Re: [PATCH 4.19 00/42] 4.19.4-stable review

[Guenter Roeck]

On 11/22/18 11:45 PM, Greg Kroah-Hartman wrote:
> On Thu, Nov 22, 2018 at 02:30:58PM -0800, Guenter Roeck wrote:
> > On 11/22/18 2:01 PM, Thomas Voegtle wrote:
> > > On Thu, 22 Nov 2018, Thomas Voegtle wrote:
> > >
> > > > Doesn't compile for me on OpenSuSE 15.0 (gcc 7.3.1):
> > > >
> > > >  Â CALLÂÂÂ scripts/checksyscalls.sh
> > > >  Â DESCENDÂ objtool
> > > >  Â CHKÂÂÂÂ include/generated/compile.h
> > > >  Â CC [M]Â drivers/gpu/drm/i915/i915_gem_gtt.o
> > > > drivers/gpu/drm/i915/i915_gem_gtt.c: In function âgen6_dump_ppgttâ:
> > > > drivers/gpu/drm/i915/i915_gem_gtt.c:1771:41: error: format â%llxâ expects
> > > > argument of type âlong long unsigned intâ, but argument 5 has type âlong
> > > > unsigned intâ [-Werror=format=]
> > > >  ÂÂ seq_printf(m, "\t\t(%03d, %04d) %08llx: ",
> > > >  ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ ~~~~~ ^
> > > >  ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ %08lx
> > > > cc1: all warnings being treated as errors
> > > > make[4]: *** [scripts/Makefile.build:306:
> > > > drivers/gpu/drm/i915/i915_gem_gtt.o] Error 1
> > > > make[3]: *** [scripts/Makefile.build:546: drivers/gpu/drm/i915] Error 2
> > > > make[2]: *** [scripts/Makefile.build:546: drivers/gpu/drm] Error 2
> > > > make[1]: *** [scripts/Makefile.build:546: drivers/gpu] Error 2
> > > > make: *** [Makefile:1052: drivers] Error 2
> > > >
> > > >
> > > >
> > > > 4.19.3 is broken for me, too. 4.19.2 is works.
> > > > 4.20-rc tree works.
> > > >
> > > > Config attached.
> > >
> > >
> > > quick bisect shows problem is:
> > >
> > > commit 1a25e1a1be71a49ee7f34fb14b5a26191e6cf501
> > > Author: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> > > Date:ÂÂ Thu Oct 25 10:18:22 2018 +0100
> > >
> > >   ÂÂÂ drm/i915: Mark up GTT sizes as u64
> > >
> > >   ÂÂÂ commit c58281056a8b26d5d9dc15c19859a7880835ef44 upstream.
> > >
> > >
> > > reverted on 4.19.4-rc1 it compiles for me. I guess something is missing here?
> > AFAICS someone was not careful with the backport. Upstream doesn't have %llx
> > at that place.
>
> Yes it does, unless I'm reading the code wrong?

Ah, I was looking at v4.20-rc1, which didn't have it. Wrong branch, sorry.

> It seems that this does not trigger on upstream due to f6e35cda6614
> ("drm/i915: Replace some PAGE_SIZE with I915_GTT_PAGE_SIZE") being
> there, which changed the type of the variable being displayed here.
>  > I've now backported that as well, as it should help out over time, and
> the build warning seems to be gone.
Makes sense.

Thanks,
Guenter