From: "Prakhya, Sai Praneeth" <sai.praneeth.prakhya@intel.com>
To: Joerg Roedel <joro@8bytes.org>
Cc: "iommu@lists.linux-foundation.org" <iommu@lists.linux-foundation.org>
Subject: RE: [PATCH] iommu: Remove functions that support private domain
Date: Fri, 15 May 2020 18:35:54 +0000 [thread overview]
Message-ID: <FFF73D592F13FD46B8700F0A279B802F573A6855@ORSMSX114.amr.corp.intel.com> (raw)
In-Reply-To: <20200515095919.GQ18353@8bytes.org>
Hi Joerg,
> -----Original Message-----
> From: Joerg Roedel <joro@8bytes.org>
> Sent: Friday, May 15, 2020 2:59 AM
> To: Prakhya, Sai Praneeth <sai.praneeth.prakhya@intel.com>
> Cc: iommu@lists.linux-foundation.org; Lu Baolu <baolu.lu@linux.intel.com>
> Subject: Re: [PATCH] iommu: Remove functions that support private domain
>
> On Thu, May 14, 2020 at 11:12:52PM +0000, Prakhya, Sai Praneeth wrote:
> > +static int is_driver_bound(struct device *dev, void *not_used) {
> > + int ret = 0;
> > +
> > + device_lock(dev);
> > + if (device_is_bound(dev))
> > + ret = 1;
> > + device_unlock(dev);
> > + return ret;
> > +}
>
> This locks only one device, so without lock-conversion there could be a driver
> probe after the device_unlock(), while we are probing the other devices of the
> group.
>
> > [SNIP]
> >
> > + /*
> > + * Check if any device in the group still has a driver binded to it.
> > + * This might race with device driver probing code and unfortunately
> > + * there is no clean way out of that either, locking all devices in the
> > + * group and then do the re-attach will introduce a lock-inversion with
> > + * group->mutex - Joerg.
> > + */
> > + if (iommu_group_for_each_dev(group, NULL, is_driver_bound)) {
> > + pr_err("Active drivers exist for devices in the group\n");
> > + return -EBUSY;
> > + }
>
> The lock inversion comes into the picture when this code is called from
> device(-driver) core through the bus-notifiers. The notifiers are called with the
> device already locked.
Make sense. I will look through that code.
> > Another question I have is.. if it's racy then it should be racy even
> > for one device iommu groups.. right? Why would it be racy only with
> > multiple devices iommu group?
>
> Valid point. So the device needs to be locked _while_ the default domain change
> happens. If triggered by sysfs there should be no locking problems, I guess. But
> you better try it out.
I will try this out and will update you.
Regards,
Sai
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu
next prev parent reply other threads:[~2020-05-15 18:36 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-13 22:47 [PATCH] iommu: Remove functions that support private domain Sai Praneeth Prakhya
2020-05-14 13:13 ` Joerg Roedel
2020-05-14 17:51 ` Prakhya, Sai Praneeth
2020-05-14 18:32 ` Joerg Roedel
2020-05-14 18:44 ` Prakhya, Sai Praneeth
2020-05-14 19:56 ` Joerg Roedel
2020-05-14 23:12 ` Prakhya, Sai Praneeth
2020-05-15 9:59 ` Joerg Roedel
2020-05-15 12:55 ` Lu Baolu
2020-05-15 15:46 ` Joerg Roedel
2020-05-17 8:29 ` Prakhya, Sai Praneeth
2020-05-25 13:56 ` Joerg Roedel
2020-05-28 19:31 ` Prakhya, Sai Praneeth
2020-05-15 18:35 ` Prakhya, Sai Praneeth [this message]
2020-05-15 10:01 ` Joerg Roedel
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=FFF73D592F13FD46B8700F0A279B802F573A6855@ORSMSX114.amr.corp.intel.com \
--to=sai.praneeth.prakhya@intel.com \
--cc=iommu@lists.linux-foundation.org \
--cc=joro@8bytes.org \
/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).