[dpdk-dev] [PATCH v5 3/5] eal: Fix memory leaks and needless increment of pci_map_addr

[Tetsuya Mukawa]

On 2015/06/25 18:18, David Marchand wrote:
> On Thu, Jun 25, 2015 at 5:19 AM, Tetsuya Mukawa  > wrote:
>
> From: "Tetsuya.Mukawa" mailto:mukawa at 
> igel.co.jp>>
>
> This patch fixes following memory leaks.
> - When open() is failed, uio_res and fds won't be freed in
>   pci_uio_map_resource().
> - When pci_map_resource() is failed but path is allocated correctly,
>   path and fds won't be freed in pci_uio_map_recource().
> - When pci_uio_unmap() is called, path should be freed.
>
> Also, fixes below.
> - When pci_map_resource() is failed, mapaddr will be MAP_FAILED.
>   In this case, pci_map_addr should not be incremented in
>   pci_uio_map_resource().
> - To shrink code, move close().
> - Remove fail variable.
>
> Signed-off-by: Tetsuya Mukawa  >
> ---
>  lib/librte_eal/bsdapp/eal/eal_pci.c   | 14 +++--
>  lib/librte_eal/linuxapp/eal/eal_pci_uio.c | 51
> ---
>  2 files changed, 44 insertions(+), 21 deletions(-)
>
> diff --git a/lib/librte_eal/bsdapp/eal/eal_pci.c
> b/lib/librte_eal/bsdapp/eal/eal_pci.c
> index 8e24fd1..b071f07 100644
> --- a/lib/librte_eal/bsdapp/eal/eal_pci.c
> +++ b/lib/librte_eal/bsdapp/eal/eal_pci.c
> @@ -235,7 +235,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> if ((uio_res = rte_zmalloc("UIO_RES", sizeof (*uio_res),
> 0)) == NULL) {
> RTE_LOG(ERR, EAL,
> "%s(): cannot store uio mmap details\n",
> __func__);
> -   return -1;
> +   goto close_fd;
> }
>
> snprintf(uio_res->path, sizeof(uio_res->path), "%s", devname);
> @@ -262,8 +262,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> (mapaddr = pci_map_resource(NULL, devname,
> (off_t)offset,
> (size_t)maps[j].size)
> ) == NULL) {
> -   rte_free(uio_res);
> -   return -1;
> +   goto free_uio_res;
> }
>
> maps[j].addr = mapaddr;
> @@ -274,6 +273,15 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> TAILQ_INSERT_TAIL(uio_res_list, uio_res, next);
>
> return 0;
> +
> +free_uio_res:
> +   rte_free(uio_res);
> +close_fd:
> +   close(dev->intr_handle.fd);
> +   dev->intr_handle.fd = -1;
> +   dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
> +
> +   return -1;
>  }
>
>
> Thinking about it, when something fails, don't you need to unmap pci
> resources in uio_res->maps before freeing ?

Yes, you are right. I will add it to linux code.
So far BSD doesn't have unmap function. It will be added through other
patch series I've already submitted.

Regards,
Tetsuya

>
>
> -- 
> David Marchand

[dpdk-dev] [PATCH v5 3/5] eal: Fix memory leaks and needless increment of pci_map_addr

[Tetsuya Mukawa]

On 2015/06/25 18:16, David Marchand wrote:
> Hello Tetsuya, 
>
>
> On Thu, Jun 25, 2015 at 5:19 AM, Tetsuya Mukawa  > wrote:
>
> From: "Tetsuya.Mukawa" mailto:mukawa at 
> igel.co.jp>>
>
> This patch fixes following memory leaks.
> - When open() is failed, uio_res and fds won't be freed in
>   pci_uio_map_resource().
> - When pci_map_resource() is failed but path is allocated correctly,
>   path and fds won't be freed in pci_uio_map_recource().
> - When pci_uio_unmap() is called, path should be freed.
>
> Also, fixes below.
> - When pci_map_resource() is failed, mapaddr will be MAP_FAILED.
>   In this case, pci_map_addr should not be incremented in
>   pci_uio_map_resource().
> - To shrink code, move close().
> - Remove fail variable.
>
> Signed-off-by: Tetsuya Mukawa  >
> ---
>  lib/librte_eal/bsdapp/eal/eal_pci.c   | 14 +++--
>  lib/librte_eal/linuxapp/eal/eal_pci_uio.c | 51
> ---
>  2 files changed, 44 insertions(+), 21 deletions(-)
>
> [snip]
> diff --git a/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
> b/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
> index 34316b6..2dd83d3 100644
> --- a/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
> +++ b/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
> @@ -308,7 +308,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> if (dev->intr_handle.uio_cfg_fd < 0) {
> RTE_LOG(ERR, EAL, "Cannot open %s: %s\n",
> cfgname, strerror(errno));
> -   return -1;
> +   goto close_fd;
> }
>
> if (dev->kdrv == RTE_KDRV_IGB_UIO)
>
>
>
> You missed a return here :

Hi David,

Thanks for reviewing
It seems I needed to fix it while rebasing.

Regards,
Tetsuya

>
> if (dev->kdrv == RTE_KDRV_IGB_UIO)
>   
> dev->intr_handle.type = RTE_INTR_HANDLE_UIO;  
>   
> else {
>   
> dev->intr_handle.type = RTE_INTR_HANDLE_UIO_INTX;
>
>  
>
> /* set bus master that is not done by uio_pci_generic
> */ 
> if
> (pci_uio_set_bus_master(dev->intr_handle.uio_cfg_fd)) {   
> RTE_LOG(ERR, EAL, "Cannot set up bus
> mastering!\n"); 
> return -1;
>   
> }
>
> }
>
>
>
>
> @@ -328,7 +328,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> if (uio_res == NULL) {
> RTE_LOG(ERR, EAL,
> "%s(): cannot store uio mmap details\n",
> __func__);
> -   return -1;
> +   goto close_fd;
> }
>
> snprintf(uio_res->path, sizeof(uio_res->path), "%s", devname);
> @@ -338,7 +338,6 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> maps = uio_res->maps;
> for (i = 0, map_idx = 0; i != PCI_MAX_RESOURCE; i++) {
> int fd;
> -   int fail = 0;
>
> /* skip empty BAR */
> phaddr = dev->mem_resource[i].phys_addr;
>
>
>
> The rest looks good to me.
>
>
> -- 
> David Marchand 

[dpdk-dev] [PATCH v5 3/5] eal: Fix memory leaks and needless increment of pci_map_addr

[Tetsuya Mukawa]

From: "Tetsuya.Mukawa" 

This patch fixes following memory leaks.
- When open() is failed, uio_res and fds won't be freed in
  pci_uio_map_resource().
- When pci_map_resource() is failed but path is allocated correctly,
  path and fds won't be freed in pci_uio_map_recource().
- When pci_uio_unmap() is called, path should be freed.

Also, fixes below.
- When pci_map_resource() is failed, mapaddr will be MAP_FAILED.
  In this case, pci_map_addr should not be incremented in
  pci_uio_map_resource().
- To shrink code, move close().
- Remove fail variable.

Signed-off-by: Tetsuya Mukawa 
---
 lib/librte_eal/bsdapp/eal/eal_pci.c   | 14 +++--
 lib/librte_eal/linuxapp/eal/eal_pci_uio.c | 51 ---
 2 files changed, 44 insertions(+), 21 deletions(-)

diff --git a/lib/librte_eal/bsdapp/eal/eal_pci.c 
b/lib/librte_eal/bsdapp/eal/eal_pci.c
index 8e24fd1..b071f07 100644
--- a/lib/librte_eal/bsdapp/eal/eal_pci.c
+++ b/lib/librte_eal/bsdapp/eal/eal_pci.c
@@ -235,7 +235,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
if ((uio_res = rte_zmalloc("UIO_RES", sizeof (*uio_res), 0)) == NULL) {
RTE_LOG(ERR, EAL,
"%s(): cannot store uio mmap details\n", __func__);
-   return -1;
+   goto close_fd;
}

snprintf(uio_res->path, sizeof(uio_res->path), "%s", devname);
@@ -262,8 +262,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
(mapaddr = pci_map_resource(NULL, devname, (off_t)offset,
(size_t)maps[j].size)
) == NULL) {
-   rte_free(uio_res);
-   return -1;
+   goto free_uio_res;
}

maps[j].addr = mapaddr;
@@ -274,6 +273,15 @@ pci_uio_map_resource(struct rte_pci_device *dev)
TAILQ_INSERT_TAIL(uio_res_list, uio_res, next);

return 0;
+
+free_uio_res:
+   rte_free(uio_res);
+close_fd:
+   close(dev->intr_handle.fd);
+   dev->intr_handle.fd = -1;
+   dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
+
+   return -1;
 }

 /* Scan one pci sysfs entry, and fill the devices list from it. */
diff --git a/lib/librte_eal/linuxapp/eal/eal_pci_uio.c 
b/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
index 34316b6..2dd83d3 100644
--- a/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
+++ b/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
@@ -308,7 +308,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
if (dev->intr_handle.uio_cfg_fd < 0) {
RTE_LOG(ERR, EAL, "Cannot open %s: %s\n",
cfgname, strerror(errno));
-   return -1;
+   goto close_fd;
}

if (dev->kdrv == RTE_KDRV_IGB_UIO)
@@ -328,7 +328,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
if (uio_res == NULL) {
RTE_LOG(ERR, EAL,
"%s(): cannot store uio mmap details\n", __func__);
-   return -1;
+   goto close_fd;
}

snprintf(uio_res->path, sizeof(uio_res->path), "%s", devname);
@@ -338,7 +338,6 @@ pci_uio_map_resource(struct rte_pci_device *dev)
maps = uio_res->maps;
for (i = 0, map_idx = 0; i != PCI_MAX_RESOURCE; i++) {
int fd;
-   int fail = 0;

/* skip empty BAR */
phaddr = dev->mem_resource[i].phys_addr;
@@ -352,6 +351,11 @@ pci_uio_map_resource(struct rte_pci_device *dev)
loc->domain, loc->bus, loc->devid, 
loc->function,
i);

+   /* allocate memory to keep path */
+   maps[map_idx].path = rte_malloc(NULL, strlen(devname) + 1, 0);
+   if (maps[map_idx].path == NULL)
+   goto free_uio_res;
+
/*
 * open resource file, to mmap it
 */
@@ -359,7 +363,8 @@ pci_uio_map_resource(struct rte_pci_device *dev)
if (fd < 0) {
RTE_LOG(ERR, EAL, "Cannot open %s: %s\n",
devname, strerror(errno));
-   return -1;
+   rte_free(maps[map_idx].path);
+   goto free_uio_res;
}

/* try mapping somewhere close to the end of hugepages */
@@ -368,23 +373,15 @@ pci_uio_map_resource(struct rte_pci_device *dev)

mapaddr = pci_map_resource(pci_map_addr, fd, 0,
(size_t)dev->mem_resource[i].len, 0);
-   if (mapaddr == MAP_FAILED)
-   fail = 1;
+   close(fd);
+   if (mapaddr == MAP_FAILED) {
+   rte_free(maps[map_idx].path);
+   goto free_uio_res;
+   }

pci_map_addr = RTE_PTR_ADD(mapaddr,

[dpdk-dev] [PATCH v5 3/5] eal: Fix memory leaks and needless increment of pci_map_addr

[David Marchand]

On Thu, Jun 25, 2015 at 5:19 AM, Tetsuya Mukawa  wrote:

> From: "Tetsuya.Mukawa" 
>
> This patch fixes following memory leaks.
> - When open() is failed, uio_res and fds won't be freed in
>   pci_uio_map_resource().
> - When pci_map_resource() is failed but path is allocated correctly,
>   path and fds won't be freed in pci_uio_map_recource().
> - When pci_uio_unmap() is called, path should be freed.
>
> Also, fixes below.
> - When pci_map_resource() is failed, mapaddr will be MAP_FAILED.
>   In this case, pci_map_addr should not be incremented in
>   pci_uio_map_resource().
> - To shrink code, move close().
> - Remove fail variable.
>
> Signed-off-by: Tetsuya Mukawa 
> ---
>  lib/librte_eal/bsdapp/eal/eal_pci.c   | 14 +++--
>  lib/librte_eal/linuxapp/eal/eal_pci_uio.c | 51
> ---
>  2 files changed, 44 insertions(+), 21 deletions(-)
>
> diff --git a/lib/librte_eal/bsdapp/eal/eal_pci.c
> b/lib/librte_eal/bsdapp/eal/eal_pci.c
> index 8e24fd1..b071f07 100644
> --- a/lib/librte_eal/bsdapp/eal/eal_pci.c
> +++ b/lib/librte_eal/bsdapp/eal/eal_pci.c
> @@ -235,7 +235,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> if ((uio_res = rte_zmalloc("UIO_RES", sizeof (*uio_res), 0)) ==
> NULL) {
> RTE_LOG(ERR, EAL,
> "%s(): cannot store uio mmap details\n", __func__);
> -   return -1;
> +   goto close_fd;
> }
>
> snprintf(uio_res->path, sizeof(uio_res->path), "%s", devname);
> @@ -262,8 +262,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> (mapaddr = pci_map_resource(NULL, devname,
> (off_t)offset,
> (size_t)maps[j].size)
> ) == NULL) {
> -   rte_free(uio_res);
> -   return -1;
> +   goto free_uio_res;
> }
>
> maps[j].addr = mapaddr;
> @@ -274,6 +273,15 @@ pci_uio_map_resource(struct rte_pci_device *dev)
> TAILQ_INSERT_TAIL(uio_res_list, uio_res, next);
>
> return 0;
> +
> +free_uio_res:
> +   rte_free(uio_res);
> +close_fd:
> +   close(dev->intr_handle.fd);
> +   dev->intr_handle.fd = -1;
> +   dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
> +
> +   return -1;
>  }
>
>
Thinking about it, when something fails, don't you need to unmap pci
resources in uio_res->maps before freeing ?


-- 
David Marchand