[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <82578dd4-a00d-44aa-9059-c5c5215aeb0b@app.fastmail.com>
Date: Fri, 25 Nov 2022 11:06:18 +0100
From: "Arnd Bergmann" <arnd@...db.de>
To: "Lee Jones" <lee@...nel.org>,
"Andrew Morton" <akpm@...ux-foundation.org>,
"Nathan Chancellor" <nathan@...nel.org>,
"Maarten Lankhorst" <maarten.lankhorst@...ux.intel.com>,
mripard@...nel.org, "Thomas Zimmermann" <tzimmermann@...e.de>,
"Dave Airlie" <airlied@...il.com>,
"Daniel Vetter" <daniel@...ll.ch>,
"Nick Desaulniers" <ndesaulniers@...gle.com>,
"Tom Rix" <trix@...hat.com>,
"Harry Wentland" <harry.wentland@....com>,
"Leo Li" <sunpeng.li@....com>,
"Rodrigo Siqueira" <Rodrigo.Siqueira@....com>,
"Alex Deucher" <alexander.deucher@....com>,
Christian König <christian.koenig@....com>,
Xinhui.Pan@....com
Cc: linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
llvm@...ts.linux.dev, amd-gfx@...ts.freedesktop.org,
dri-devel@...ts.freedesktop.org
Subject: Re: [PATCH 2/3] drm/amdgpu: Temporarily disable broken Clang builds due to
blown stack-frame
On Fri, Nov 25, 2022, at 10:25, Lee Jones wrote:
> calculate_bandwidth() is presently broken on all !(X86_64 || SPARC64 || ARM64)
> architectures built with Clang (all released versions), whereby the stack
> frame gets blown up to well over 5k. This would cause an immediate kernel
> panic on most architectures. We'll revert this when the following bug report
> has been resolved: https://github.com/llvm/llvm-project/issues/41896.
>
> Suggested-by: Arnd Bergmann <arnd@...db.de>
> Signed-off-by: Lee Jones <lee@...nel.org>
> ---
> drivers/gpu/drm/Kconfig | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> index 34f5a092c99e7..1fa7b9760adb8 100644
> --- a/drivers/gpu/drm/Kconfig
> +++ b/drivers/gpu/drm/Kconfig
> @@ -265,6 +265,7 @@ source "drivers/gpu/drm/radeon/Kconfig"
>
> config DRM_AMDGPU
> tristate "AMD GPU"
> + depends on BROKEN || !CC_IS_CLANG || !(X86_64 || SPARC64 || ARM64)
The logic looks fine, this has been broken for so long without anyone
paying attention that limiting the broken code to the working architectures
is probably the best way to avoid trouble.
However, as far as I can tell, the problem doesn't affect the
entire driver, only the "new" display engine, so I would probably
try to limit this to turning off CONFIG_DRM_AMD_DC for architectures
that don't have a fixed clang.
Arnd
Powered by blists - more mailing lists