Re: [PATCH RFCv2 2/3] lib/vsprintf.c: make %pD print full path for file

[Matthew Wilcox <willy@infradead.org>]

On Fri, May 28, 2021 at 02:22:01PM +0000, Justin He wrote:
> > On Fri, May 28, 2021 at 07:39:50PM +0800, Jia He wrote:
> > > We have '%pD' for printing a filename. It may not be perfect (by
> > > default it only prints one component.)
> > >
> > > As suggested by Linus at [1]:
> > > A dentry has a parent, but at the same time, a dentry really does
> > > inherently have "one name" (and given just the dentry pointers, you
> > > can't show mount-related parenthood, so in many ways the "show just
> > > one name" makes sense for "%pd" in ways it doesn't necessarily for
> > > "%pD"). But while a dentry arguably has that "one primary component",
> > > a _file_ is certainly not exclusively about that last component.
> > >
> > > Hence "file_dentry_name()" simply shouldn't use "dentry_name()" at all.
> > > Despite that shared code origin, and despite that similar letter
> > > choice (lower-vs-upper case), a dentry and a file really are very
> > > different from a name standpoint.
> > >
> > > Here stack space is preferred for file_d_path_name() because it is
> > > much safer. The stack size 256 is a compromise between stack overflow
> > > and too short full path.
> >
> > How is it "safer"?  You already have a buffer passed from the caller.
> > Are you saying that d_path_fast() might overrun a really small buffer
> > but won't overrun a 256 byte buffer?
> No, it won't overrun a 256 byte buf. When the full path size is larger than 256, the p->len is < 0 in prepend_name, and this overrun will be
> dectected in extract_string() with "-ENAMETOOLONG".
> 
> Each printk contains 2 vsnprintf. vsnprintf() returns the required size after formatting the string.
> 1. vprintk_store() will invoke 1st vsnprintf() will 8 bytes space to get the reserve_size. In this case, the _buf_ could be less than _end_ by design.
> 2. Then it invokes 2nd printk_sprint()->vscnprintf()->vsnprintf() to really fill the space.

I think you need to explain _that_ in the commit log, not make some
nebulous claim of "safer".

> If we choose the stack space, it can meet above 2 cases.
> 
> If we pass the parameter like:
> p = d_path_fast(path, buf, end - buf);
> We need to handle the complicated logic in prepend_name()
> I have tried this way in local test, the code logic is very complicated
> and not so graceful.
> e.g. I need to firstly go through the loop and get the full path size of
> that file. And then return reserved_size for that 1st vsnprintf

I'm not sure why it's so complicated.  p->len records how many bytes
are needed for the entire path; can't you just return -p->len ?