From: "Christian König" <christian.koenig@amd.com>
To: Icenowy Zheng <uwu@icenowy.me>,
Jiaxun Yang <jiaxun.yang@flygoat.com>,
Huang Rui <ray.huang@amd.com>,
Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
Maxime Ripard <mripard@kernel.org>,
Thomas Zimmermann <tzimmermann@suse.de>,
David Airlie <airlied@gmail.com>, Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org
Subject: Re: [RFC PATCH 2/2] drm/ttm: downgrade cached to write_combined when snooping not available
Date: Mon, 1 Jul 2024 13:40:34 +0200 [thread overview]
Message-ID: <7e30177b-ff13-4fed-aa51-47a9cbd5d572@amd.com> (raw)
In-Reply-To: <2760BA02-8FF8-4B29-BFE2-1322B5BFB6EC@icenowy.me>
Am 29.06.24 um 22:51 schrieb Icenowy Zheng:
>
> 于 2024年6月30日 GMT+08:00 03:57:47,Jiaxun Yang <jiaxun.yang@flygoat.com> 写道:
>>
>> 在2024年6月29日六月 上午6:22,Icenowy Zheng写道:
>> [...]
>>> @@ -302,6 +302,10 @@ pgprot_t ttm_io_prot(struct ttm_buffer_object *bo,
>>> struct ttm_resource *res,
>>> caching = res->bus.caching;
>>> }
>>>
>>> + /* Downgrade cached mapping for non-snooping devices */
>>> + if (!bo->bdev->dma_coherent && caching == ttm_cached)
>>> + caching = ttm_write_combined;
>> Hi Icenowy,
>>
>> Thanks for your patch! You saved many non-coh PCIe host implementations a day!.
Ah, wait a second.
Such a thing as non-coherent PCIe implementation doesn't exist. The PCIe
specification makes it mandatory for memory access to be cache coherent.
There are a bunch of non-compliant PCIe implementations which have
broken cache coherency, but those explicitly violate the specification
and because of that are not supported.
Regards,
Christian.
>>
>> Unfortunately I don't think we can safely ttm_cached to ttm_write_comnined, we've
>> had enough drama with write combine behaviour on all different platforms.
>>
>> See drm_arch_can_wc_memory in drm_cache.h.
>>
> Yes this really sounds like an issue.
>
> Maybe the behavior of ttm_write_combined should furtherly be decided
> by drm_arch_can_wc_memory() in case of quirks?
>
>> Thanks
>>
>>> +
>>> return ttm_prot_from_caching(caching, tmp);
>>> }
>>> EXPORT_SYMBOL(ttm_io_prot);
>>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
>>> index 7b00ddf0ce49f..3335df45fba5e 100644
>>> --- a/drivers/gpu/drm/ttm/ttm_tt.c
>>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c
>>> @@ -152,6 +152,10 @@ static void ttm_tt_init_fields(struct ttm_tt *ttm,
>>> enum ttm_caching caching,
>>> unsigned long extra_pages)
>>> {
>>> + /* Downgrade cached mapping for non-snooping devices */
>>> + if (!bo->bdev->dma_coherent && caching == ttm_cached)
>>> + caching = ttm_write_combined;
>>> +
>>> ttm->num_pages = (PAGE_ALIGN(bo->base.size) >> PAGE_SHIFT) + extra_pages;
>>> ttm->page_flags = page_flags;
>>> ttm->dma_address = NULL;
>>> diff --git a/include/drm/ttm/ttm_caching.h b/include/drm/ttm/ttm_caching.h
>>> index a18f43e93abab..f92d7911f50e4 100644
>>> --- a/include/drm/ttm/ttm_caching.h
>>> +++ b/include/drm/ttm/ttm_caching.h
>>> @@ -47,7 +47,8 @@ enum ttm_caching {
>>>
>>> /**
>>> * @ttm_cached: Fully cached like normal system memory, requires that
>>> - * devices snoop the CPU cache on accesses.
>>> + * devices snoop the CPU cache on accesses. Downgraded to
>>> + * ttm_write_combined when the snooping capaiblity is missing.
>>> */
>>> ttm_cached
>>> };
>>> --
>>> 2.45.2
next prev parent reply other threads:[~2024-07-01 11:40 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-06-29 5:22 [RFC PATCH 0/2] drm/ttm: support device w/o coherency Icenowy Zheng
2024-06-29 5:22 ` [RFC PATCH 1/2] drm/ttm: save the device's DMA coherency status in ttm_device Icenowy Zheng
2024-06-29 5:22 ` [RFC PATCH 2/2] drm/ttm: downgrade cached to write_combined when snooping not available Icenowy Zheng
2024-06-29 19:57 ` Jiaxun Yang
2024-06-29 20:51 ` Icenowy Zheng
2024-07-01 11:40 ` Christian König [this message]
2024-07-01 11:52 ` Jiaxun Yang
2024-07-01 12:10 ` Christian König
2024-07-02 1:46 ` Icenowy Zheng
[not found] ` <ff1bf596-83cb-4b3e-a33a-621ac2c8171c@amd.com>
2024-07-02 9:06 ` Icenowy Zheng
[not found] ` <ae7085fd-3bca-4a4a-b465-5e4941011877@amd.com>
2024-07-02 10:01 ` Icenowy Zheng
2024-07-02 10:03 ` Jiaxun Yang
2024-07-03 8:52 ` PCIe coherency in spec (was: [RFC PATCH 2/2] drm/ttm: downgrade cached to write_combined when snooping not available) Jiaxun Yang
2024-07-03 21:08 ` Bjorn Helgaas
2024-07-04 2:00 ` Icenowy Zheng
2024-07-04 6:11 ` Christoph Hellwig
2024-07-04 6:40 ` Icenowy Zheng
2024-07-04 6:44 ` Christoph Hellwig
[not found] ` <4e958aa5-2a56-4754-b3dd-273da95f1d14@amd.com>
2024-07-02 9:36 ` [RFC PATCH 2/2] drm/ttm: downgrade cached to write_combined when snooping not available Icenowy Zheng
2024-06-30 6:52 ` Icenowy Zheng
2024-07-01 11:37 ` [RFC PATCH 0/2] drm/ttm: support device w/o coherency Christian König
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=7e30177b-ff13-4fed-aa51-47a9cbd5d572@amd.com \
--to=christian.koenig@amd.com \
--cc=airlied@gmail.com \
--cc=daniel@ffwll.ch \
--cc=dri-devel@lists.freedesktop.org \
--cc=jiaxun.yang@flygoat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=maarten.lankhorst@linux.intel.com \
--cc=mripard@kernel.org \
--cc=ray.huang@amd.com \
--cc=tzimmermann@suse.de \
--cc=uwu@icenowy.me \
/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
all inboxes | Powered by JetHome®