[PATCH] nvme: fix uninitialized-variable warning

[PATCH] nvme: fix uninitialized-variable warning

[Arnd Bergmann]

gcc -O3 adds a bogus warning in this file:

In file included from include/linux/compiler_types.h:68,
                 from <command-line>:
drivers/nvme/host/core.c: In function 'nvme_set_queue_count':
include/linux/compiler-gcc.h:71:45: error: 'res.u32' may be used uninitialized in this function [-Werror=maybe-uninitialized]
 #define __UNIQUE_ID(prefix) __PASTE(__PASTE(__UNIQUE_ID_, prefix), __COUNTER__)
                                             ^~~~~~~~~~~~
drivers/nvme/host/core.c:1167:20: note: 'res.u32' was declared here
  union nvme_result res;
                    ^~~

Slightly rearrange the code, enough to let gcc understand
better what is going on and not warn about it.

Fixes: mmtom ("init/Kconfig: enable -O3 for all arches")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 drivers/nvme/host/core.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 667f18f465be..6f0991e8c5cc 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -825,14 +825,15 @@ int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
 	int ret;
 
 	req = nvme_alloc_request(q, cmd, flags, qid);
-	if (IS_ERR(req))
-		return PTR_ERR(req);
+	ret = PTR_ERR_OR_ZERO(req);
+	if (ret < 0)
+		return ret;
 
 	req->timeout = timeout ? timeout : ADMIN_TIMEOUT;
 
 	if (buffer && bufflen) {
 		ret = blk_rq_map_kern(q, req, buffer, bufflen, GFP_KERNEL);
-		if (ret)
+		if (ret < 0)
 			goto out;
 	}
 
-- 
2.20.0


_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: [PATCH] nvme: fix uninitialized-variable warning

[Christoph Hellwig]

On Tue, Jan 07, 2020 at 10:42:08PM +0100, Arnd Bergmann wrote:
> Fixes: mmtom ("init/Kconfig: enable -O3 for all arches")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
>  drivers/nvme/host/core.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
> index 667f18f465be..6f0991e8c5cc 100644
> --- a/drivers/nvme/host/core.c
> +++ b/drivers/nvme/host/core.c
> @@ -825,14 +825,15 @@ int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
>  	int ret;
>  
>  	req = nvme_alloc_request(q, cmd, flags, qid);
> -	if (IS_ERR(req))
> -		return PTR_ERR(req);
> +	ret = PTR_ERR_OR_ZERO(req);
> +	if (ret < 0)
> +		return ret;

This one is just gross.  I think we'll need to find some other fix
that doesn't obsfucate the code as much.

>  
>  	req->timeout = timeout ? timeout : ADMIN_TIMEOUT;
>  
>  	if (buffer && bufflen) {
>  		ret = blk_rq_map_kern(q, req, buffer, bufflen, GFP_KERNEL);
> -		if (ret)
> +		if (ret < 0)

OTOH if this shuts up a compiler warning I'd be perfectly fine with it.

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: [PATCH] nvme: fix uninitialized-variable warning

[Arnd Bergmann]

On Thu, Jan 30, 2020 at 4:04 PM Christoph Hellwig <hch@infradead.org> wrote:
>
> On Tue, Jan 07, 2020 at 10:42:08PM +0100, Arnd Bergmann wrote:
> > Fixes: mmtom ("init/Kconfig: enable -O3 for all arches")
> > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> > ---
> >  drivers/nvme/host/core.c | 7 ++++---
> >  1 file changed, 4 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
> > index 667f18f465be..6f0991e8c5cc 100644
> > --- a/drivers/nvme/host/core.c
> > +++ b/drivers/nvme/host/core.c
> > @@ -825,14 +825,15 @@ int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
> >       int ret;
> >
> >       req = nvme_alloc_request(q, cmd, flags, qid);
> > -     if (IS_ERR(req))
> > -             return PTR_ERR(req);
> > +     ret = PTR_ERR_OR_ZERO(req);
> > +     if (ret < 0)
> > +             return ret;
>
> This one is just gross.  I think we'll need to find some other fix
> that doesn't obsfucate the code as much.

Initializing the nvme_result in nvme_features() would do it, as would
setting it in the error path in __nvme_submit_sync_cmd() -- either
way the compiler cannot be confused about whether it is initialized
later on.

       Arnd

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: [PATCH] nvme: fix uninitialized-variable warning

[Christoph Hellwig]

On Thu, Jan 30, 2020 at 04:36:48PM +0100, Arnd Bergmann wrote:
> > This one is just gross.  I think we'll need to find some other fix
> > that doesn't obsfucate the code as much.
> 
> Initializing the nvme_result in nvme_features() would do it, as would
> setting it in the error path in __nvme_submit_sync_cmd() -- either
> way the compiler cannot be confused about whether it is initialized
> later on.

Given that this is outside the hot path we can just zero the whole
structure before submitting the I/O.

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: [PATCH] nvme: fix uninitialized-variable warning

[Keith Busch]

On Thu, Jan 30, 2020 at 07:48:15AM -0800, Christoph Hellwig wrote:
> On Thu, Jan 30, 2020 at 04:36:48PM +0100, Arnd Bergmann wrote:
> > > This one is just gross.  I think we'll need to find some other fix
> > > that doesn't obsfucate the code as much.
> > 
> > Initializing the nvme_result in nvme_features() would do it, as would
> > setting it in the error path in __nvme_submit_sync_cmd() -- either
> > way the compiler cannot be confused about whether it is initialized
> > later on.
> 
> Given that this is outside the hot path we can just zero the whole
> structure before submitting the I/O.

I think this should be okay:

---
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 7f05deada7f4..4aeed750dab2 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1165,8 +1165,8 @@ static int nvme_identify_ns(struct nvme_ctrl *ctrl,
 static int nvme_features(struct nvme_ctrl *dev, u8 op, unsigned int fid,
 		unsigned int dword11, void *buffer, size_t buflen, u32 *result)
 {
+	union nvme_result res = { 0 };
 	struct nvme_command c;
-	union nvme_result res;
 	int ret;
 
 	memset(&c, 0, sizeof(c));
--

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: [PATCH] nvme: fix uninitialized-variable warning

[Christoph Hellwig]

On Fri, Feb 14, 2020 at 04:51:06AM +0900, Keith Busch wrote:
> On Thu, Jan 30, 2020 at 07:48:15AM -0800, Christoph Hellwig wrote:
> > On Thu, Jan 30, 2020 at 04:36:48PM +0100, Arnd Bergmann wrote:
> > > > This one is just gross.  I think we'll need to find some other fix
> > > > that doesn't obsfucate the code as much.
> > > 
> > > Initializing the nvme_result in nvme_features() would do it, as would
> > > setting it in the error path in __nvme_submit_sync_cmd() -- either
> > > way the compiler cannot be confused about whether it is initialized
> > > later on.
> > 
> > Given that this is outside the hot path we can just zero the whole
> > structure before submitting the I/O.
> 
> I think this should be okay:

This looks good.  Can you send a formal patch?

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

[PATCH] nvme: Fix uninitialized-variable warning

[Keith Busch]

gcc may detect a false positive on nvme using an unintialized variable
if setting features fails. Since this is not a fast path, explicitly
initialize this variable to suppress the warning.

Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
---
 drivers/nvme/host/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index ada59df642d2..a4d8c90ee7cc 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1165,8 +1165,8 @@ static int nvme_identify_ns(struct nvme_ctrl *ctrl,
 static int nvme_features(struct nvme_ctrl *dev, u8 op, unsigned int fid,
 		unsigned int dword11, void *buffer, size_t buflen, u32 *result)
 {
+	union nvme_result res = { 0 };
 	struct nvme_command c;
-	union nvme_result res;
 	int ret;
 
 	memset(&c, 0, sizeof(c));
-- 
2.21.0


_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: [PATCH] nvme: Fix uninitialized-variable warning

[Christoph Hellwig]

On Thu, Feb 20, 2020 at 01:19:50AM +0900, Keith Busch wrote:
> gcc may detect a false positive on nvme using an unintialized variable
> if setting features fails. Since this is not a fast path, explicitly
> initialize this variable to suppress the warning.
> 
> Reported-by: Arnd Bergmann <arnd@arndb.de>
> Signed-off-by: Keith Busch <kbusch@kernel.org>

Looks good,

Reviewed-by: Christoph Hellwig <hch@lst.de>

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: [PATCH] nvme: Fix uninitialized-variable warning

[Sagi Grimberg]

Reviewed-by: Sagi Grimberg <sagi@grimberg.me>

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme