* [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
@ 2014-04-08 20:42 Matei Oprea
2014-04-09 0:04 ` Joe Perches
2014-04-09 10:48 ` Boaz Harrosh
0 siblings, 2 replies; 12+ messages in thread
From: Matei Oprea @ 2014-04-08 20:42 UTC (permalink / raw)
To: jbottomley; +Cc: linux-scsi, linux-kernel, Matei Oprea, ROSEdu Kernel Community
It's easier to use kcalloc for allocating arrays. While at it
also remove useless casting value.
Signed-off-by: Matei Oprea <eu@opreamatei.ro>
Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
---
drivers/scsi/lpfc/lpfc_init.c | 14 ++++++++------
1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index 635eeb3..1528f34 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -4742,12 +4742,14 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
phba->cfg_sg_seg_cnt = LPFC_DEFAULT_MENLO_SG_SEG_CNT;
}
- if (!phba->sli.ring)
- phba->sli.ring = (struct lpfc_sli_ring *)
- kzalloc(LPFC_SLI3_MAX_RING *
- sizeof(struct lpfc_sli_ring), GFP_KERNEL);
- if (!phba->sli.ring)
- return -ENOMEM;
+ if (unlikely(!phba->sli.ring)) {
+ phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
+ sizeof(struct lpfc_sli_ring),
+ GFP_KERNEL);
+
+ if (!phba->sli.ring)
+ return -ENOMEM;
+ }
/*
* Since lpfc_sg_seg_cnt is module parameter, the sg_dma_buf_size
--
1.9.1
^ permalink raw reply related [flat|nested] 12+ messages in thread* Re: [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
2014-04-08 20:42 [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory Matei Oprea
@ 2014-04-09 0:04 ` Joe Perches
2014-04-09 10:48 ` Boaz Harrosh
1 sibling, 0 replies; 12+ messages in thread
From: Joe Perches @ 2014-04-09 0:04 UTC (permalink / raw)
To: Matei Oprea; +Cc: jbottomley, linux-scsi, linux-kernel, ROSEdu Kernel Community
On Tue, 2014-04-08 at 23:42 +0300, Matei Oprea wrote:
> It's easier to use kcalloc for allocating arrays. While at it
> also remove useless casting value.
[]
> diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
[]
> @@ -4742,12 +4742,14 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
[]
> - if (!phba->sli.ring)
[]
> + if (unlikely(!phba->sli.ring)) {
Why add the unlikely?
Glancing at the code it seems more like that this
should be likely than unlikely.
> + phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
> + sizeof(struct lpfc_sli_ring),
> + GFP_KERNEL);
It'd be nicer looking if aligned better
phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
sizeof(struct lpfc_sli_ring),
GFP_KERNEL);
^ permalink raw reply [flat|nested] 12+ messages in thread* Re: [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
2014-04-08 20:42 [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory Matei Oprea
2014-04-09 0:04 ` Joe Perches
@ 2014-04-09 10:48 ` Boaz Harrosh
1 sibling, 0 replies; 12+ messages in thread
From: Boaz Harrosh @ 2014-04-09 10:48 UTC (permalink / raw)
To: Matei Oprea, jbottomley; +Cc: linux-scsi, linux-kernel, ROSEdu Kernel Community
On 04/08/2014 11:42 PM, Matei Oprea wrote:
> It's easier to use kcalloc for allocating arrays. While at it
> also remove useless casting value.
>
> Signed-off-by: Matei Oprea <eu@opreamatei.ro>
> Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
> ---
> drivers/scsi/lpfc/lpfc_init.c | 14 ++++++++------
> 1 file changed, 8 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
> index 635eeb3..1528f34 100644
> --- a/drivers/scsi/lpfc/lpfc_init.c
> +++ b/drivers/scsi/lpfc/lpfc_init.c
> @@ -4742,12 +4742,14 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
> phba->cfg_sg_seg_cnt = LPFC_DEFAULT_MENLO_SG_SEG_CNT;
> }
>
> - if (!phba->sli.ring)
> - phba->sli.ring = (struct lpfc_sli_ring *)
> - kzalloc(LPFC_SLI3_MAX_RING *
> - sizeof(struct lpfc_sli_ring), GFP_KERNEL);
> - if (!phba->sli.ring)
> - return -ENOMEM;
> + if (unlikely(!phba->sli.ring)) {
As Joe Perches said this one is not unlikely, it might or it might not
depending on state.
> + phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
> + sizeof(struct lpfc_sli_ring),
> + GFP_KERNEL);
> +
> + if (!phba->sli.ring)
but this one is unlikely. This is the error path so we want
the hot path to be decoded by the CPU.
> + return -ENOMEM;
> + }
>
> /*
> * Since lpfc_sg_seg_cnt is module parameter, the sg_dma_buf_size
>
Cheers
Boaz
^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
@ 2014-04-11 9:21 Matei Oprea
0 siblings, 0 replies; 12+ messages in thread
From: Matei Oprea @ 2014-04-11 9:21 UTC (permalink / raw)
To: jbottomley; +Cc: linux-scsi, linux-kernel, Matei Oprea, ROSEdu Kernel Community
It's easier to use kcalloc for allocating arrays. While at it
also remove useless casting value.
Signed-off-by: Matei Oprea <eu@opreamatei.ro>
Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
---
drivers/scsi/lpfc/lpfc_init.c | 13 +++++++------
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index 635eeb3..7e58c22 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -4742,12 +4742,13 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
phba->cfg_sg_seg_cnt = LPFC_DEFAULT_MENLO_SG_SEG_CNT;
}
- if (!phba->sli.ring)
- phba->sli.ring = (struct lpfc_sli_ring *)
- kzalloc(LPFC_SLI3_MAX_RING *
- sizeof(struct lpfc_sli_ring), GFP_KERNEL);
- if (!phba->sli.ring)
- return -ENOMEM;
+ if (!phba->sli.ring) {
+ phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
+ sizeof(struct lpfc_sli_ring),
+ GFP_KERNEL);
+ if (!phba->sli.ring)
+ return -ENOMEM;
+ }
/*
* Since lpfc_sg_seg_cnt is module parameter, the sg_dma_buf_size
--
1.9.1
^ permalink raw reply related [flat|nested] 12+ messages in thread* [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
@ 2014-04-08 20:18 Matei Oprea
2014-04-08 20:20 ` Joe Perches
0 siblings, 1 reply; 12+ messages in thread
From: Matei Oprea @ 2014-04-08 20:18 UTC (permalink / raw)
To: jbottomley; +Cc: linux-scsi, linux-kernel, Matei Oprea, ROSEdu Kernel Community
It's easier to use kcalloc for allocating arrays. While at it
also remove useless casting value.
Signed-off-by: Matei Oprea <eu@opreamatei.ro>
Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
---
drivers/scsi/lpfc/lpfc_init.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index e286909..2b1a786 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -4742,12 +4742,14 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
phba->cfg_sg_seg_cnt = LPFC_DEFAULT_MENLO_SG_SEG_CNT;
}
- if (!phba->sli.ring) {
+ if (unlikely(!phba->sli.ring)) {
phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
sizeof(struct lpfc_sli_ring),
GFP_KERNEL);
- return -ENOMEM;
}
+
+ if (!phba->sli.ring)
+ return -ENOMEM;
/*
* Since lpfc_sg_seg_cnt is module parameter, the sg_dma_buf_size
--
1.9.1
^ permalink raw reply related [flat|nested] 12+ messages in thread* Re: [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
2014-04-08 20:18 Matei Oprea
@ 2014-04-08 20:20 ` Joe Perches
2014-04-08 20:21 ` Matei Oprea
0 siblings, 1 reply; 12+ messages in thread
From: Joe Perches @ 2014-04-08 20:20 UTC (permalink / raw)
To: Matei Oprea; +Cc: jbottomley, linux-scsi, linux-kernel, ROSEdu Kernel Community
On Tue, 2014-04-08 at 23:18 +0300, Matei Oprea wrote:
> It's easier to use kcalloc for allocating arrays. While at it
> also remove useless casting value.
Don't send a patch against your bad patch.
Send a new patch with just the appropriate
changes.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
2014-04-08 20:20 ` Joe Perches
@ 2014-04-08 20:21 ` Matei Oprea
0 siblings, 0 replies; 12+ messages in thread
From: Matei Oprea @ 2014-04-08 20:21 UTC (permalink / raw)
To: Joe Perches; +Cc: jbottomley, linux-scsi, linux-kernel, ROSEdu Kernel Community
Ok. Sorry.
ᐧ
On Tue, Apr 8, 2014 at 11:20 PM, Joe Perches <joe@perches.com> wrote:
> On Tue, 2014-04-08 at 23:18 +0300, Matei Oprea wrote:
>> It's easier to use kcalloc for allocating arrays. While at it
>> also remove useless casting value.
>
> Don't send a patch against your bad patch.
> Send a new patch with just the appropriate
> changes.
>
>
--
Oprea Matei
^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
@ 2014-04-08 19:44 Matei Oprea
2014-04-08 19:48 ` Joe Perches
0 siblings, 1 reply; 12+ messages in thread
From: Matei Oprea @ 2014-04-08 19:44 UTC (permalink / raw)
To: jbottomley; +Cc: linux-scsi, linux-kernel, Matei Oprea, ROSEdu Kernel Community
It's easier to use kcalloc for allocating arrays. While at it
also remove useless casting value.
Signed-off-by: Matei Oprea <eu@opreamatei.ro>
Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
---
drivers/scsi/lpfc/lpfc_init.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index 635eeb3..e286909 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -4742,12 +4742,12 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
phba->cfg_sg_seg_cnt = LPFC_DEFAULT_MENLO_SG_SEG_CNT;
}
- if (!phba->sli.ring)
- phba->sli.ring = (struct lpfc_sli_ring *)
- kzalloc(LPFC_SLI3_MAX_RING *
- sizeof(struct lpfc_sli_ring), GFP_KERNEL);
- if (!phba->sli.ring)
+ if (!phba->sli.ring) {
+ phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
+ sizeof(struct lpfc_sli_ring),
+ GFP_KERNEL);
return -ENOMEM;
+ }
/*
* Since lpfc_sg_seg_cnt is module parameter, the sg_dma_buf_size
--
1.9.1
^ permalink raw reply related [flat|nested] 12+ messages in thread* Re: [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
2014-04-08 19:44 Matei Oprea
@ 2014-04-08 19:48 ` Joe Perches
0 siblings, 0 replies; 12+ messages in thread
From: Joe Perches @ 2014-04-08 19:48 UTC (permalink / raw)
To: Matei Oprea; +Cc: jbottomley, linux-scsi, linux-kernel, ROSEdu Kernel Community
On Tue, 2014-04-08 at 22:44 +0300, Matei Oprea wrote:
> It's easier to use kcalloc for allocating arrays. While at it
> also remove useless casting value.
>
> Signed-off-by: Matei Oprea <eu@opreamatei.ro>
> Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
> ---
> drivers/scsi/lpfc/lpfc_init.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
> index 635eeb3..e286909 100644
> --- a/drivers/scsi/lpfc/lpfc_init.c
> +++ b/drivers/scsi/lpfc/lpfc_init.c
> @@ -4742,12 +4742,12 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
> phba->cfg_sg_seg_cnt = LPFC_DEFAULT_MENLO_SG_SEG_CNT;
> }
>
> - if (!phba->sli.ring)
> - phba->sli.ring = (struct lpfc_sli_ring *)
> - kzalloc(LPFC_SLI3_MAX_RING *
> - sizeof(struct lpfc_sli_ring), GFP_KERNEL);
> - if (!phba->sli.ring)
> + if (!phba->sli.ring) {
> + phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
> + sizeof(struct lpfc_sli_ring),
> + GFP_KERNEL);
> return -ENOMEM;
> + }
You seem to be returning -ENOMEM after kcalloc.
You need to put back the
if (!phba->sli.ring)
return -ENOMEM;
after the alloc.
^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
@ 2014-03-18 20:51 Matei Oprea
2014-04-07 12:37 ` Matei Oprea
2014-04-08 6:36 ` Boaz Harrosh
0 siblings, 2 replies; 12+ messages in thread
From: Matei Oprea @ 2014-03-18 20:51 UTC (permalink / raw)
To: jbottomley
Cc: james.smart, linux-scsi, linux-kernel, Matei Oprea,
ROSEdu Kernel Community
It's easier to use kcalloc for allocating arrays. While at it
also remove useless casting value.
Signed-off-by: Matei Oprea <eu@opreamatei.ro>
Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
---
drivers/scsi/lpfc/lpfc_init.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index 68c94cc..0a51ca5 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -4731,9 +4731,9 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
}
if (!phba->sli.ring)
- phba->sli.ring = (struct lpfc_sli_ring *)
- kzalloc(LPFC_SLI3_MAX_RING *
+ phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
sizeof(struct lpfc_sli_ring), GFP_KERNEL);
+
if (!phba->sli.ring)
return -ENOMEM;
--
1.7.9.5
^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
2014-03-18 20:51 Matei Oprea
@ 2014-04-07 12:37 ` Matei Oprea
2014-04-08 6:36 ` Boaz Harrosh
1 sibling, 0 replies; 12+ messages in thread
From: Matei Oprea @ 2014-04-07 12:37 UTC (permalink / raw)
To: jbottomley
Cc: james.smart, linux-scsi, linux-kernel, Matei Oprea,
ROSEdu Kernel Community
Any comments on this ?
ᐧ
On Tue, Mar 18, 2014 at 10:51 PM, Matei Oprea <eu@opreamatei.ro> wrote:
> It's easier to use kcalloc for allocating arrays. While at it
> also remove useless casting value.
>
> Signed-off-by: Matei Oprea <eu@opreamatei.ro>
> Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
> ---
> drivers/scsi/lpfc/lpfc_init.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
> index 68c94cc..0a51ca5 100644
> --- a/drivers/scsi/lpfc/lpfc_init.c
> +++ b/drivers/scsi/lpfc/lpfc_init.c
> @@ -4731,9 +4731,9 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
> }
>
> if (!phba->sli.ring)
> - phba->sli.ring = (struct lpfc_sli_ring *)
> - kzalloc(LPFC_SLI3_MAX_RING *
> + phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
> sizeof(struct lpfc_sli_ring), GFP_KERNEL);
> +
> if (!phba->sli.ring)
> return -ENOMEM;
>
> --
> 1.7.9.5
>
--
Oprea Matei
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory
2014-03-18 20:51 Matei Oprea
2014-04-07 12:37 ` Matei Oprea
@ 2014-04-08 6:36 ` Boaz Harrosh
1 sibling, 0 replies; 12+ messages in thread
From: Boaz Harrosh @ 2014-04-08 6:36 UTC (permalink / raw)
To: Matei Oprea, jbottomley
Cc: james.smart, linux-scsi, linux-kernel, ROSEdu Kernel Community
On 03/18/2014 10:51 PM, Matei Oprea wrote:
> It's easier to use kcalloc for allocating arrays. While at it
> also remove useless casting value.
>
> Signed-off-by: Matei Oprea <eu@opreamatei.ro>
> Cc: ROSEdu Kernel Community <firefly@lists.rosedu.org>
> ---
> drivers/scsi/lpfc/lpfc_init.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
> index 68c94cc..0a51ca5 100644
> --- a/drivers/scsi/lpfc/lpfc_init.c
> +++ b/drivers/scsi/lpfc/lpfc_init.c
> @@ -4731,9 +4731,9 @@ lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
> }
>
> if (!phba->sli.ring)
> - phba->sli.ring = (struct lpfc_sli_ring *)
> - kzalloc(LPFC_SLI3_MAX_RING *
> + phba->sli.ring = kcalloc(LPFC_SLI3_MAX_RING,
> sizeof(struct lpfc_sli_ring), GFP_KERNEL);
> +
Just a nit please put this recheck inside the if(){} so in the hot path
when phba->sli.ring is already allocated it will not test twice.
> if (!phba->sli.ring)
and unlikely() on this if
> return -ENOMEM;
>
>
thanks
Boaz
^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2014-04-11 9:21 UTC | newest]
Thread overview: 12+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-04-08 20:42 [PATCH] scsi: lpfc: lpfc_init: use kcalloc for allocating memory Matei Oprea
2014-04-09 0:04 ` Joe Perches
2014-04-09 10:48 ` Boaz Harrosh
-- strict thread matches above, loose matches on Subject: below --
2014-04-11 9:21 Matei Oprea
2014-04-08 20:18 Matei Oprea
2014-04-08 20:20 ` Joe Perches
2014-04-08 20:21 ` Matei Oprea
2014-04-08 19:44 Matei Oprea
2014-04-08 19:48 ` Joe Perches
2014-03-18 20:51 Matei Oprea
2014-04-07 12:37 ` Matei Oprea
2014-04-08 6:36 ` Boaz Harrosh
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).