[Qemu-devel] [PATCH 0/2] virtio-blk: clean up around g_free() and inline a few functions

[Qemu-devel] [PATCH 0/2] virtio-blk: clean up around g_free() and inline a few functions

[Fam Zheng]

Fam Zheng (2):
  virtio-blk: Remove useless condition around g_free()
  virtio-blk: Inline request init, complete and free functions

 hw/block/virtio-blk.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

-- 
2.9.3

[Qemu-devel] [PATCH 1/2] virtio-blk: Remove useless condition around g_free()

[Fam Zheng]

Laszlo spotted and studied this wasteful "if". He pointed out:

The original virtio_blk_free_request needed an "if" as it accesses one
field, since 671ec3f05655 ("virtio-blk: Convert VirtIOBlockReq.elem to
pointer", 2014-06-11); later on in f897bf751fbd ("virtio-blk: embed
VirtQueueElement in VirtIOBlockReq", 2014-07-09) the field became
embedded, so the "if" became unnecessary (at which point we were using
g_slice_free(), but it is the same.

Now drop it.

Reported-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
---
 hw/block/virtio-blk.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
index 702eda8..2858c31 100644
--- a/hw/block/virtio-blk.c
+++ b/hw/block/virtio-blk.c
@@ -42,9 +42,7 @@ static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
 
 static void virtio_blk_free_request(VirtIOBlockReq *req)
 {
-    if (req) {
-        g_free(req);
-    }
+    g_free(req);
 }
 
 static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
-- 
2.9.3

[Qemu-devel] [PATCH 2/2] virtio-blk: Inline request init, complete and free functions

[Fam Zheng]

These are used in each request handling, inline them.

Signed-off-by: Fam Zheng <famz@redhat.com>
---
 hw/block/virtio-blk.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
index 2858c31..1da9570 100644
--- a/hw/block/virtio-blk.c
+++ b/hw/block/virtio-blk.c
@@ -29,8 +29,8 @@
 #include "hw/virtio/virtio-bus.h"
 #include "hw/virtio/virtio-access.h"
 
-static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
-                                    VirtIOBlockReq *req)
+static inline void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
+                                           VirtIOBlockReq *req)
 {
     req->dev = s;
     req->vq = vq;
@@ -40,12 +40,13 @@ static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
     req->mr_next = NULL;
 }
 
-static void virtio_blk_free_request(VirtIOBlockReq *req)
+static inline void virtio_blk_free_request(VirtIOBlockReq *req)
 {
     g_free(req);
 }
 
-static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
+static inline void virtio_blk_req_complete(VirtIOBlockReq *req,
+                                           unsigned char status)
 {
     VirtIOBlock *s = req->dev;
     VirtIODevice *vdev = VIRTIO_DEVICE(s);
-- 
2.9.3

Re: [Qemu-devel] [PATCH 1/2] virtio-blk: Remove useless condition around g_free()

[Laszlo Ersek]

On 02/07/17 14:27, Fam Zheng wrote:
> Laszlo spotted and studied this wasteful "if". He pointed out:
> 
> The original virtio_blk_free_request needed an "if" as it accesses one
> field, since 671ec3f05655 ("virtio-blk: Convert VirtIOBlockReq.elem to
> pointer", 2014-06-11); later on in f897bf751fbd ("virtio-blk: embed
> VirtQueueElement in VirtIOBlockReq", 2014-07-09) the field became
> embedded, so the "if" became unnecessary (at which point we were using
> g_slice_free(), but it is the same.
> 
> Now drop it.
> 
> Reported-by: Laszlo Ersek <lersek@redhat.com>
> Signed-off-by: Fam Zheng <famz@redhat.com>
> ---
>  hw/block/virtio-blk.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
> index 702eda8..2858c31 100644
> --- a/hw/block/virtio-blk.c
> +++ b/hw/block/virtio-blk.c
> @@ -42,9 +42,7 @@ static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
>  
>  static void virtio_blk_free_request(VirtIOBlockReq *req)
>  {
> -    if (req) {
> -        g_free(req);
> -    }
> +    g_free(req);
>  }
>  
>  static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
> 

Reviewed-by: Laszlo Ersek <lersek@redhat.com>

Thanks!
Laszlo

Re: [Qemu-devel] [PATCH 2/2] virtio-blk: Inline request init, complete and free functions

[Laszlo Ersek]

On 02/07/17 14:27, Fam Zheng wrote:
> These are used in each request handling, inline them.
> 
> Signed-off-by: Fam Zheng <famz@redhat.com>
> ---
>  hw/block/virtio-blk.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
> index 2858c31..1da9570 100644
> --- a/hw/block/virtio-blk.c
> +++ b/hw/block/virtio-blk.c
> @@ -29,8 +29,8 @@
>  #include "hw/virtio/virtio-bus.h"
>  #include "hw/virtio/virtio-access.h"
>  
> -static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
> -                                    VirtIOBlockReq *req)
> +static inline void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
> +                                           VirtIOBlockReq *req)
>  {
>      req->dev = s;
>      req->vq = vq;
> @@ -40,12 +40,13 @@ static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
>      req->mr_next = NULL;
>  }
>  
> -static void virtio_blk_free_request(VirtIOBlockReq *req)
> +static inline void virtio_blk_free_request(VirtIOBlockReq *req)
>  {
>      g_free(req);
>  }
>  
> -static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
> +static inline void virtio_blk_req_complete(VirtIOBlockReq *req,
> +                                           unsigned char status)
>  {
>      VirtIOBlock *s = req->dev;
>      VirtIODevice *vdev = VIRTIO_DEVICE(s);
> 

Hm, virtio_blk_req_complete() looks a bit too "meaty" and seems to be
called from a little too many places for me to feel convenient about
inlining it. I guess I'd leave it to the compiler to optimize the
function call. Does the explicit hint offer a noticeable perf improvement?

Inlining virtio_blk_free_request() looks reasonable.

virtio_blk_init_request() looks okay too.

Other reviewers should feel free to override my concerns :) My view on
this is distant.

Thanks
Laszlo

Re: [Qemu-devel] [Qemu-block] [PATCH 1/2] virtio-blk: Remove useless condition around g_free()

[Stefan Hajnoczi]

[-- Attachment #1: Type: text/plain, Size: 797 bytes --]

On Tue, Feb 07, 2017 at 09:27:22PM +0800, Fam Zheng wrote:
> Laszlo spotted and studied this wasteful "if". He pointed out:
> 
> The original virtio_blk_free_request needed an "if" as it accesses one
> field, since 671ec3f05655 ("virtio-blk: Convert VirtIOBlockReq.elem to
> pointer", 2014-06-11); later on in f897bf751fbd ("virtio-blk: embed
> VirtQueueElement in VirtIOBlockReq", 2014-07-09) the field became
> embedded, so the "if" became unnecessary (at which point we were using
> g_slice_free(), but it is the same.
> 
> Now drop it.
> 
> Reported-by: Laszlo Ersek <lersek@redhat.com>
> Signed-off-by: Fam Zheng <famz@redhat.com>
> ---
>  hw/block/virtio-blk.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [Qemu-devel] [PATCH 2/2] virtio-blk: Inline request init, complete and free functions

[Stefan Hajnoczi]

[-- Attachment #1: Type: text/plain, Size: 2231 bytes --]

On Tue, Feb 07, 2017 at 02:52:38PM +0100, Laszlo Ersek wrote:
> On 02/07/17 14:27, Fam Zheng wrote:
> > These are used in each request handling, inline them.
> > 
> > Signed-off-by: Fam Zheng <famz@redhat.com>
> > ---
> >  hw/block/virtio-blk.c | 9 +++++----
> >  1 file changed, 5 insertions(+), 4 deletions(-)
> > 
> > diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
> > index 2858c31..1da9570 100644
> > --- a/hw/block/virtio-blk.c
> > +++ b/hw/block/virtio-blk.c
> > @@ -29,8 +29,8 @@
> >  #include "hw/virtio/virtio-bus.h"
> >  #include "hw/virtio/virtio-access.h"
> >  
> > -static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
> > -                                    VirtIOBlockReq *req)
> > +static inline void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
> > +                                           VirtIOBlockReq *req)
> >  {
> >      req->dev = s;
> >      req->vq = vq;
> > @@ -40,12 +40,13 @@ static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
> >      req->mr_next = NULL;
> >  }
> >  
> > -static void virtio_blk_free_request(VirtIOBlockReq *req)
> > +static inline void virtio_blk_free_request(VirtIOBlockReq *req)
> >  {
> >      g_free(req);
> >  }
> >  
> > -static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
> > +static inline void virtio_blk_req_complete(VirtIOBlockReq *req,
> > +                                           unsigned char status)
> >  {
> >      VirtIOBlock *s = req->dev;
> >      VirtIODevice *vdev = VIRTIO_DEVICE(s);
> > 
> 
> Hm, virtio_blk_req_complete() looks a bit too "meaty" and seems to be
> called from a little too many places for me to feel convenient about
> inlining it. I guess I'd leave it to the compiler to optimize the
> function call. Does the explicit hint offer a noticeable perf improvement?
> 
> Inlining virtio_blk_free_request() looks reasonable.
> 
> virtio_blk_init_request() looks okay too.
> 
> Other reviewers should feel free to override my concerns :) My view on
> this is distant.

I'm not a big fan of manually inlining functions.  Let the compiler
decide whether these static functions should be inlined.

Stefan

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [Qemu-devel] [PATCH 2/2] virtio-blk: Inline request init, complete and free functions

[Fam Zheng]

On Mon, 02/13 14:28, Stefan Hajnoczi wrote:
> On Tue, Feb 07, 2017 at 02:52:38PM +0100, Laszlo Ersek wrote:
> > On 02/07/17 14:27, Fam Zheng wrote:
> > > These are used in each request handling, inline them.
> > > 
> > > Signed-off-by: Fam Zheng <famz@redhat.com>
> > > ---
> > >  hw/block/virtio-blk.c | 9 +++++----
> > >  1 file changed, 5 insertions(+), 4 deletions(-)
> > > 
> > > diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
> > > index 2858c31..1da9570 100644
> > > --- a/hw/block/virtio-blk.c
> > > +++ b/hw/block/virtio-blk.c
> > > @@ -29,8 +29,8 @@
> > >  #include "hw/virtio/virtio-bus.h"
> > >  #include "hw/virtio/virtio-access.h"
> > >  
> > > -static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
> > > -                                    VirtIOBlockReq *req)
> > > +static inline void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
> > > +                                           VirtIOBlockReq *req)
> > >  {
> > >      req->dev = s;
> > >      req->vq = vq;
> > > @@ -40,12 +40,13 @@ static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
> > >      req->mr_next = NULL;
> > >  }
> > >  
> > > -static void virtio_blk_free_request(VirtIOBlockReq *req)
> > > +static inline void virtio_blk_free_request(VirtIOBlockReq *req)
> > >  {
> > >      g_free(req);
> > >  }
> > >  
> > > -static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
> > > +static inline void virtio_blk_req_complete(VirtIOBlockReq *req,
> > > +                                           unsigned char status)
> > >  {
> > >      VirtIOBlock *s = req->dev;
> > >      VirtIODevice *vdev = VIRTIO_DEVICE(s);
> > > 
> > 
> > Hm, virtio_blk_req_complete() looks a bit too "meaty" and seems to be
> > called from a little too many places for me to feel convenient about
> > inlining it. I guess I'd leave it to the compiler to optimize the
> > function call. Does the explicit hint offer a noticeable perf improvement?
> > 
> > Inlining virtio_blk_free_request() looks reasonable.
> > 
> > virtio_blk_init_request() looks okay too.
> > 
> > Other reviewers should feel free to override my concerns :) My view on
> > this is distant.
> 
> I'm not a big fan of manually inlining functions.  Let the compiler
> decide whether these static functions should be inlined.
> 

Fair enough, let's drop this one.

Fam

Re: [Qemu-devel] [PATCH 1/2] virtio-blk: Remove useless condition around g_free()

[Fam Zheng]

On Tue, 02/07 21:27, Fam Zheng wrote:
> Laszlo spotted and studied this wasteful "if". He pointed out:
> 
> The original virtio_blk_free_request needed an "if" as it accesses one
> field, since 671ec3f05655 ("virtio-blk: Convert VirtIOBlockReq.elem to
> pointer", 2014-06-11); later on in f897bf751fbd ("virtio-blk: embed
> VirtQueueElement in VirtIOBlockReq", 2014-07-09) the field became
> embedded, so the "if" became unnecessary (at which point we were using
> g_slice_free(), but it is the same.
> 
> Now drop it.
> 
> Reported-by: Laszlo Ersek <lersek@redhat.com>
> Signed-off-by: Fam Zheng <famz@redhat.com>
> ---
>  hw/block/virtio-blk.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
> index 702eda8..2858c31 100644
> --- a/hw/block/virtio-blk.c
> +++ b/hw/block/virtio-blk.c
> @@ -42,9 +42,7 @@ static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
>  
>  static void virtio_blk_free_request(VirtIOBlockReq *req)
>  {
> -    if (req) {
> -        g_free(req);
> -    }
> +    g_free(req);
>  }
>  
>  static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
> -- 
> 2.9.3
> 
> 

Cc: qemu-trivial@nongnu.org

(Let's drop 2/2 and perhaps merge this via trivial)

Fam

Re: [Qemu-devel] [PATCH 1/2] virtio-blk: Remove useless condition around g_free()

[Michael Tokarev]

07.02.2017 16:27, Fam Zheng wrote:
> Laszlo spotted and studied this wasteful "if". He pointed out:
> 
> The original virtio_blk_free_request needed an "if" as it accesses one
> field, since 671ec3f05655 ("virtio-blk: Convert VirtIOBlockReq.elem to
> pointer", 2014-06-11); later on in f897bf751fbd ("virtio-blk: embed
> VirtQueueElement in VirtIOBlockReq", 2014-07-09) the field became
> embedded, so the "if" became unnecessary (at which point we were using
> g_slice_free(), but it is the same.
> 
> Now drop it.

Applied to -trivial, thanks!

/mjt