[PATCH] misc: fastrpc: fix potential fastrpc_invoke_ctx leak

[PATCH] misc: fastrpc: fix potential fastrpc_invoke_ctx leak

[Srinivas Kandagatla]

fastrpc_invoke_ctx can have refcount of 2 in error path where
rpmsg_send() fails to send invoke message. decrement the refcount
properly in the error path to fix this leak.

This also fixes below static checker warning:

drivers/misc/fastrpc.c:990 fastrpc_internal_invoke()
warn: 'ctx->refcount.refcount.ref.counter' not decremented on lines: 990.

Fixes: c68cfb718c8f ("misc: fastrpc: Add support for context")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
---
 drivers/misc/fastrpc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/misc/fastrpc.c b/drivers/misc/fastrpc.c
index 9065d3e71ff7..07065728e39f 100644
--- a/drivers/misc/fastrpc.c
+++ b/drivers/misc/fastrpc.c
@@ -949,8 +949,10 @@ static int fastrpc_internal_invoke(struct fastrpc_user *fl,  u32 kernel,
 	dma_wmb();
 	/* Send invoke buffer to remote dsp */
 	err = fastrpc_invoke_send(fl->sctx, ctx, kernel, handle);
-	if (err)
+	if (err) {
+		fastrpc_context_put(ctx);
 		goto bail;
+	}
 
 	if (kernel) {
 		if (!wait_for_completion_timeout(&ctx->work, 10 * HZ))
-- 
2.21.0

Re: [PATCH] misc: fastrpc: fix potential fastrpc_invoke_ctx leak

[Bjorn Andersson]

On Mon 11 May 09:29 PDT 2020, Srinivas Kandagatla wrote:

> fastrpc_invoke_ctx can have refcount of 2 in error path where
> rpmsg_send() fails to send invoke message. decrement the refcount
> properly in the error path to fix this leak.
> 
> This also fixes below static checker warning:
> 
> drivers/misc/fastrpc.c:990 fastrpc_internal_invoke()
> warn: 'ctx->refcount.refcount.ref.counter' not decremented on lines: 990.
> 
> Fixes: c68cfb718c8f ("misc: fastrpc: Add support for context")
> Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
> Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
> ---
>  drivers/misc/fastrpc.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/misc/fastrpc.c b/drivers/misc/fastrpc.c
> index 9065d3e71ff7..07065728e39f 100644
> --- a/drivers/misc/fastrpc.c
> +++ b/drivers/misc/fastrpc.c
> @@ -949,8 +949,10 @@ static int fastrpc_internal_invoke(struct fastrpc_user *fl,  u32 kernel,
>  	dma_wmb();
>  	/* Send invoke buffer to remote dsp */
>  	err = fastrpc_invoke_send(fl->sctx, ctx, kernel, handle);
> -	if (err)
> +	if (err) {
> +		fastrpc_context_put(ctx);

So we refcount ctx once for the invoke function and once between send
and callback. And this fastrpc_context_put() would counter the fact that
rpmsg_send() failed, so we will not get the "remote's" put().

I think that if you moved this call inside fastrpc_invoke_send() it's
relationship to the failing rpmsg_send() would be obvious.

Regards,
Bjorn

>  		goto bail;
> +	}
>  
>  	if (kernel) {
>  		if (!wait_for_completion_timeout(&ctx->work, 10 * HZ))
> -- 
> 2.21.0
> 

Re: [PATCH] misc: fastrpc: fix potential fastrpc_invoke_ctx leak

[Srinivas Kandagatla]

On 12/05/2020 02:18, Bjorn Andersson wrote:
> On Mon 11 May 09:29 PDT 2020, Srinivas Kandagatla wrote:
> 
>> fastrpc_invoke_ctx can have refcount of 2 in error path where
>> rpmsg_send() fails to send invoke message. decrement the refcount
>> properly in the error path to fix this leak.
>>
>> This also fixes below static checker warning:
>>
>> drivers/misc/fastrpc.c:990 fastrpc_internal_invoke()
>> warn: 'ctx->refcount.refcount.ref.counter' not decremented on lines: 990.
>>
>> Fixes: c68cfb718c8f ("misc: fastrpc: Add support for context")
>> Reported-by: Dan Carpenter<dan.carpenter@oracle.com>
>> Signed-off-by: Srinivas Kandagatla<srinivas.kandagatla@linaro.org>
>> ---
>>   drivers/misc/fastrpc.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/misc/fastrpc.c b/drivers/misc/fastrpc.c
>> index 9065d3e71ff7..07065728e39f 100644
>> --- a/drivers/misc/fastrpc.c
>> +++ b/drivers/misc/fastrpc.c
>> @@ -949,8 +949,10 @@ static int fastrpc_internal_invoke(struct fastrpc_user *fl,  u32 kernel,
>>   	dma_wmb();
>>   	/* Send invoke buffer to remote dsp */
>>   	err = fastrpc_invoke_send(fl->sctx, ctx, kernel, handle);
>> -	if (err)
>> +	if (err) {
>> +		fastrpc_context_put(ctx);
> So we refcount ctx once for the invoke function and once between send
> and callback. And this fastrpc_context_put() would counter the fact that
> rpmsg_send() failed, so we will not get the "remote's" put().
> 
> I think that if you moved this call inside fastrpc_invoke_send() it's
> relationship to the failing rpmsg_send() would be obvious.

That works as well, and sounds much cleaner!


thanks,
srini

> 
> Regards,