* [dpdk-dev] [PATCH] i40e: fix the issue of not freeing memzone
@ 2015-11-06 3:26 Helin Zhang
2015-11-06 5:39 ` Wu, Jingjing
2015-11-06 7:57 ` [dpdk-dev] [PATCH v2] " Helin Zhang
0 siblings, 2 replies; 11+ messages in thread
From: Helin Zhang @ 2015-11-06 3:26 UTC (permalink / raw)
To: dev
This fixes the issue of not freeing memzone in a call to free the
memory for adminq DMA.
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
---
doc/guides/rel_notes/release_2_2.rst | 5 +++++
drivers/net/i40e/base/i40e_osdep.h | 2 +-
drivers/net/i40e/i40e_ethdev.c | 12 +++++++-----
3 files changed, 13 insertions(+), 6 deletions(-)
diff --git a/doc/guides/rel_notes/release_2_2.rst b/doc/guides/rel_notes/release_2_2.rst
index 59dda59..eaa906c 100644
--- a/doc/guides/rel_notes/release_2_2.rst
+++ b/doc/guides/rel_notes/release_2_2.rst
@@ -150,6 +150,11 @@ Drivers
Added discarding packets on VSI to the stats and rectify the old statistics.
+* **i40e: Fixed issue of not freeing memzone.**
+
+ Fixed the issue of not freeing memzone in the call to free the memory for
+ adminq DMA.
+
* **vhost: Fixed Qemu shutdown.**
Fixed issue with libvirt ``virsh destroy`` not killing the VM.
diff --git a/drivers/net/i40e/base/i40e_osdep.h b/drivers/net/i40e/base/i40e_osdep.h
index 70d2721..71077f0 100644
--- a/drivers/net/i40e/base/i40e_osdep.h
+++ b/drivers/net/i40e/base/i40e_osdep.h
@@ -146,7 +146,7 @@ struct i40e_dma_mem {
void *va;
u64 pa;
u32 size;
- u64 id;
+ const void *zone;
} __attribute__((packed));
#define i40e_allocate_dma_mem(h, m, unused, s, a) \
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index ddf3d38..8d6c0fa 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -2910,15 +2910,13 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
u64 size,
u32 alignment)
{
- static uint64_t id = 0;
const struct rte_memzone *mz = NULL;
char z_name[RTE_MEMZONE_NAMESIZE];
if (!mem)
return I40E_ERR_PARAM;
- id++;
- snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, id);
+ snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, rte_rand());
#ifdef RTE_LIBRTE_XEN_DOM0
mz = rte_memzone_reserve_bounded(z_name, size, SOCKET_ID_ANY, 0,
alignment, RTE_PGSIZE_2M);
@@ -2929,7 +2927,6 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
if (!mz)
return I40E_ERR_NO_MEMORY;
- mem->id = id;
mem->size = size;
mem->va = mz->addr;
#ifdef RTE_LIBRTE_XEN_DOM0
@@ -2937,6 +2934,8 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
#else
mem->pa = mz->phys_addr;
#endif
+ mem->zone = (const void *)mz;
+ PMD_DRV_LOG(DEBUG, "memzone allocated: %p", mem->zone);
return I40E_SUCCESS;
}
@@ -2950,9 +2949,12 @@ enum i40e_status_code
i40e_free_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
struct i40e_dma_mem *mem)
{
- if (!mem || !mem->va)
+ if (!mem)
return I40E_ERR_PARAM;
+ PMD_DRV_LOG(DEBUG, "memzone to be freed: %p", mem->zone);
+ rte_memzone_free((const struct rte_memzone *)mem->zone);
+ mem->zone = NULL;
mem->va = NULL;
mem->pa = (u64)0;
--
1.9.3
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH] i40e: fix the issue of not freeing memzone
2015-11-06 3:26 [dpdk-dev] [PATCH] i40e: fix the issue of not freeing memzone Helin Zhang
@ 2015-11-06 5:39 ` Wu, Jingjing
2015-11-06 6:00 ` Zhang, Helin
2015-11-06 7:57 ` [dpdk-dev] [PATCH v2] " Helin Zhang
1 sibling, 1 reply; 11+ messages in thread
From: Wu, Jingjing @ 2015-11-06 5:39 UTC (permalink / raw)
To: Zhang, Helin, dev
> - static uint64_t id = 0;
> const struct rte_memzone *mz = NULL;
> char z_name[RTE_MEMZONE_NAMESIZE];
>
> if (!mem)
> return I40E_ERR_PARAM;
>
> - id++;
> - snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, id);
> + snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, rte_rand());
Why change id++ to rte_rand() ?
> #ifdef RTE_LIBRTE_XEN_DOM0
> mz = rte_memzone_reserve_bounded(z_name, size,
> SOCKET_ID_ANY, 0,
> alignment, RTE_PGSIZE_2M);
> @@ -2929,7 +2927,6 @@
> i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
> if (!mz)
> return I40E_ERR_NO_MEMORY;
>
> - mem->id = id;
> mem->size = size;
> mem->va = mz->addr;
> #ifdef RTE_LIBRTE_XEN_DOM0
> @@ -2937,6 +2934,8 @@
> i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
> #else
> mem->pa = mz->phys_addr;
> #endif
> + mem->zone = (const void *)mz;
> + PMD_DRV_LOG(DEBUG, "memzone allocated: %p", mem->zone);
>
Why not print the name of mem_zone instead of pointer?
Thanks
Jingjing
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH] i40e: fix the issue of not freeing memzone
2015-11-06 5:39 ` Wu, Jingjing
@ 2015-11-06 6:00 ` Zhang, Helin
0 siblings, 0 replies; 11+ messages in thread
From: Zhang, Helin @ 2015-11-06 6:00 UTC (permalink / raw)
To: Wu, Jingjing, dev
> -----Original Message-----
> From: Wu, Jingjing
> Sent: Friday, November 6, 2015 1:40 PM
> To: Zhang, Helin; dev@dpdk.org
> Subject: RE: [PATCH] i40e: fix the issue of not freeing memzone
>
> > - static uint64_t id = 0;
> > const struct rte_memzone *mz = NULL;
> > char z_name[RTE_MEMZONE_NAMESIZE];
> >
> > if (!mem)
> > return I40E_ERR_PARAM;
> >
> > - id++;
> > - snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, id);
> > + snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, rte_rand());
>
> Why change id++ to rte_rand() ?
Don't need to maintain the ID, which may have race condition issue.
Get a random data is good enough. Some other PMDs are using tsc count for the similar purpose.
>
> > #ifdef RTE_LIBRTE_XEN_DOM0
> > mz = rte_memzone_reserve_bounded(z_name, size, SOCKET_ID_ANY, 0,
> > alignment, RTE_PGSIZE_2M);
> > @@ -2929,7 +2927,6 @@
> > i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
> > if (!mz)
> > return I40E_ERR_NO_MEMORY;
> >
> > - mem->id = id;
> > mem->size = size;
> > mem->va = mz->addr;
> > #ifdef RTE_LIBRTE_XEN_DOM0
> > @@ -2937,6 +2934,8 @@
> > i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
> > #else
> > mem->pa = mz->phys_addr;
> > #endif
> > + mem->zone = (const void *)mz;
> > + PMD_DRV_LOG(DEBUG, "memzone allocated: %p", mem->zone);
> >
> Why not print the name of mem_zone instead of pointer?
Good idea to print the name instead, and possible physical address, virtual address, etc.
Regards,
Helin
>
>
> Thanks
> Jingjing
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH v2] i40e: fix the issue of not freeing memzone
2015-11-06 3:26 [dpdk-dev] [PATCH] i40e: fix the issue of not freeing memzone Helin Zhang
2015-11-06 5:39 ` Wu, Jingjing
@ 2015-11-06 7:57 ` Helin Zhang
2015-11-06 8:32 ` Wu, Jingjing
` (2 more replies)
1 sibling, 3 replies; 11+ messages in thread
From: Helin Zhang @ 2015-11-06 7:57 UTC (permalink / raw)
To: dev
This fixes the issue of not freeing memzone in a call to free the
memory for adminq DMA.
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
---
doc/guides/rel_notes/release_2_2.rst | 5 +++++
drivers/net/i40e/base/i40e_osdep.h | 2 +-
drivers/net/i40e/i40e_ethdev.c | 14 +++++++++-----
3 files changed, 15 insertions(+), 6 deletions(-)
v2 changes:
Reworked debug messages.
diff --git a/doc/guides/rel_notes/release_2_2.rst b/doc/guides/rel_notes/release_2_2.rst
index 59dda59..eaa906c 100644
--- a/doc/guides/rel_notes/release_2_2.rst
+++ b/doc/guides/rel_notes/release_2_2.rst
@@ -150,6 +150,11 @@ Drivers
Added discarding packets on VSI to the stats and rectify the old statistics.
+* **i40e: Fixed issue of not freeing memzone.**
+
+ Fixed the issue of not freeing memzone in the call to free the memory for
+ adminq DMA.
+
* **vhost: Fixed Qemu shutdown.**
Fixed issue with libvirt ``virsh destroy`` not killing the VM.
diff --git a/drivers/net/i40e/base/i40e_osdep.h b/drivers/net/i40e/base/i40e_osdep.h
index 70d2721..71077f0 100644
--- a/drivers/net/i40e/base/i40e_osdep.h
+++ b/drivers/net/i40e/base/i40e_osdep.h
@@ -146,7 +146,7 @@ struct i40e_dma_mem {
void *va;
u64 pa;
u32 size;
- u64 id;
+ const void *zone;
} __attribute__((packed));
#define i40e_allocate_dma_mem(h, m, unused, s, a) \
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index ddf3d38..9f06ec2 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -2910,15 +2910,13 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
u64 size,
u32 alignment)
{
- static uint64_t id = 0;
const struct rte_memzone *mz = NULL;
char z_name[RTE_MEMZONE_NAMESIZE];
if (!mem)
return I40E_ERR_PARAM;
- id++;
- snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, id);
+ snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, rte_rand());
#ifdef RTE_LIBRTE_XEN_DOM0
mz = rte_memzone_reserve_bounded(z_name, size, SOCKET_ID_ANY, 0,
alignment, RTE_PGSIZE_2M);
@@ -2929,7 +2927,6 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
if (!mz)
return I40E_ERR_NO_MEMORY;
- mem->id = id;
mem->size = size;
mem->va = mz->addr;
#ifdef RTE_LIBRTE_XEN_DOM0
@@ -2937,6 +2934,9 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
#else
mem->pa = mz->phys_addr;
#endif
+ mem->zone = (const void *)mz;
+ PMD_DRV_LOG(DEBUG, "memzone %s allocated with physical address: %p",
+ mz->name, mem->pa);
return I40E_SUCCESS;
}
@@ -2950,9 +2950,13 @@ enum i40e_status_code
i40e_free_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
struct i40e_dma_mem *mem)
{
- if (!mem || !mem->va)
+ if (!mem)
return I40E_ERR_PARAM;
+ PMD_DRV_LOG(DEBUG, "memzone %s to be freed with physical address: %p",
+ ((const struct rte_memzone *)mem->zone)->name, mem->pa);
+ rte_memzone_free((const struct rte_memzone *)mem->zone);
+ mem->zone = NULL;
mem->va = NULL;
mem->pa = (u64)0;
--
1.9.3
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH v2] i40e: fix the issue of not freeing memzone
2015-11-06 7:57 ` [dpdk-dev] [PATCH v2] " Helin Zhang
@ 2015-11-06 8:32 ` Wu, Jingjing
2015-11-06 10:19 ` Thomas Monjalon
2015-11-09 1:20 ` [dpdk-dev] [PATCH v3] " Helin Zhang
2 siblings, 0 replies; 11+ messages in thread
From: Wu, Jingjing @ 2015-11-06 8:32 UTC (permalink / raw)
To: Zhang, Helin, dev
> -----Original Message-----
> From: Zhang, Helin
> Sent: Friday, November 06, 2015 3:57 PM
> To: dev@dpdk.org
> Cc: Wu, Jingjing; Zhang, Helin
> Subject: [PATCH v2] i40e: fix the issue of not freeing memzone
>
> This fixes the issue of not freeing memzone in a call to free the memory for
> adminq DMA.
>
> Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH v2] i40e: fix the issue of not freeing memzone
2015-11-06 7:57 ` [dpdk-dev] [PATCH v2] " Helin Zhang
2015-11-06 8:32 ` Wu, Jingjing
@ 2015-11-06 10:19 ` Thomas Monjalon
2015-11-09 1:20 ` Zhang, Helin
2015-11-09 1:20 ` [dpdk-dev] [PATCH v3] " Helin Zhang
2 siblings, 1 reply; 11+ messages in thread
From: Thomas Monjalon @ 2015-11-06 10:19 UTC (permalink / raw)
To: Helin Zhang; +Cc: dev
2015-11-06 15:57, Helin Zhang:
> This fixes the issue of not freeing memzone in a call to free the
> memory for adminq DMA.
>
> Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Please could you add a "Fixes:" line?
Thanks
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH v3] i40e: fix the issue of not freeing memzone
2015-11-06 7:57 ` [dpdk-dev] [PATCH v2] " Helin Zhang
2015-11-06 8:32 ` Wu, Jingjing
2015-11-06 10:19 ` Thomas Monjalon
@ 2015-11-09 1:20 ` Helin Zhang
2015-11-10 15:53 ` Thomas Monjalon
2015-11-11 5:28 ` [dpdk-dev] [PATCH v4] " Helin Zhang
2 siblings, 2 replies; 11+ messages in thread
From: Helin Zhang @ 2015-11-09 1:20 UTC (permalink / raw)
To: dev
This fixes the issue of not freeing memzone in a call to free the
memory for adminq DMA.
Fixes: 4861cde46116 ("i40e: new poll mode driver")
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
---
doc/guides/rel_notes/release_2_2.rst | 5 +++++
drivers/net/i40e/base/i40e_osdep.h | 2 +-
drivers/net/i40e/i40e_ethdev.c | 14 +++++++++-----
3 files changed, 15 insertions(+), 6 deletions(-)
diff --git a/doc/guides/rel_notes/release_2_2.rst b/doc/guides/rel_notes/release_2_2.rst
index 59dda59..eaa906c 100644
--- a/doc/guides/rel_notes/release_2_2.rst
+++ b/doc/guides/rel_notes/release_2_2.rst
@@ -150,6 +150,11 @@ Drivers
Added discarding packets on VSI to the stats and rectify the old statistics.
+* **i40e: Fixed issue of not freeing memzone.**
+
+ Fixed the issue of not freeing memzone in the call to free the memory for
+ adminq DMA.
+
* **vhost: Fixed Qemu shutdown.**
Fixed issue with libvirt ``virsh destroy`` not killing the VM.
diff --git a/drivers/net/i40e/base/i40e_osdep.h b/drivers/net/i40e/base/i40e_osdep.h
index 70d2721..71077f0 100644
--- a/drivers/net/i40e/base/i40e_osdep.h
+++ b/drivers/net/i40e/base/i40e_osdep.h
@@ -146,7 +146,7 @@ struct i40e_dma_mem {
void *va;
u64 pa;
u32 size;
- u64 id;
+ const void *zone;
} __attribute__((packed));
#define i40e_allocate_dma_mem(h, m, unused, s, a) \
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index ddf3d38..9f06ec2 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -2910,15 +2910,13 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
u64 size,
u32 alignment)
{
- static uint64_t id = 0;
const struct rte_memzone *mz = NULL;
char z_name[RTE_MEMZONE_NAMESIZE];
if (!mem)
return I40E_ERR_PARAM;
- id++;
- snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, id);
+ snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, rte_rand());
#ifdef RTE_LIBRTE_XEN_DOM0
mz = rte_memzone_reserve_bounded(z_name, size, SOCKET_ID_ANY, 0,
alignment, RTE_PGSIZE_2M);
@@ -2929,7 +2927,6 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
if (!mz)
return I40E_ERR_NO_MEMORY;
- mem->id = id;
mem->size = size;
mem->va = mz->addr;
#ifdef RTE_LIBRTE_XEN_DOM0
@@ -2937,6 +2934,9 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
#else
mem->pa = mz->phys_addr;
#endif
+ mem->zone = (const void *)mz;
+ PMD_DRV_LOG(DEBUG, "memzone %s allocated with physical address: %p",
+ mz->name, mem->pa);
return I40E_SUCCESS;
}
@@ -2950,9 +2950,13 @@ enum i40e_status_code
i40e_free_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
struct i40e_dma_mem *mem)
{
- if (!mem || !mem->va)
+ if (!mem)
return I40E_ERR_PARAM;
+ PMD_DRV_LOG(DEBUG, "memzone %s to be freed with physical address: %p",
+ ((const struct rte_memzone *)mem->zone)->name, mem->pa);
+ rte_memzone_free((const struct rte_memzone *)mem->zone);
+ mem->zone = NULL;
mem->va = NULL;
mem->pa = (u64)0;
--
1.9.3
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH v2] i40e: fix the issue of not freeing memzone
2015-11-06 10:19 ` Thomas Monjalon
@ 2015-11-09 1:20 ` Zhang, Helin
0 siblings, 0 replies; 11+ messages in thread
From: Zhang, Helin @ 2015-11-09 1:20 UTC (permalink / raw)
To: Thomas Monjalon; +Cc: dev
> -----Original Message-----
> From: Thomas Monjalon [mailto:thomas.monjalon@6wind.com]
> Sent: Friday, November 6, 2015 6:20 PM
> To: Zhang, Helin
> Cc: dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2] i40e: fix the issue of not freeing memzone
>
> 2015-11-06 15:57, Helin Zhang:
> > This fixes the issue of not freeing memzone in a call to free the
> > memory for adminq DMA.
> >
> > Signed-off-by: Helin Zhang <helin.zhang@intel.com>
>
> Please could you add a "Fixes:" line?
> Thanks
Sure, will send a new patch for that. Thanks!
Helin
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH v3] i40e: fix the issue of not freeing memzone
2015-11-09 1:20 ` [dpdk-dev] [PATCH v3] " Helin Zhang
@ 2015-11-10 15:53 ` Thomas Monjalon
2015-11-11 5:28 ` [dpdk-dev] [PATCH v4] " Helin Zhang
1 sibling, 0 replies; 11+ messages in thread
From: Thomas Monjalon @ 2015-11-10 15:53 UTC (permalink / raw)
To: Helin Zhang; +Cc: dev
Helin,
There is a compilation error.
2015-11-09 09:20, Helin Zhang:
> + mem->zone = (const void *)mz;
> + PMD_DRV_LOG(DEBUG, "memzone %s allocated with physical address: %p",
> + mz->name, mem->pa);
error: format ‘%p’ expects argument of type ‘void *’, but argument 6 has type ‘u64 {aka long unsigned int}’
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH v4] i40e: fix the issue of not freeing memzone
2015-11-09 1:20 ` [dpdk-dev] [PATCH v3] " Helin Zhang
2015-11-10 15:53 ` Thomas Monjalon
@ 2015-11-11 5:28 ` Helin Zhang
2015-11-11 18:10 ` Thomas Monjalon
1 sibling, 1 reply; 11+ messages in thread
From: Helin Zhang @ 2015-11-11 5:28 UTC (permalink / raw)
To: dev
This fixes the issue of not freeing memzone in a call to free the
memory for adminq DMA.
Fixes: 4861cde46116 ("i40e: new poll mode driver")
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
---
doc/guides/rel_notes/release_2_2.rst | 5 +++++
drivers/net/i40e/base/i40e_osdep.h | 2 +-
drivers/net/i40e/i40e_ethdev.c | 15 ++++++++++-----
3 files changed, 16 insertions(+), 6 deletions(-)
v2 changes:
Reworked the debug logs.
v3 changes:
Added more info into the commit logs.
v4 changes:
Fixed the compile issues in debug mode.
diff --git a/doc/guides/rel_notes/release_2_2.rst b/doc/guides/rel_notes/release_2_2.rst
index 59dda59..eaa906c 100644
--- a/doc/guides/rel_notes/release_2_2.rst
+++ b/doc/guides/rel_notes/release_2_2.rst
@@ -150,6 +150,11 @@ Drivers
Added discarding packets on VSI to the stats and rectify the old statistics.
+* **i40e: Fixed issue of not freeing memzone.**
+
+ Fixed the issue of not freeing memzone in the call to free the memory for
+ adminq DMA.
+
* **vhost: Fixed Qemu shutdown.**
Fixed issue with libvirt ``virsh destroy`` not killing the VM.
diff --git a/drivers/net/i40e/base/i40e_osdep.h b/drivers/net/i40e/base/i40e_osdep.h
index 70d2721..71077f0 100644
--- a/drivers/net/i40e/base/i40e_osdep.h
+++ b/drivers/net/i40e/base/i40e_osdep.h
@@ -146,7 +146,7 @@ struct i40e_dma_mem {
void *va;
u64 pa;
u32 size;
- u64 id;
+ const void *zone;
} __attribute__((packed));
#define i40e_allocate_dma_mem(h, m, unused, s, a) \
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index ddf3d38..c156665 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -2910,15 +2910,13 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
u64 size,
u32 alignment)
{
- static uint64_t id = 0;
const struct rte_memzone *mz = NULL;
char z_name[RTE_MEMZONE_NAMESIZE];
if (!mem)
return I40E_ERR_PARAM;
- id++;
- snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, id);
+ snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, rte_rand());
#ifdef RTE_LIBRTE_XEN_DOM0
mz = rte_memzone_reserve_bounded(z_name, size, SOCKET_ID_ANY, 0,
alignment, RTE_PGSIZE_2M);
@@ -2929,7 +2927,6 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
if (!mz)
return I40E_ERR_NO_MEMORY;
- mem->id = id;
mem->size = size;
mem->va = mz->addr;
#ifdef RTE_LIBRTE_XEN_DOM0
@@ -2937,6 +2934,9 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
#else
mem->pa = mz->phys_addr;
#endif
+ mem->zone = (const void *)mz;
+ PMD_DRV_LOG(DEBUG, "memzone %s allocated with physical address: "
+ "%"PRIu64, mz->name, mem->pa);
return I40E_SUCCESS;
}
@@ -2950,9 +2950,14 @@ enum i40e_status_code
i40e_free_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw,
struct i40e_dma_mem *mem)
{
- if (!mem || !mem->va)
+ if (!mem)
return I40E_ERR_PARAM;
+ PMD_DRV_LOG(DEBUG, "memzone %s to be freed with physical address: "
+ "%"PRIu64, ((const struct rte_memzone *)mem->zone)->name,
+ mem->pa);
+ rte_memzone_free((const struct rte_memzone *)mem->zone);
+ mem->zone = NULL;
mem->va = NULL;
mem->pa = (u64)0;
--
1.9.3
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH v4] i40e: fix the issue of not freeing memzone
2015-11-11 5:28 ` [dpdk-dev] [PATCH v4] " Helin Zhang
@ 2015-11-11 18:10 ` Thomas Monjalon
0 siblings, 0 replies; 11+ messages in thread
From: Thomas Monjalon @ 2015-11-11 18:10 UTC (permalink / raw)
To: Helin Zhang; +Cc: dev
2015-11-11 13:28, Helin Zhang:
> This fixes the issue of not freeing memzone in a call to free the
> memory for adminq DMA.
>
> Fixes: 4861cde46116 ("i40e: new poll mode driver")
>
> Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Applied, thanks
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2015-11-11 18:11 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-11-06 3:26 [dpdk-dev] [PATCH] i40e: fix the issue of not freeing memzone Helin Zhang
2015-11-06 5:39 ` Wu, Jingjing
2015-11-06 6:00 ` Zhang, Helin
2015-11-06 7:57 ` [dpdk-dev] [PATCH v2] " Helin Zhang
2015-11-06 8:32 ` Wu, Jingjing
2015-11-06 10:19 ` Thomas Monjalon
2015-11-09 1:20 ` Zhang, Helin
2015-11-09 1:20 ` [dpdk-dev] [PATCH v3] " Helin Zhang
2015-11-10 15:53 ` Thomas Monjalon
2015-11-11 5:28 ` [dpdk-dev] [PATCH v4] " Helin Zhang
2015-11-11 18:10 ` Thomas Monjalon
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).