From: Will Deacon <will@kernel.org>
To: Tian Tao <tiantao6@hisilicon.com>
Cc: jean-philippe@linaro.org, iommu@lists.linux-foundation.org,
robin.murphy@arm.com
Subject: Re: [PATCH v2] iommu/arm-smmu-v3: Fix not checking return value about devm_add_action
Date: Tue, 8 Dec 2020 13:53:50 +0000 [thread overview]
Message-ID: <20201208135350.GA6274@willie-the-truck> (raw)
In-Reply-To: <1607429722-54861-1-git-send-email-tiantao6@hisilicon.com>
On Tue, Dec 08, 2020 at 08:15:22PM +0800, Tian Tao wrote:
> Use devm_add_action_or_reset to avoid the situation where the release
> function is not called when devm_add_action returns an error.
>
> Signed-off-by: Tian Tao <tiantao6@hisilicon.com>
> ---
> v2:
> repositioning devm_add_action_or_reset in the function
> arm_smmu_setup_msis, and check the return value.
> ---
> drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 13 +++++++++----
> 1 file changed, 9 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> index 2ddf5ec..b4d3b7d 100644
> --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> @@ -2680,7 +2680,8 @@ static int arm_smmu_cmdq_init(struct arm_smmu_device *smmu)
> ret = -ENOMEM;
> } else {
> cmdq->valid_map = bitmap;
> - devm_add_action(smmu->dev, arm_smmu_cmdq_free_bitmap, bitmap);
> + ret = devm_add_action_or_reset(smmu->dev,
> + arm_smmu_cmdq_free_bitmap, bitmap);
> }
>
> return ret;
> @@ -2921,6 +2922,13 @@ static void arm_smmu_setup_msis(struct arm_smmu_device *smmu)
> return;
> }
>
> + /* Add callback to free MSIs on teardown */
> + ret = devm_add_action_or_reset(dev, arm_smmu_free_msis, dev);
> + if (ret) {
> + dev_warn(dev, "failed to add callback to free MSIs on teardown\n");
> + return;
Honestly, wouldn't we just be better leaking memory in this case? Tearing
down the SMMU is a pretty specialist sport _anyway_, but this seems to throw
the baby out with the bath water by failing to initialise because we can't
defer freeing something that we've already allocated. I think we're better
off continuing and trying to get the device up and running.
In fact, the same applies to the cmdq 'valid_map' too -- why do we care?
WIll
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu
prev parent reply other threads:[~2020-12-08 13:53 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-08 12:15 [PATCH v2] iommu/arm-smmu-v3: Fix not checking return value about devm_add_action Tian Tao
2020-12-08 13:53 ` Will Deacon [this message]
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=20201208135350.GA6274@willie-the-truck \
--to=will@kernel.org \
--cc=iommu@lists.linux-foundation.org \
--cc=jean-philippe@linaro.org \
--cc=robin.murphy@arm.com \
--cc=tiantao6@hisilicon.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).