public inbox for linux-mm@kvack.org
 help / color / mirror / Atom feed
From: Samiullah Khawaja <skhawaja@google.com>
To: Pasha Tatashin <pasha.tatashin@soleen.com>
Cc: rppt@kernel.org, akpm@linux-foundation.org, linux-mm@kvack.org,
	 linux-kernel@vger.kernel.org, dmatlack@google.com,
	pratyush@kernel.org
Subject: Re: [PATCH v3 03/10] liveupdate: Protect file handler list with rwsem
Date: Mon, 30 Mar 2026 16:48:19 +0000	[thread overview]
Message-ID: <acqnQ64U8Cxst-ce@google.com> (raw)
In-Reply-To: <20260327033335.696621-4-pasha.tatashin@soleen.com>

On Fri, Mar 27, 2026 at 03:33:27AM +0000, Pasha Tatashin wrote:
>Because liveupdate file handlers will no longer hold a module reference
>when registered, we must ensure that the access to the handler list is
>protected against concurrent module unloading.

Nit: Here we make an assumption that the file (and flb) handler
lifecycle is bound with the module lifecycle. It is a fair assumption,
but maybe this can be documented somewhere?
>
>Utilize the global luo_register_rwlock to protect the global registry of
>file handlers. Read locks are taken during list traversals in
>luo_preserve_file() and luo_file_deserialize(). Write locks are taken
>during registration and unregistration.
>
>Signed-off-by: Pasha Tatashin <pasha.tatashin@soleen.com>
>---
> kernel/liveupdate/luo_core.c     |  6 ++++++
> kernel/liveupdate/luo_file.c     | 22 +++++++++++++++++-----
> kernel/liveupdate/luo_internal.h |  2 ++
> 3 files changed, 25 insertions(+), 5 deletions(-)
>
>diff --git a/kernel/liveupdate/luo_core.c b/kernel/liveupdate/luo_core.c
>index dda7bb57d421..f9ae9364a962 100644
>--- a/kernel/liveupdate/luo_core.c
>+++ b/kernel/liveupdate/luo_core.c
>@@ -54,6 +54,7 @@
> #include <linux/liveupdate.h>
> #include <linux/miscdevice.h>
> #include <linux/mm.h>
>+#include <linux/rwsem.h>
> #include <linux/sizes.h>
> #include <linux/string.h>
> #include <linux/unaligned.h>
>@@ -68,6 +69,11 @@ static struct {
> 	u64 liveupdate_num;
> } luo_global;
>
>+/*
>+ * luo_register_rwlock - Protects registration of file handlers and FLBs.
>+ */
>+DECLARE_RWSEM(luo_register_rwlock);
>+
> static int __init early_liveupdate_param(char *buf)
> {
> 	return kstrtobool(buf, &luo_global.enabled);
>diff --git a/kernel/liveupdate/luo_file.c b/kernel/liveupdate/luo_file.c
>index a6d98fc75d25..4aea17a94b4f 100644
>--- a/kernel/liveupdate/luo_file.c
>+++ b/kernel/liveupdate/luo_file.c
>@@ -277,12 +277,14 @@ int luo_preserve_file(struct luo_file_set *file_set, u64 token, int fd)
> 		goto  err_fput;
>
> 	err = -ENOENT;
>+	down_read(&luo_register_rwlock);
> 	list_private_for_each_entry(fh, &luo_file_handler_list, list) {
> 		if (fh->ops->can_preserve(fh, file)) {
> 			err = 0;
> 			break;
> 		}
> 	}
>+	up_read(&luo_register_rwlock);

We took the read lock here when running can_preserve, but then we use
the fh without taking the lock later before calling file_preserve. This
is safe since the module reference is taken and fh will not go away
(based on the assumption I mentioned above). Maybe add a comment here
that documents this assumption.
>
> 	/* err is still -ENOENT if no handler was found */
> 	if (err)
>@@ -777,12 +779,14 @@ int luo_file_deserialize(struct luo_file_set *file_set,
> 		bool handler_found = false;
> 		struct luo_file *luo_file;
>
>+		down_read(&luo_register_rwlock);
> 		list_private_for_each_entry(fh, &luo_file_handler_list, list) {
> 			if (!strcmp(fh->compatible, file_ser[i].compatible)) {
> 				handler_found = true;
> 				break;
> 			}
> 		}
>+		up_read(&luo_register_rwlock);
>
> 		if (!handler_found) {
> 			pr_warn("No registered handler for compatible '%.*s'\n",
>@@ -851,32 +855,36 @@ int liveupdate_register_file_handler(struct liveupdate_file_handler *fh)
> 	if (!luo_session_quiesce())
> 		return -EBUSY;
>
>+	down_write(&luo_register_rwlock);
> 	/* Check for duplicate compatible strings */
> 	list_private_for_each_entry(fh_iter, &luo_file_handler_list, list) {
> 		if (!strcmp(fh_iter->compatible, fh->compatible)) {
> 			pr_err("File handler registration failed: Compatible string '%s' already registered.\n",
> 			       fh->compatible);
> 			err = -EEXIST;
>-			goto err_resume;
>+			goto err_unlock;
> 		}
> 	}
>
> 	/* Pin the module implementing the handler */
> 	if (!try_module_get(fh->ops->owner)) {
> 		err = -EAGAIN;
>-		goto err_resume;
>+		goto err_unlock;
> 	}
>
> 	INIT_LIST_HEAD(&ACCESS_PRIVATE(fh, flb_list));
> 	INIT_LIST_HEAD(&ACCESS_PRIVATE(fh, list));
> 	list_add_tail(&ACCESS_PRIVATE(fh, list), &luo_file_handler_list);
>+	up_write(&luo_register_rwlock);
>+
> 	luo_session_resume();
>
> 	liveupdate_test_register(fh);
>
> 	return 0;
>
>-err_resume:
>+err_unlock:
>+	up_write(&luo_register_rwlock);
> 	luo_session_resume();
> 	return err;
> }
>@@ -910,16 +918,20 @@ int liveupdate_unregister_file_handler(struct liveupdate_file_handler *fh)
> 	if (!luo_session_quiesce())
> 		goto err_register;
>
>+	down_write(&luo_register_rwlock);
> 	if (!list_empty(&ACCESS_PRIVATE(fh, flb_list)))
>-		goto err_resume;
>+		goto err_unlock;
>
> 	list_del(&ACCESS_PRIVATE(fh, list));
>+	up_write(&luo_register_rwlock);
>+
> 	module_put(fh->ops->owner);
> 	luo_session_resume();
>
> 	return 0;
>
>-err_resume:
>+err_unlock:
>+	up_write(&luo_register_rwlock);
> 	luo_session_resume();
> err_register:
> 	liveupdate_test_register(fh);
>diff --git a/kernel/liveupdate/luo_internal.h b/kernel/liveupdate/luo_internal.h
>index 8083d8739b09..4bfe00ac8866 100644
>--- a/kernel/liveupdate/luo_internal.h
>+++ b/kernel/liveupdate/luo_internal.h
>@@ -77,6 +77,8 @@ struct luo_session {
> 	struct mutex mutex;
> };
>
>+extern struct rw_semaphore luo_register_rwlock;
>+
> int luo_session_create(const char *name, struct file **filep);
> int luo_session_retrieve(const char *name, struct file **filep);
> int __init luo_session_setup_outgoing(void *fdt);
>-- 
>2.43.0
>


  reply	other threads:[~2026-03-30 16:48 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-27  3:33 [PATCH v3 00/10] liveupdate: Fix module unloading and unregister API Pasha Tatashin
2026-03-27  3:33 ` [PATCH v3 01/10] liveupdate: Safely print untrusted strings Pasha Tatashin
2026-03-27 13:16   ` Pasha Tatashin
2026-03-31  9:40     ` Pratyush Yadav
2026-03-31  9:50   ` Pratyush Yadav
2026-03-31 16:35     ` Pasha Tatashin
2026-03-27  3:33 ` [PATCH v3 02/10] liveupdate: Synchronize lazy initialization of FLB private state Pasha Tatashin
2026-03-31 10:38   ` Pratyush Yadav
2026-03-31 16:41     ` Pasha Tatashin
2026-03-31 19:22       ` Pratyush Yadav
2026-03-31 19:38         ` Pasha Tatashin
2026-03-27  3:33 ` [PATCH v3 03/10] liveupdate: Protect file handler list with rwsem Pasha Tatashin
2026-03-30 16:48   ` Samiullah Khawaja [this message]
2026-03-30 19:32     ` Pasha Tatashin
2026-03-31 19:24   ` Pratyush Yadav
2026-03-27  3:33 ` [PATCH v3 04/10] liveupdate: Protect FLB lists with luo_register_rwlock Pasha Tatashin
2026-03-31 19:33   ` Pratyush Yadav
2026-03-27  3:33 ` [PATCH v3 05/10] liveupdate: Defer FLB module refcounting to active sessions Pasha Tatashin
2026-03-30 16:56   ` Samiullah Khawaja
2026-03-30 19:28     ` Pasha Tatashin
2026-04-02 16:21   ` Pratyush Yadav
2026-03-27  3:33 ` [PATCH v3 06/10] liveupdate: Remove luo_session_quiesce() Pasha Tatashin
2026-04-02 16:27   ` Pratyush Yadav
2026-03-27  3:33 ` [PATCH v3 07/10] liveupdate: Auto unregister FLBs on file handler unregistration Pasha Tatashin
2026-04-03 10:17   ` Pratyush Yadav
2026-03-27  3:33 ` [PATCH v3 08/10] liveupdate: Remove liveupdate_test_unregister() Pasha Tatashin
2026-04-03 10:20   ` Pratyush Yadav
2026-03-27  3:33 ` [PATCH v3 09/10] liveupdate: Make unregister functions return void Pasha Tatashin
2026-03-27 14:41   ` Pasha Tatashin
2026-04-03 10:41   ` Pratyush Yadav
2026-03-27  3:33 ` [PATCH v3 10/10] liveupdate: Defer file handler module refcounting to active sessions Pasha Tatashin
2026-03-27 17:14   ` Andrew Morton
2026-04-03 10:42   ` Pratyush Yadav
2026-03-27 17:24 ` [PATCH v3 00/10] liveupdate: Fix module unloading and unregister API Andrew Morton

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=acqnQ64U8Cxst-ce@google.com \
    --to=skhawaja@google.com \
    --cc=akpm@linux-foundation.org \
    --cc=dmatlack@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=pasha.tatashin@soleen.com \
    --cc=pratyush@kernel.org \
    --cc=rppt@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