campbell@sauron.UUCP (02/26/87)
Looking through the 3B2 V.3 code, it appears to me that there exists a bug in
the handling of 0410 files (note: I realize that these are not the default).
If a 0410 file were exec'ed that had a text (or data) size that was larger
than physical memory but smaller than the swap space the exec would "hang"
forever. This is due to the fact that the entire appropo section of the file
is read in at once in the kernel via "readi()".
If anyone knows of this problem, or knows that this *isn't* a problem, I'd
appreciate your letting me know. Thanks.
--
Mark Campbell
{}!ncsu!ncrcae!sauron!campbelldf@nud.UUCP (02/27/87)
In article <831@sauron.Columbia.NCR.COM> campbell@sauron.Columbia.NCR.COM (Mark Campbell) writes: >Looking through the 3B2 V.3 code, it appears to me that there exists a bug in >the handling of 0410 files (note: I realize that these are not the default). >If a 0410 file were exec'ed that had a text (or data) size that was larger >than physical memory but smaller than the swap space the exec would "hang" >forever. This is due to the fact that the entire appropo section of the file >is read in at once in the kernel via "readi()". This is NOT a problem if while "faulting in" pages during readi(), you allow previously read in pages to be "paged out" via the vhand or sched processes. I'm pretty sure that's the way it works on the 3B2; it does work that way on System V/68 R3V1, our 68020/68851 port of the 3B2 code. -Dale -- Dale Farnsworth (NO7K) seismo!noao!mcdsun!nud!df (602)438-5739 ihnp4!mot!nud!df