Now that sounds like a real problem. Question is: is that nfs that is
broken or /proc. I can see arguments both ways.
Does nfs do a stat on each file before a read? What if it knows the file
pointer is at EOF, does it refuse to read the file? Surely, if you ask nfs
to read, it has to either go out and stat and then refuse or just try the
read. Sounds like trying the read would be less work.
Another problem with reporting sizes on the textual files is that you can
write to them. Having a size of 1 and allowing you to write arbitrary text
into the file without disturbing that 1 seems strange. Makes a little more
sense with 0.
I suppose the best way of doing it would be to make them pipes instead of
files since their behaviour in many cases is more like a pipe into the
kernel than a file.
-- [======================================================================] [ Kevin Lentin Email: K.Lentin@cs.monash.edu.au ] [ finger kevinl@fangorn.cs.monash.edu.au for PGP public key block. ] [ KeyId: 06808EED FingerPrint: 6024308DE1F84314 811B511DBA6FD596 ] [======================================================================]- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu