Re: large file support? (fwd)

Martin von Loewis (martin@mira.isdn.cs.tu-berlin.de)
Sun, 17 Aug 1997 21:47:47 +0200


> But what for open64/read64/write64? None of them give the
> user a hint as to what the file pointer offset is. And we are
> still on a 32-bit system, with a 32-bit VM space, so writing
> 8GB in one go is not possible.

open64 is a short-hand for open with the flag O_LARGEFILE.
Here's the summary of the new interfaces:
aio(aio_read64,...): The aio structure contains an offset field
locking(flock64,fcntl): support for locking ranges of a file
pointer(ftell,fseek...): offset is 64 bits
dir access(readdir,ftw): d_ion is 64 bits
info(stat,statfs,...): file size, number of blocks, number of files, ...
mmap: mapping ranges
resource limits(getrlimit): maximum file size per process

A new mount option 'nolargefiles' denies mounting file systems if
the fs driver cannot guarantee that all files are small.

Regards,
Martin