From: Jonathan Marek <jonathan@marek.ca>
To: linux-arm-msm@vger.kernel.org
Cc: srinivas.kandagatla@linaro.org, Arnd Bergmann <arnd@arndb.de>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
linux-kernel@vger.kernel.org (open list)
Subject: [PATCH 2/3] fastrpc: move fl->maps list removal into fastrpc_free_map
Date: Wed, 17 Feb 2021 22:20:54 -0500 [thread overview]
Message-ID: <20210218032055.28247-3-jonathan@marek.ca> (raw)
In-Reply-To: <20210218032055.28247-1-jonathan@marek.ca>
This fixes the incredibly broken behavior of fastrpc_context_free(),
which calls fastrpc_map_put() but does not remove the map from the list
when it is free'd, causing use-after-free errors.
fl->lock needs to be held not just for list_del(), but also kref_put, to
avoid a race condition with fastrpc_map_find() logic.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
---
drivers/misc/fastrpc.c | 27 +++++++++++++--------------
1 file changed, 13 insertions(+), 14 deletions(-)
diff --git a/drivers/misc/fastrpc.c b/drivers/misc/fastrpc.c
index 4fabea0c1551..170352b43ab6 100644
--- a/drivers/misc/fastrpc.c
+++ b/drivers/misc/fastrpc.c
@@ -245,19 +245,21 @@ static void fastrpc_free_map(struct kref *ref)
dma_buf_put(map->buf);
}
+ list_del(&map->node);
+
kfree(map);
}
static void fastrpc_map_put(struct fastrpc_map *map)
{
- if (map)
- kref_put(&map->refcount, fastrpc_free_map);
+ spin_lock(&map->fl->lock);
+ kref_put(&map->refcount, fastrpc_free_map);
+ spin_unlock(&map->fl->lock);
}
static void fastrpc_map_get(struct fastrpc_map *map)
{
- if (map)
- kref_get(&map->refcount);
+ kref_get(&map->refcount);
}
static int fastrpc_map_find(struct fastrpc_user *fl, int fd,
@@ -351,8 +353,10 @@ static void fastrpc_context_free(struct kref *ref)
ctx = container_of(ref, struct fastrpc_invoke_ctx, refcount);
cctx = ctx->cctx;
- for (i = 0; i < ctx->nscalars; i++)
- fastrpc_map_put(ctx->maps[i]);
+ for (i = 0; i < ctx->nscalars; i++) {
+ if (ctx->maps[i])
+ fastrpc_map_put(ctx->maps[i]);
+ }
if (ctx->buf)
fastrpc_buf_free(ctx->buf);
@@ -1103,12 +1107,8 @@ static int fastrpc_init_create_process(struct fastrpc_user *fl,
fl->init_mem = NULL;
fastrpc_buf_free(imem);
err_alloc:
- if (map) {
- spin_lock(&fl->lock);
- list_del(&map->node);
- spin_unlock(&fl->lock);
+ if (map)
fastrpc_map_put(map);
- }
err:
kfree(args);
@@ -1185,10 +1185,9 @@ static int fastrpc_device_release(struct inode *inode, struct file *file)
fastrpc_context_put(ctx);
}
- list_for_each_entry_safe(map, m, &fl->maps, node) {
- list_del(&map->node);
+ list_for_each_entry_safe(map, m, &fl->maps, node)
fastrpc_map_put(map);
- }
+ WARN_ON(!list_empty(&fl->maps));
list_for_each_entry_safe(buf, b, &fl->mmaps, node) {
list_del(&buf->node);
--
2.26.1
next prev parent reply other threads:[~2021-02-18 3:23 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-18 3:20 [PATCH 0/3] fastrpc: fix dmabuf arguments Jonathan Marek
2021-02-18 3:20 ` [PATCH 1/3] fastrpc: always use fl->lock and remove fl->mutex Jonathan Marek
2021-02-18 3:20 ` Jonathan Marek [this message]
2021-02-18 3:20 ` [PATCH 3/3] fastrpc: remove redundant fastrpc_map_create() call Jonathan Marek
2021-02-22 12:37 ` Srinivas Kandagatla
2021-02-22 13:53 ` Jonathan Marek
2021-02-23 9:43 ` Srinivas Kandagatla
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=20210218032055.28247-3-jonathan@marek.ca \
--to=jonathan@marek.ca \
--cc=arnd@arndb.de \
--cc=gregkh@linuxfoundation.org \
--cc=linux-arm-msm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=srinivas.kandagatla@linaro.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).