dri-devel.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: Icenowy Zheng <uwu@icenowy.me>
To: Drew Fustini <fustini@kernel.org>, Guo Ren <guoren@kernel.org>,
	Fu Wei <wefu@redhat.com>, Lucas Stach <l.stach@pengutronix.de>,
	Russell King <linux+etnaviv@armlinux.org.uk>,
	Christian Gmeiner <christian.gmeiner@gmail.com>,
	 David Airlie <airlied@gmail.com>,
	Simona Vetter <simona@ffwll.ch>
Cc: linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
	 etnaviv@lists.freedesktop.org, dri-devel@lists.freedesktop.org
Subject: Re: [PATCH 6/7] drm/etnaviv: add shared context support for iommuv2
Date: Tue, 02 Sep 2025 23:40:02 +0800	[thread overview]
Message-ID: <05ef7c0df0a2235277030b9e33f34082e8938faa.camel@icenowy.me> (raw)
In-Reply-To: <20250816074757.2559055-7-uwu@icenowy.me>

在 2025-08-16星期六的 15:47 +0800,Icenowy Zheng写道:
> Unfortunately the GC620 GPU seems to have broken PTA capibility, and
> switching page table ID in command stream after it's running won't
> work.
> As directly switching mtlb isn't working either, there will be no
> reliable way to switch page table in the command stream, and a shared
> context, like iommuv1, is needed.
> 
> Add support for this shared context situation. Shared context is set
> when the broken PTA is known, and the context allocation code will be
> made short circuit when a shared context is set.
> 
> Signed-off-by: Icenowy Zheng <uwu@icenowy.me>
> ---
>  drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c | 8 ++++++++
>  drivers/gpu/drm/etnaviv/etnaviv_mmu.c      | 1 +
>  drivers/gpu/drm/etnaviv/etnaviv_mmu.h      | 2 ++
>  3 files changed, 11 insertions(+)
> 
> diff --git a/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
> b/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
> index 5654a604c70cf..960ba3d659dc5 100644
> --- a/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
> +++ b/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c

Well, I forgot to clean shared_context in etnaviv_iommuv2_free() when
the shared context is torn down...

> @@ -273,6 +273,12 @@ etnaviv_iommuv2_context_alloc(struct
> etnaviv_iommu_global *global)
>         struct etnaviv_iommu_context *context;
>  
>         mutex_lock(&global->lock);
> +       if (global->shared_context) {
> +               context = global->shared_context;
> +               etnaviv_iommu_context_get(context);
> +               mutex_unlock(&global->lock);
> +               return context;
> +       }
>  
>         v2_context = vzalloc(sizeof(*v2_context));
>         if (!v2_context)
> @@ -301,6 +307,8 @@ etnaviv_iommuv2_context_alloc(struct
> etnaviv_iommu_global *global)
>         mutex_init(&context->lock);
>         INIT_LIST_HEAD(&context->mappings);
>         drm_mm_init(&context->mm, SZ_4K, (u64)SZ_1G * 4 - SZ_4K);
> +       if (global->v2.broken_pta)
> +               global->shared_context = context;
>  
>         mutex_unlock(&global->lock);
>         return context;
> diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
> b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
> index df5192083b201..a0f9c950504e0 100644
> --- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
> +++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
> @@ -504,6 +504,7 @@ int etnaviv_iommu_global_init(struct etnaviv_gpu
> *gpu)
>         memset32(global->bad_page_cpu, 0xdead55aa, SZ_4K /
> sizeof(u32));
>  
>         if (version == ETNAVIV_IOMMU_V2) {
> +               global->v2.broken_pta = gpu->identity.model ==
> chipModel_GC620;
>                 global->v2.pta_cpu = dma_alloc_wc(dev,
> ETNAVIV_PTA_SIZE,
>                                                &global->v2.pta_dma,
> GFP_KERNEL);
>                 if (!global->v2.pta_cpu)
> diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.h
> b/drivers/gpu/drm/etnaviv/etnaviv_mmu.h
> index 2ec4acda02bc6..5627d2a0d0237 100644
> --- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.h
> +++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.h
> @@ -55,6 +55,8 @@ struct etnaviv_iommu_global {
>                 u64 *pta_cpu;
>                 dma_addr_t pta_dma;
>                 DECLARE_BITMAP(pta_alloc, ETNAVIV_PTA_ENTRIES);
> +               /* Whether runtime switching page table ID will fail
> */
> +               bool broken_pta;
>         } v2;
>  };
>  


  reply	other threads:[~2025-09-02 15:40 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-08-16  7:47 [PATCH 0/7] drm/etnaviv: add support for GC620 on T-Head TH1520 Icenowy Zheng
2025-08-16  7:47 ` [PATCH 1/7] drm/etnaviv: add HWDB entry for GC620 r5552 c20b Icenowy Zheng
2025-08-16  7:47 ` [PATCH 2/7] drm/etnaviv: add handle for GPUs with only SECURITY_AHB flag Icenowy Zheng
2025-08-16  7:47 ` [PATCH 3/7] drm/etnaviv: setup DEC400EX on GC620 r5552 Icenowy Zheng
2025-08-16  7:47 ` [PATCH 4/7] drm/etnaviv: protect whole iommuv2 ctx alloc func under global mutex Icenowy Zheng
2025-08-16  7:47 ` [PATCH 5/7] drm/etnaviv: prepare for shared_context support for iommuv2 Icenowy Zheng
2025-08-16  7:47 ` [PATCH 6/7] drm/etnaviv: add shared context " Icenowy Zheng
2025-09-02 15:40   ` Icenowy Zheng [this message]
2025-08-16  7:47 ` [PATCH 7/7] [NOT FOR UPSTREAM] riscv: dts: thead: enable GC620 G2D on TH1520 Icenowy Zheng

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=05ef7c0df0a2235277030b9e33f34082e8938faa.camel@icenowy.me \
    --to=uwu@icenowy.me \
    --cc=airlied@gmail.com \
    --cc=christian.gmeiner@gmail.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=etnaviv@lists.freedesktop.org \
    --cc=fustini@kernel.org \
    --cc=guoren@kernel.org \
    --cc=l.stach@pengutronix.de \
    --cc=linux+etnaviv@armlinux.org.uk \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=simona@ffwll.ch \
    --cc=wefu@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).