>> At 09:27 AM 11/30/00 -0500, Mark E. wrote:
>> >> Does the IOCTL subfunction or poking the SFT allow to set the
>> >> SH_DENY* bits for files that are already open?
>>
>> >You can fiddle with those flags in the SFT, but I haven't
>> experimented >with them.
>>
>> OK, you folk are beyond my level of knowledge here. Are you saying
>> there is an alternate way to set up read locks? Assuming that these
>> bits can be set on an open file, how would it work?
>
>A file handle's SH_DENY* flags are stored in the JFT like a lot of other
>flags passed to open(). If you manipulate them using undocumented
>methods, you just might be able to have a read lock for the whole file
>that can then be unlocked by fiddling with the handle's open flagst.
>However, I haven't tried this and I doubt it's worth the effort to work
>on this just to get one special case of read locks working.
OK, then we stay with the current F_RDLCK workaround, and allow whole-file
read locks as previously discussed?
BTW, I was looking through RBIL, and I didn't see anything that suggested
SH_DENY* equivalents in the SFT, but maybe I'm not reading the right
section.
And I also don't see anything in a JFT except the SFT number. Am I missing something there? It's only one byte per file in the JFT, right?
---------------------------------------------------------
Peter J. Farley III (pjfarley AT dorsai DOT org OR
pjfarley AT banet DOT net)
- Raw text -