linux-kernel - Re: [RFC PATCH 1/4] drm/msm: register a fault handler for display mmu faults

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Tue, 21 May 2024 00:12:39 -0400
From: Stephen Boyd <swboyd@...omium.org>
To: Abhinav Kumar <quic_abhinavk@...cinc.com>, Daniel Vetter <daniel@...ll.ch>, 
	David Airlie <airlied@...il.com>, Dmitry Baryshkov <dmitry.baryshkov@...aro.org>, 
	Marijn Suijten <marijn.suijten@...ainline.org>, Rob Clark <robdclark@...il.com>, 
	Sean Paul <sean@...rly.run>, freedreno@...ts.freedesktop.org
Cc: dri-devel@...ts.freedesktop.org, seanpaul@...omium.org, 
	dianders@...omium.org, quic_jesszhan@...cinc.com, 
	linux-arm-msm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH 1/4] drm/msm: register a fault handler for display mmu faults

Quoting Abhinav Kumar (2024-05-17 16:37:56)
> diff --git a/drivers/gpu/drm/msm/msm_kms.c b/drivers/gpu/drm/msm/msm_kms.c
> index af6a6fcb1173..62c8e6163e81 100644
> --- a/drivers/gpu/drm/msm/msm_kms.c
> +++ b/drivers/gpu/drm/msm/msm_kms.c
> @@ -200,6 +200,28 @@ struct msm_gem_address_space *msm_kms_init_aspace(struct drm_device *dev)
>         return aspace;
>  }
>
> +static int msm_kms_fault_handler(void *arg, unsigned long iova, int flags, void *data)
> +{
> +       struct msm_kms *kms = arg;
> +       struct msm_disp_state *state;
> +       int ret;
> +
> +       ret = mutex_lock_interruptible(&kms->dump_mutex);

>From past experience I've seen the smmu fault handler called in hardirq
context, so it can't sleep. Is there some way to grab the register
contents without sleeping? Otherwise this will have to fork off
somewhere else that can take locks, runtime PM resume, etc.

> +       if (ret)
> +               return ret;
> +
> +       state = msm_disp_snapshot_state_sync(kms);
> +
> +       mutex_unlock(&kms->dump_mutex);
> +