From: Marek Szyprowski <m.szyprowski@samsung.com>
To: Sricharan <sricharan@codeaurora.org>,
linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org,
iommu@lists.linux-foundation.org,
linux-samsung-soc@vger.kernel.org
Cc: "'Tomeu Vizoso'" <tomeu.vizoso@collabora.com>,
"'Bartlomiej Zolnierkiewicz'" <b.zolnierkie@samsung.com>,
"'Greg Kroah-Hartman'" <gregkh@linuxfoundation.org>,
"'Kevin Hilman'" <khilman@kernel.org>,
"'Rafael J. Wysocki'" <rjw@rjwysocki.net>,
"'Tomasz Figa'" <tomasz.figa@gmail.com>,
"'Krzysztof Kozlowski'" <krzk@kernel.org>,
"'Inki Dae'" <inki.dae@samsung.com>,
"'Tobias Jakobi'" <tjakobi@math.uni-bielefeld.de>,
"'Luis R. Rodriguez'" <mcgrof@kernel.org>,
"'Kukjin Kim'" <kgene@kernel.org>,
"'Mark Brown'" <broonie@kernel.org>,
"'Lukas Wunner'" <lukas@wunner.de>
Subject: Re: [PATCH v5 7/7] iommu/exynos: Use device dependency links to control runtime pm
Date: Mon, 24 Oct 2016 14:39:15 +0200 [thread overview]
Message-ID: <bff28d5b-3b39-a5bb-b381-2d7626566a2d@samsung.com> (raw)
In-Reply-To: <004a01d22df2$45ecd450$d1c67cf0$@codeaurora.org>
Hi Sricharan,
On 2016-10-24 14:29, Sricharan wrote:
>>>> This patch uses recently introduced device dependency links to track the
>>>> runtime pm state of the master's device. This way each SYSMMU controller
>>>> is set to runtime active only when its master's device is active and can
>>>> restore or save its state instead of being activated all the time when
>>>> attached to the given master device. This way SYSMMU controllers no longer
>>>> prevents respective power domains to be turned off when master's device
>>>> is not being used.
>>>>
>>>> Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
>>>> ---
>>>> drivers/iommu/exynos-iommu.c | 16 ++++++++--------
>>>> 1 file changed, 8 insertions(+), 8 deletions(-)
>>>>
>>>> diff --git a/drivers/iommu/exynos-iommu.c b/drivers/iommu/exynos-iommu.c
>>>> index 5e6d7bbf9b70..59b4f2ce4f5f 100644
>>>> --- a/drivers/iommu/exynos-iommu.c
>>>> +++ b/drivers/iommu/exynos-iommu.c
>>>> @@ -781,10 +781,6 @@ static void exynos_iommu_detach_device(struct iommu_domain *iommu_domain,
>>>> if (!has_sysmmu(dev) || owner->domain != iommu_domain)
>>>> return;
>>>>
>>>> - list_for_each_entry(data, &owner->controllers, owner_node) {
>>>> - pm_runtime_put_sync(data->sysmmu);
>>>> - }
>>>> -
>>>> mutex_lock(&owner->rpm_lock);
>>>>
>>>> list_for_each_entry(data, &owner->controllers, owner_node) {
>>>> @@ -848,10 +844,6 @@ static int exynos_iommu_attach_device(struct iommu_domain *iommu_domain,
>>>>
>>>> mutex_unlock(&owner->rpm_lock);
>>>>
>>>> - list_for_each_entry(data, &owner->controllers, owner_node) {
>>>> - pm_runtime_get_sync(data->sysmmu);
>>>> - }
>>>> -
>>>> dev_dbg(dev, "%s: Attached IOMMU with pgtable %pa\n", __func__,
>>>> &pagetable);
>>>>
>>>> @@ -1232,6 +1224,14 @@ static int exynos_iommu_of_xlate(struct device *dev,
>>>>
>>>> list_add_tail(&data->owner_node, &owner->controllers);
>>>> data->master = dev;
>>>> +
>>>> + /*
>>>> + * SYSMMU will be runtime activated via device link (dependency) to its
>>>> + * master device, so there are no direct calls to pm_runtime_get/put
>>>> + * in this driver.
>>>> + */
>>>> + device_link_add(dev, data->sysmmu, DL_FLAG_PM_RUNTIME);
>>>> +
>>> So in the case of master with multiple sids, this would be called multiple times
>>> for the same master ?
>> I don't know what is "multiple sids" case, but if given SYSMMU master
>> device (supplier)
>> has multiple SYSMMU controllers (consumers), then links will be created
>> for each SYSMMU
>> controller. Please note that this code is based on vanilla v4.9-rc1,
>> which calls
>> of_xlate() callback only once for every iommu for given master device.
>> Your IOMMU
>> deferred probe patches change this, but I already posted a fix for
>> Exynos IOMMU driver
>> to handle such case.
> By multiple sids, i meant iommus = <&phandle sid1 sid2 .. sidn> case,
> so xlate would be called multiples for the same master without deferred
> probing also. But the fix that you showed on the other thread would work
> here as well or maybe if you dont have masters with multiple sids you wont
> have any issues as well.
Exynos SYSMMU driver always use "#iommu-cells = <0>", so it doesn't support
multiple sids. However there is a case with 2 SYSMMU controllers attached
to the same master device: "iommus = <&sysmmu_mfc_l>, <&sysmmu_mfc_r>;".
Best regards
--
Marek Szyprowski, PhD
Samsung R&D Institute Poland
next prev parent reply other threads:[~2016-10-24 12:39 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <CGME20161020072330eucas1p2b09ad8d091171edbac9449815fdc0fb7@eucas1p2.samsung.com>
2016-10-20 7:22 ` [PATCH v5 0/7] Exynos IOMMU: proper runtime PM support (use device dependencies) Marek Szyprowski
[not found] ` <CGME20161020072331eucas1p1af7dc7270b0b19168b949f3416eda474@eucas1p1.samsung.com>
2016-10-20 7:22 ` [PATCH v5 1/7] iommu/exynos: Remove excessive, useless debug Marek Szyprowski
[not found] ` <CGME20161020072332eucas1p1d980c1659979bd5bc2918bfc9d40a415@eucas1p1.samsung.com>
2016-10-20 7:22 ` [PATCH v5 2/7] iommu/exynos: Remove dead code Marek Szyprowski
[not found] ` <CGME20161020072332eucas1p26960035de3007724498d59057329683d@eucas1p2.samsung.com>
2016-10-20 7:22 ` [PATCH v5 3/7] iommu/exynos: Simplify internal enable/disable functions Marek Szyprowski
[not found] ` <CGME20161020072333eucas1p25b638379091939f10b3c9eb5d89a031e@eucas1p2.samsung.com>
2016-10-20 7:22 ` [PATCH v5 4/7] iommu/exynos: Set master device once on boot Marek Szyprowski
[not found] ` <CGME20161020072334eucas1p2a159a25a2875611eff208381ebdb2e84@eucas1p2.samsung.com>
2016-10-20 7:22 ` [PATCH v5 5/7] iommu/exynos: Rework and fix internal locking Marek Szyprowski
[not found] ` <CGME20161020072335eucas1p209675d6fbf39e5045281e8023fa9d234@eucas1p2.samsung.com>
2016-10-20 7:22 ` [PATCH v5 6/7] iommu/exynos: Add runtime pm support Marek Szyprowski
2016-10-22 5:50 ` Sricharan
2016-10-24 5:19 ` Marek Szyprowski
2016-10-24 12:15 ` Sricharan
[not found] ` <CGME20161020072336eucas1p24a2b020f69b6ae1f55e1760e6e0e94f9@eucas1p2.samsung.com>
2016-10-20 7:22 ` [PATCH v5 7/7] iommu/exynos: Use device dependency links to control runtime pm Marek Szyprowski
2016-10-23 9:49 ` Sricharan
2016-10-24 5:30 ` Marek Szyprowski
2016-10-24 12:29 ` Sricharan
2016-10-24 12:39 ` Marek Szyprowski [this message]
2016-10-25 6:53 ` Sricharan
2016-11-07 21:47 ` Luis R. Rodriguez
2016-11-08 7:27 ` Marek Szyprowski
2016-11-08 15:30 ` Lukas Wunner
2016-11-09 23:55 ` Luis R. Rodriguez
2016-11-10 0:05 ` Rafael J. Wysocki
2016-11-10 0:12 ` Luis R. Rodriguez
2016-11-10 0:20 ` Rafael J. Wysocki
2016-11-09 23:56 ` Rafael J. Wysocki
2016-11-16 9:30 ` Lukas Wunner
2016-11-19 11:11 ` Lukas Wunner
2016-11-21 13:11 ` Marek Szyprowski
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=bff28d5b-3b39-a5bb-b381-2d7626566a2d@samsung.com \
--to=m.szyprowski@samsung.com \
--cc=b.zolnierkie@samsung.com \
--cc=broonie@kernel.org \
--cc=gregkh@linuxfoundation.org \
--cc=inki.dae@samsung.com \
--cc=iommu@lists.linux-foundation.org \
--cc=kgene@kernel.org \
--cc=khilman@kernel.org \
--cc=krzk@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=linux-samsung-soc@vger.kernel.org \
--cc=lukas@wunner.de \
--cc=mcgrof@kernel.org \
--cc=rjw@rjwysocki.net \
--cc=sricharan@codeaurora.org \
--cc=tjakobi@math.uni-bielefeld.de \
--cc=tomasz.figa@gmail.com \
--cc=tomeu.vizoso@collabora.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).