Re: d_off field in struct dirent and 32-on-64 emulation

* Andy Lutomirski:

quoted

On Dec 27, 2018, at 10:18 AM, Florian Weimer [off-list ref] wrote:

We have a bit of an interesting problem with respect to the d_off
field in struct dirent.

When running a 64-bit kernel on certain file systems, notably ext4,
this field uses the full 63 bits even for small directories (strace -v
output, wrapped here for readability):

getdents(3, [
 {d_ino=1494304, d_off=3901177228673045825, d_reclen=40,
d_name="authorized_keys", d_type=DT_REG},
 {d_ino=1494277, d_off=7491915799041650922, d_reclen=24, d_name=".",
d_type=DT_DIR},
 {d_ino=1314655, d_off=9223372036854775807, d_reclen=24,
d_name="..", d_type=DT_DIR}
], 32768) = 88

When running in 32-bit compat mode, this value is somehow truncated to
31 bits, for both the getdents and the getdents64 (!) system call (at
least on i386).

I imagine you’re encountering this bug:

https://lkml.org/lkml/2018/10/18/859

It's definitely in this area.  However, the original collision problem
with 32-bit hashes is also real, so I can see the desire to use more
bits.

Presumably the right fix involves modifying the relevant VFS file
operations to indicate the relevant ABI to the implementations.

Not sure.  How does NFS solve this problem when access happens from a
32-bit process and the rest (client kernel, transport, server kernel)
is 64-bit all the way?

I would guess that 9p is triggering the “not really in the syscall you
think you’re in” issue.

I think the issue is more like the networking case for 9p.  In this
scenario, the server shouldn't have to care whether the client process
is in 32-bit mode or 64-bit mode.  But maybe the only solution is to
pass through some sort of flag, as Peter Maydell has just suggested.