Re: [RFC PATCH] checks: Use source position information for check failures

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



On Thu, Jan 11, 2018 at 10:19:06AM -0600, Rob Herring wrote:
> On Wed, Jan 10, 2018 at 5:43 PM, David Gibson
> <david@xxxxxxxxxxxxxxxxxxxxx> wrote:
> > On Wed, Jan 10, 2018 at 10:41:54AM -0600, Rob Herring wrote:
> >> On Tue, Jan 9, 2018 at 11:30 PM, David Gibson
> >> <david@xxxxxxxxxxxxxxxxxxxxx> wrote:
> >> > On Mon, Jan 08, 2018 at 04:49:27PM -0600, Rob Herring wrote:
> >> >> Now that we retain source position information of nodes and properties,
> >> >> make that the preferred file name (and position) to print out in check
> >> >> failures. This will greatly simplify finding and fixing check errors
> >> >> because most errors are in included source .dtsi files and they get
> >> >> duplicated every time the source file is included.
> >> >>
> >> >> For now, only converting a few locations and using a new macro name. I
> >> >> will convert all FAIL occurences once we agree on the new syntax. Also,
> >> >> after this, some checks may need some rework to have more specific
> >> >> line numbers of properties rather than nodes.
> 
> [...]
> 
> >> >> @@ -1049,7 +1065,7 @@ static void check_avoid_unnecessary_addr_size(struct check *c, struct dt_info *d
> >> >>       }
> >> >>
> >> >>       if (!has_reg)
> >> >> -             FAIL(c, dti, "unnecessary #address-cells/#size-cells without \"ranges\" or child \"reg\" property in %s",
> >> >> +             FAIL_POS(c, dti, node->srcpos, "unnecessary #address-cells/#size-cells without \"ranges\" or child \"reg\" property in %s",
> >> >>                    node->fullpath);
> >> >
> >> > Checks are already associated with a node, would it make more sense to
> >> > print the position information from the general code?
> >>
> >> Not sure I follow the question. You mean pass in the struct node and
> >> get the srcpos and fullpath inside check_msg?
> >
> > Basically, yes.
> 
> That would help getting the messages to be a more consistent form
> like: 'source/file.dts:123: (ERROR|WARNING): /full/path/of/node: bad
> news'
> 
> That had been something I wanted to do. The downside is it makes for
> really long lines, but many messages already have the full path in
> them. I guess we could go to 2 line messages where the 1st line is the
> error and the 2nd line is the node path. The downside to that is I
> typically do 'sort -u' (stripping the the dtb name) to de-duplicate
> the errors as 10 boards including 1 SoC dtsi file gives me 10 of the
> same error. Of course, printing the dts filename instead fixes that
> problem.

I suspect it might also make a bunch of existing testcases a bit of a
pain in the arse, too.

> It probably makes sense to do that in one step rather than reword
> error messages twice.
> 
> >> While checks are associated with nodes, specific error messages may be
> >> associated with properties. This is one example where we could make
> >> the error message be the exact line (of the #address-cells or
> >> #size-cells), but that would require re-working the check a bit to get
> >> the property structs (and srcpos).

-- 
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_!
http://www.ozlabs.org/~dgibson

Attachment: signature.asc
Description: PGP signature


[Index of Archives]     [Device Tree]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]

  Powered by Linux