[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v2 1/3] device_tree: Add a helper function for string arrays

From: David Gibson
Subject: Re: [PATCH v2 1/3] device_tree: Add a helper function for string arrays
Date: Sun, 10 Nov 2019 19:44:24 +0000
User-agent: Mutt/1.12.1 (2019-06-15)

On Sat, Nov 09, 2019 at 03:59:24PM +0000, Peter Maydell wrote:
> On Fri, 8 Nov 2019 at 19:48, Palmer Dabbelt <address@hidden> wrote:
> >
> > The device tree format allows for arrays of strings, which are encoded
> > with '\0's inside regular strings.  These are ugly to represent in C, so
> > the helper function represents them as strings with internal '\0's that
> > are terminated with a double '\0'.  In other words, the array
> > ["string1", "string2"] is represeted as "string1\0string2\0".
> >
> > The DTB generated by this function is accepted by DTC and produces an
> > array of strings, but I can't find any explicit line in the DT
> > specification that defines how these are encoded.
> > +/*
> > + * This uses a particularly odd encoding: "strings" is a list of strings 
> > that
> > + * must be terminated by two back-to-back '\0' characters.
> > + */
> > +int qemu_fdt_setprop_strings(void *fdt, const char *node_path,
> > +                             const char *property, const char *strings);
> The clean API for this would be to use varargs so you could write
> qemu_fdt_setprop_stringlist(fdt, node, prop, "arm,armv8-timer",
>                             "arm,armv7-timer");
> and have it do the assembly into the encoding that fdt expects.
> That would require us to do a bit of allocation-and-freeing
> to assemble the string, of course, but then we only do fdt
> creation at startup.

Right, I really don't see the value in this interface.  Using
"foo\0bar" is a little ugly, but not really any uglier than
"foo\0bar\0".  The existing interface would be a drag if you had
dynamically created entries in the list (because getting the size
can't be done with sizeof() then), but I don't think that's actually a
very likely usecase.

> NB: I think that this is a good idea but not-for-4.2 material,
> so if you wanted your sifive board change to go into 4.2 you
> should probably start with the simple approach and leave the
> refactoring for the next release cycle.

I concur.

David Gibson                    | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au  | minimalist, thank you.  NOT _the_ _other_
                                | _way_ _around_!

Attachment: signature.asc
Description: PGP signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]