[PATCH v1] aarch64 tests: remove irrelevant information messages from error tests

[Alice Carlotti]

On Thu, Aug 07, 2025 at 10:14:11AM +0100, Matthieu Longo wrote:
> On 2025-08-06 18:09, Alice Carlotti wrote:
> > On Mon, Jul 21, 2025 at 05:19:55PM +0100, Matthieu Longo wrote:
> > > Some existing error tests for system registers produce information level
> > > messages that don't contain any information useful for the test.
> > > 
> > > This patch turns off those information messages using the --no-info option.
> > 
> > If it were a new test file I wouldn't mind it, but I don't think it's worth the
> > churn for changing existing tests.
> > 
> > In any case, I have a large WIP patch that affects sysreg tests, so I'd prefer
> > to avoid any unnecessary merge conflicts with that for now.
> > 
> > I'll also note that in some other cases (though not this one) I have found the
> > line numbers in info messages helpful for identifying the actual assembly line
> > used in a test case.
> > 
> > Alice
> 
> I am not denying that those info lines might have a value, but they have
> none in the case of the system register tests. The info line and the
> warning/error line have exactly the same number, so they are redundant.

They do not have the same line number (and also have the wrong filename; it's
PR32738 and I have a fix for it). However, I do agree that they aren't
particularly useful in this case (because it's easy to match the unique
register names).

> 
> A previous patch (92b97af5ef0301ef983df5b3f52860309f64d9d3) was applied to
> clean up some of them, but new ones were added in the meantime.

I thought that there weren't any existing uses, but it turns out I just got the
arguments to grep the wrong way around.  So existing precedent supports this.

> This patch does not intend to clean up all of them, but targets the tests
> which require being changed for the upcoming GICv5 patch series.

Ah, this is the context I was missing.  In which case this should be fine as
part of that patch series, but please wait for now to avoid unnecessary
conflict with my own in-progress series (assuming that mine ends up being
merged first).

> 
> Matthieu