[PATCH] cxl: Fix leaking pid refs in some error paths

From: Vaibhav Jain <hidden>
Date: 2016-10-21 05:46:51

In some error paths in functions cxl_start_context and
afu_ioctl_start_work pid references to the current & group-leader tasks
can leak after they are taken. This patch fixes these error paths to
release these pid references before exiting the error path.

This patch is based on earlier patch "cxl: Prevent adapter reset
if an active context exists" at
https://patchwork.ozlabs.org/patch/682187/

Fixes: 7b8ad495("cxl: Fix DSI misses when the context owning task exits")
Reported-by: Frederic Barrat <redacted>
Signed-off-by: Vaibhav Jain <redacted>
---
 drivers/misc/cxl/api.c  |  2 ++
 drivers/misc/cxl/file.c | 22 +++++++++++++---------
 2 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/drivers/misc/cxl/api.c b/drivers/misc/cxl/api.c
index af23d7d..2e5233b 100644
--- a/drivers/misc/cxl/api.c
+++ b/drivers/misc/cxl/api.c

@@ -247,7 +247,9 @@ int cxl_start_context(struct cxl_context *ctx, u64 wed,
 	cxl_ctx_get();
 
 	if ((rc = cxl_ops->attach_process(ctx, kernel, wed, 0))) {
+		put_pid(ctx->glpid);
 		put_pid(ctx->pid);
+		ctx->glpid = ctx->pid = NULL;
 		cxl_adapter_context_put(ctx->afu->adapter);
 		cxl_ctx_put();
 		goto out;

diff --git a/drivers/misc/cxl/file.c b/drivers/misc/cxl/file.c
index d0b421f..77080cc 100644
--- a/drivers/misc/cxl/file.c
+++ b/drivers/misc/cxl/file.c

@@ -194,6 +194,16 @@ static long afu_ioctl_start_work(struct cxl_context *ctx,
 	ctx->mmio_err_ff = !!(work.flags & CXL_START_WORK_ERR_FF);
 
 	/*
+	 * Increment the mapped context count for adapter. This also checks
+	 * if adapter_context_lock is taken.
+	 */
+	rc = cxl_adapter_context_get(ctx->afu->adapter);
+	if (rc) {
+		afu_release_irqs(ctx, ctx);
+		goto out;
+	}
+
+	/*
 	 * We grab the PID here and not in the file open to allow for the case
 	 * where a process (master, some daemon, etc) has opened the chardev on
 	 * behalf of another process, so the AFU's mm gets bound to the process

@@ -205,15 +215,6 @@ static long afu_ioctl_start_work(struct cxl_context *ctx,
 	ctx->pid = get_task_pid(current, PIDTYPE_PID);
 	ctx->glpid = get_task_pid(current->group_leader, PIDTYPE_PID);
 
-	/*
-	 * Increment the mapped context count for adapter. This also checks
-	 * if adapter_context_lock is taken.
-	 */
-	rc = cxl_adapter_context_get(ctx->afu->adapter);
-	if (rc) {
-		afu_release_irqs(ctx, ctx);
-		goto out;
-	}
 
 	trace_cxl_attach(ctx, work.work_element_descriptor, work.num_interrupts, amr);

@@ -221,6 +222,9 @@ static long afu_ioctl_start_work(struct cxl_context *ctx,
 							amr))) {
 		afu_release_irqs(ctx, ctx);
 		cxl_adapter_context_put(ctx->afu->adapter);
+		put_pid(ctx->glpid);
+		put_pid(ctx->pid);
+		ctx->glpid = ctx->pid = NULL;
 		goto out;
 	}

-- 
2.7.4

Re: [PATCH] cxl: Fix leaking pid refs in some error paths

From: Andrew Donnellan <hidden>
Date: 2016-10-21 06:19:15

On 21/10/16 16:46, Vaibhav Jain wrote:

In some error paths in functions cxl_start_context and
afu_ioctl_start_work pid references to the current & group-leader tasks
can leak after they are taken. This patch fixes these error paths to
release these pid references before exiting the error path.

This patch is based on earlier patch "cxl: Prevent adapter reset
if an active context exists" at
https://patchwork.ozlabs.org/patch/682187/

Put this paragraph under the ---.

Fixes: 7b8ad495("cxl: Fix DSI misses when the context owning task exits")
Reported-by: Frederic Barrat <redacted>
Signed-off-by: Vaibhav Jain <redacted>

Reviewed-by: Andrew Donnellan <redacted>

quoted hunk

---
 drivers/misc/cxl/api.c  |  2 ++
 drivers/misc/cxl/file.c | 22 +++++++++++++---------
 2 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/drivers/misc/cxl/api.c b/drivers/misc/cxl/api.c
index af23d7d..2e5233b 100644
--- a/drivers/misc/cxl/api.c
+++ b/drivers/misc/cxl/api.c

@@ -247,7 +247,9 @@ int cxl_start_context(struct cxl_context *ctx, u64 wed,
 	cxl_ctx_get();

 	if ((rc = cxl_ops->attach_process(ctx, kernel, wed, 0))) {
+		put_pid(ctx->glpid);
 		put_pid(ctx->pid);
+		ctx->glpid = ctx->pid = NULL;

This is only needed if task != NULL, but I think it should be okay as 
long as ctx->[gl]pid is already NULL in that situation (which I haven't 
checked but I think that's the case).

-- 
Andrew Donnellan              OzLabs, ADL Canberra
andrew.donnellan@au1.ibm.com  IBM Australia Limited

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help