Re: [PATCH v3 3/3] 9p: skip over-long directory entry names for legacy 9p2000 too

[Dominique Martinet <asmadeus@codewreck.org>]

Haobin Wu,
Thanks, I picked up all three patches

Christian Schoenebeck wrote on Fri, Sep 18, 2026 at 02:08:00PM +0200:
> On Friday, 18 September 2026 04:48:51 CEST Haobin Wu wrote:
> > The legacy 9p2000 and 9p2000.u readdir path, v9fs_dir_readdir(), parses
> > each entry with p9stat_read() and, unlike the 9p2000.L path, never had
> > a 256-byte name limit: a name of any length up to the wire maximum was
> > handed to dir_emit(). getdents64() therefore returned names longer than
> > NAME_MAX that no later syscall can use, and failed with -EIO for names
> > of PATH_MAX bytes or more.
> > 
> > Apply the same NAME_MAX check as the previous patch does for
> > v9fs_dir_readdir_dotl(), so both protocol variants behave the same.
> > 
> > Note that this is a behaviour change for 9p2000 and 9p2000.u: entries
> > with names longer than NAME_MAX used to be listed and are now skipped.
> 
> That sentence is a bit misleading, isn't it?
> 
> Name length 256..4095:
>   - before: listed, but unusable (ENAMETOOLONG)
>   - now: entry skipped, fetch continues
> 
> Name length >=4096:
>   - before: aborts the entire fetch (similar to 9p2000.L side)
>   - now: entry skipped, fetch continues

Right, I've shamelessly copied that in when applying
https://github.com/martinetd/linux/commit/55d6732b8015705a657d96fc5bd37fbed349624a

> > @@ -126,8 +128,16 @@ static int v9fs_dir_readdir(struct file *file, struct
> > dir_context *ctx) return -EIO;
> >  			}
> > 
> > -			over = !dir_emit(ctx, st.name, strlen(st.name),
> > -					QID2INO(&st.qid), dt_type(&st));
> > +			namelen = strlen(st.name);
> > +			if (namelen > NAME_MAX) {
> 
> ... maybe better if (namelen >= PATH_MAX) here for legacy 9p2000(.u)?

Given things won't work after the name is picked up anyway (from what I
tried at least stat didn't work), I don't think it matters much, but I'm
fine breaking compatibility here.
We can bring the limit back up to PATH_MAX if someone claims it's useful
somewhere.

> > +				p9_debug(P9_DEBUG_ERROR,
> > +					 "skip dentry: name length %zu > NAME_MAX\n",
> > +					 namelen);
> > +				over = false;
> 
> Nit: alternative would be init over = false and dropping explicit assignment 
> inside loop branch here. But that's already personal taste level.

(I actually had the same comment, so made the swap locally)

-- 
Dominique Martinet | Asmadeus