From: "Danilo Krummrich" <dakr@kernel.org>
To: "Tzung-Bi Shih" <tzungbi@kernel.org>
Cc: "Benson Leung" <bleung@chromium.org>,
"Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
"Rafael J . Wysocki" <rafael@kernel.org>,
"Jonathan Corbet" <corbet@lwn.net>,
"Shuah Khan" <shuah@kernel.org>,
"Dawid Niedzwiecki" <dawidn@google.com>,
<linux-doc@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<chrome-platform@lists.linux.dev>,
<linux-kselftest@vger.kernel.org>
Subject: Re: [PATCH v3 1/5] revocable: Revocable resource management
Date: Fri, 12 Sep 2025 11:05:20 +0200 [thread overview]
Message-ID: <DCQP9ZJ0DFBO.3O3W57IDYN08I@kernel.org> (raw)
In-Reply-To: <20250912081718.3827390-2-tzungbi@kernel.org>
On Fri Sep 12, 2025 at 10:17 AM CEST, Tzung-Bi Shih wrote:
> +/**
> + * struct revocable_provider - A handle for resource provider.
> + * @srcu: The SRCU to protect the resource.
> + * @res: The pointer of resource. It can point to anything.
> + * @kref: The refcount for this handle.
> + */
> +struct revocable_provider {
> + struct srcu_struct srcu;
> + void __rcu *res;
> + struct kref kref;
> +};
I think a revocable provider should provide an optional revoke() callback where
users of the revocable provider can release the revoked resource.
But this can also be done as a follow-up.
> +/**
> + * struct revocable - A handle for resource consumer.
> + * @rp: The pointer of resource provider.
> + * @idx: The index for the RCU critical section.
> + */
> +struct revocable {
> + struct revocable_provider *rp;
> + int idx;
> +};
I think I asked about this in the previous version (but I don't remember if
there was an answer):
In Rust we get away with a single Revocable<T> structure, but we're using RCU
instead of SRCU. It seems to me that the split between struct revocable and
struct revocable_provider is only for the SRCU index? Or is there any other
reason?
> +/**
> + * revocable_provider_free() - Free struct revocable_provider.
> + * @rp: The pointer of resource provider.
> + *
> + * This sets the resource `(struct revocable_provider *)->res` to NULL to
> + * indicate the resource has gone.
> + *
> + * This drops the refcount to the resource provider. If it is the final
> + * reference, revocable_provider_release() will be called to free the struct.
> + */
> +void revocable_provider_free(struct revocable_provider *rp)
> +{
> + rcu_assign_pointer(rp->res, NULL);
> + synchronize_srcu(&rp->srcu);
> + kref_put(&rp->kref, revocable_provider_release);
> +}
> +EXPORT_SYMBOL_GPL(revocable_provider_free);
I think naming this "free" is a bit misleading, since what it basically does is
(1) Revoke access to the resource.
(2) Drop a reference count of struct revocable_provider.
In a typical application there may still be struct revocable instances that have
a reference to the provider, so we can't claim that it's freed here.
So, given that, I'd rather call this revocable_provider_revoke().
> +static void devm_revocable_provider_free(void *data)
> +{
> + struct revocable_provider *rp = data;
> +
> + revocable_provider_free(rp);
> +}
Same here, I'd call this devm_revocable_provider_revoke().
> +DEFINE_FREE(revocable, struct revocable *, if (_T) revocable_release(_T))
> +
> +#define _REVOCABLE(_rev, _label, _res) \
> + for (struct revocable *__UNIQUE_ID(name) __free(revocable) = _rev; \
> + (_res = revocable_try_access(_rev)) || true; ({ goto _label; })) \
> + if (0) { \
> +_label: \
> + break; \
> + } else
> +
> +#define REVOCABLE(_rev, _res) _REVOCABLE(_rev, __UNIQUE_ID(label), _res)
This is basically the same as Revocable::try_access_with() [1] in Rust, i.e.
try to access and run a closure.
Admittedly, REVOCABLE_TRY_ACCESS_WITH() is pretty verbose and I also do not have
a great idea to shorten it.
Maybe you have a good idea, otherwise I'm also fine with the current name.
Otherwise, maybe it's worth to link to the Rust Revocable API for reference?
With *_free() renamed to *_revoke(), feel free to add:
Acked-by: Danilo Krummrich <dakr@kernel.org>
[1] https://rust.docs.kernel.org/kernel/revocable/struct.Revocable.html#method.try_access_with
next prev parent reply other threads:[~2025-09-12 9:05 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-12 8:17 [PATCH v3 0/5] platform/chrome: Fix a possible UAF via revocable Tzung-Bi Shih
2025-09-12 8:17 ` [PATCH v3 1/5] revocable: Revocable resource management Tzung-Bi Shih
2025-09-12 9:05 ` Danilo Krummrich [this message]
2025-09-13 15:56 ` Tzung-Bi Shih
2025-09-12 13:27 ` Jonathan Corbet
2025-09-13 15:56 ` Tzung-Bi Shih
2025-09-17 5:24 ` Tzung-Bi Shih
2025-09-22 18:35 ` Simona Vetter
2025-09-12 8:17 ` [PATCH v3 2/5] revocable: Add Kunit test cases Tzung-Bi Shih
2025-09-12 8:17 ` [PATCH v3 3/5] selftests: revocable: Add kselftest cases Tzung-Bi Shih
2025-09-12 8:17 ` [PATCH v3 4/5] platform/chrome: Protect cros_ec_device lifecycle with revocable Tzung-Bi Shih
2025-09-12 8:17 ` [PATCH v3 5/5] platform/chrome: cros_ec_chardev: Consume cros_ec_device via revocable Tzung-Bi Shih
2025-09-12 8:30 ` [PATCH v3 0/5] platform/chrome: Fix a possible UAF " Greg Kroah-Hartman
2025-09-12 8:34 ` Danilo Krummrich
2025-09-12 9:20 ` Laurent Pinchart
2025-09-12 9:09 ` Krzysztof Kozlowski
2025-09-12 9:24 ` Bartosz Golaszewski
2025-09-12 12:49 ` Tzung-Bi Shih
2025-09-12 13:26 ` Laurent Pinchart
2025-09-12 13:39 ` Greg Kroah-Hartman
2025-09-12 13:45 ` Laurent Pinchart
2025-09-12 13:46 ` Bartosz Golaszewski
2025-09-12 13:59 ` Laurent Pinchart
2025-09-12 14:19 ` Greg Kroah-Hartman
2025-09-12 14:26 ` Laurent Pinchart
2025-09-12 14:40 ` Greg Kroah-Hartman
2025-09-12 14:44 ` Bartosz Golaszewski
2025-09-12 14:54 ` Laurent Pinchart
2025-09-12 16:22 ` Danilo Krummrich
2025-09-13 16:17 ` Laurent Pinchart
2025-09-22 22:43 ` dan.j.williams
2025-09-13 15:55 ` Tzung-Bi Shih
2025-09-13 16:14 ` Laurent Pinchart
2025-09-23 8:20 ` Tzung-Bi Shih
2025-09-12 14:53 ` Laurent Pinchart
2025-09-22 15:10 ` Jason Gunthorpe
2025-09-22 15:55 ` Danilo Krummrich
2025-09-22 17:40 ` Jason Gunthorpe
2025-09-22 18:42 ` Greg Kroah-Hartman
2025-09-22 20:17 ` Jason Gunthorpe
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=DCQP9ZJ0DFBO.3O3W57IDYN08I@kernel.org \
--to=dakr@kernel.org \
--cc=bleung@chromium.org \
--cc=chrome-platform@lists.linux.dev \
--cc=corbet@lwn.net \
--cc=dawidn@google.com \
--cc=gregkh@linuxfoundation.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=rafael@kernel.org \
--cc=shuah@kernel.org \
--cc=tzungbi@kernel.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).