[dpdk-dev] [PATCH 0/4] fix compilation on Fedora 22

[dpdk-dev] [PATCH 0/4] fix compilation on Fedora 22

[Bruce Richardson]

GCC version 5.1 included with Fedora 22 is running stricter array 
bounds checks which are throwing up errors in a number of components 
in the DPDK code. This patchset fixes these errors to allow compilation
with GCC on Fedora 22.

Example error:
== Build lib/librte_eal/linuxapp/eal
  CC eal_memory.o
/home/bruce/dpdk.org/lib/librte_eal/linuxapp/eal/eal_memory.c: In function 'rte_eal_hugepage_init':
/home/bruce/dpdk.org/lib/librte_eal/linuxapp/eal/eal_memory.c:1193:35: error: array subscript is above array bounds [-Werror=array-bounds]
      internal_config.hugepage_info[j].hugepage_sz) {
                                   ^
compilation terminated due to -Wfatal-errors.


Bruce Richardson (4):
  eal: fix compile on Fedora 22 (GCC 5.1)
  ip_frag: fix compile on Fedora 22 (GCC 5.1)
  i40e: fix compile on Fedora 22 (GCC 5.1)
  null: fix compile on Fedora 22 (GCC 5.1)

 drivers/net/i40e/i40e_fdir.c             |  2 +-
 drivers/net/null/rte_eth_null.c          | 12 ++++++++----
 lib/librte_eal/linuxapp/eal/eal_memory.c |  3 ++-
 lib/librte_ip_frag/ip_frag_common.h      |  2 +-
 4 files changed, 12 insertions(+), 7 deletions(-)

-- 
2.4.1

[dpdk-dev] [PATCH 1/4] eal: fix compile on Fedora 22 (GCC 5.1)

[Bruce Richardson]

On Fedora 22, with GCC 5.1, errors are reported due to array accesses
being potentially out of bounds. This commit fixes this by adding in an
extra bounds check to the loop counter.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 lib/librte_eal/linuxapp/eal/eal_memory.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/lib/librte_eal/linuxapp/eal/eal_memory.c b/lib/librte_eal/linuxapp/eal/eal_memory.c
index 5f9f92e..744635f 100644
--- a/lib/librte_eal/linuxapp/eal/eal_memory.c
+++ b/lib/librte_eal/linuxapp/eal/eal_memory.c
@@ -1188,7 +1188,8 @@ rte_eal_hugepage_init(void)
 		int socket = tmp_hp[i].socket_id;
 
 		/* find a hugepage info with right size and increment num_pages */
-		for (j = 0; j < (int) internal_config.num_hugepage_sizes; j++) {
+		for (j = 0; j < (int) internal_config.num_hugepage_sizes &&
+				j < MAX_HUGEPAGE_SIZES; j++) {
 			if (tmp_hp[i].size ==
 					internal_config.hugepage_info[j].hugepage_sz) {
 #ifdef RTE_EAL_SINGLE_FILE_SEGMENTS
-- 
2.4.1

Re: [dpdk-dev] [PATCH 1/4] eal: fix compile on Fedora 22 (GCC 5.1)

[Neil Horman]

On Fri, May 29, 2015 at 01:53:44PM +0100, Bruce Richardson wrote:
> On Fedora 22, with GCC 5.1, errors are reported due to array accesses
> being potentially out of bounds. This commit fixes this by adding in an
> extra bounds check to the loop counter.
> 
> Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> ---
>  lib/librte_eal/linuxapp/eal/eal_memory.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/lib/librte_eal/linuxapp/eal/eal_memory.c b/lib/librte_eal/linuxapp/eal/eal_memory.c
> index 5f9f92e..744635f 100644
> --- a/lib/librte_eal/linuxapp/eal/eal_memory.c
> +++ b/lib/librte_eal/linuxapp/eal/eal_memory.c
> @@ -1188,7 +1188,8 @@ rte_eal_hugepage_init(void)
>  		int socket = tmp_hp[i].socket_id;
>  
>  		/* find a hugepage info with right size and increment num_pages */
> -		for (j = 0; j < (int) internal_config.num_hugepage_sizes; j++) {
> +		for (j = 0; j < (int) internal_config.num_hugepage_sizes &&
> +				j < MAX_HUGEPAGE_SIZES; j++) {
>  			if (tmp_hp[i].size ==
Use RTE_MIN here, it will look alot cleaner I think
Neil

>  					internal_config.hugepage_info[j].hugepage_sz) {
>  #ifdef RTE_EAL_SINGLE_FILE_SEGMENTS
> -- 
> 2.4.1
> 
> 

[dpdk-dev] [PATCH 2/4] ip_frag: fix compile on Fedora 22 (GCC 5.1)

[Bruce Richardson]

On Fedora 22, with GCC 5.1, errors are reported due to array accesses
being potentially out of bounds. This commit fixes this by adding in an
extra bounds check to the loop counter.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 lib/librte_ip_frag/ip_frag_common.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/librte_ip_frag/ip_frag_common.h b/lib/librte_ip_frag/ip_frag_common.h
index 210f409..e37073b 100644
--- a/lib/librte_ip_frag/ip_frag_common.h
+++ b/lib/librte_ip_frag/ip_frag_common.h
@@ -90,7 +90,7 @@ static inline int
 ip_frag_key_is_empty(const struct ip_frag_key * key)
 {
 	uint32_t i;
-	for (i = 0; i < key->key_len; i++)
+	for (i = 0; i < key->key_len && i < RTE_DIM(key->src_dst); i++)
 		if (key->src_dst[i] != 0)
 			return 0;
 	return 1;
-- 
2.4.1

Re: [dpdk-dev] [PATCH 2/4] ip_frag: fix compile on Fedora 22 (GCC 5.1)

[Neil Horman]

On Fri, May 29, 2015 at 01:53:45PM +0100, Bruce Richardson wrote:
> On Fedora 22, with GCC 5.1, errors are reported due to array accesses
> being potentially out of bounds. This commit fixes this by adding in an
> extra bounds check to the loop counter.
> 
> Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> ---
>  lib/librte_ip_frag/ip_frag_common.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/librte_ip_frag/ip_frag_common.h b/lib/librte_ip_frag/ip_frag_common.h
> index 210f409..e37073b 100644
> --- a/lib/librte_ip_frag/ip_frag_common.h
> +++ b/lib/librte_ip_frag/ip_frag_common.h
> @@ -90,7 +90,7 @@ static inline int
>  ip_frag_key_is_empty(const struct ip_frag_key * key)
>  {
>  	uint32_t i;
> -	for (i = 0; i < key->key_len; i++)
> +	for (i = 0; i < key->key_len && i < RTE_DIM(key->src_dst); i++)
>  		if (key->src_dst[i] != 0)
>  			return 0;
>  	return 1;
> -- 
> 2.4.1
> 
> 
RTE_MIN again
Neil

[dpdk-dev] [PATCH 3/4] i40e: fix compile on Fedora 22 (GCC 5.1)

[Bruce Richardson]

On Fedora 22, with GCC 5.1, errors are reported due to array accesses
being potentially out of bounds. This commit fixes this by adding in an
extra bounds check to the loop counter.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 drivers/net/i40e/i40e_fdir.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/i40e/i40e_fdir.c b/drivers/net/i40e/i40e_fdir.c
index e688b4f..55cf2cf 100644
--- a/drivers/net/i40e/i40e_fdir.c
+++ b/drivers/net/i40e/i40e_fdir.c
@@ -551,7 +551,7 @@ i40e_set_flx_pld_cfg(struct i40e_pf *pf,
 	memset(flex_pit, 0, sizeof(flex_pit));
 	num = i40e_srcoff_to_flx_pit(cfg->src_offset, flex_pit);
 
-	for (i = 0; i < num; i++) {
+	for (i = 0; i < num && i < RTE_DIM(flex_pit); i++) {
 		field_idx = layer_idx * I40E_MAX_FLXPLD_FIED + i;
 		/* record the info in fdir structure */
 		pf->fdir.flex_set[field_idx].src_offset =
-- 
2.4.1

[dpdk-dev] [PATCH 4/4] null: fix compile on Fedora 22 (GCC 5.1)

[Bruce Richardson]

On Fedora 22, with GCC 5.1, errors are reported due to array accesses
being potentially out of bounds. This commit fixes this by adding in an
extra bounds check to the loop counter.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 drivers/net/null/rte_eth_null.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/drivers/net/null/rte_eth_null.c b/drivers/net/null/rte_eth_null.c
index 5895065..4ca7f09 100644
--- a/drivers/net/null/rte_eth_null.c
+++ b/drivers/net/null/rte_eth_null.c
@@ -298,7 +298,8 @@ eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *igb_stats)
 	internal = dev->data->dev_private;
 	memset(igb_stats, 0, sizeof(*igb_stats));
 	num_stats = RTE_MIN((unsigned)RTE_ETHDEV_QUEUE_STAT_CNTRS,
-					internal->nb_rx_queues);
+			RTE_MIN(internal->nb_rx_queues,
+				RTE_DIM(internal->rx_null_queues)));
 	for (i = 0; i < num_stats; i++) {
 		igb_stats->q_ipackets[i] =
 			internal->rx_null_queues[i].rx_pkts.cnt;
@@ -306,7 +307,8 @@ eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *igb_stats)
 	}
 
 	num_stats = RTE_MIN((unsigned)RTE_ETHDEV_QUEUE_STAT_CNTRS,
-					internal->nb_tx_queues);
+			RTE_MIN(internal->nb_tx_queues,
+				RTE_DIM(internal->tx_null_queues)));
 	for (i = 0; i < num_stats; i++) {
 		igb_stats->q_opackets[i] =
 			internal->tx_null_queues[i].tx_pkts.cnt;
@@ -331,9 +333,11 @@ eth_stats_reset(struct rte_eth_dev *dev)
 		return;
 
 	internal = dev->data->dev_private;
-	for (i = 0; i < internal->nb_rx_queues; i++)
+	for (i = 0; i < internal->nb_rx_queues &&
+			i < RTE_DIM(internal->rx_null_queues); i++)
 		internal->rx_null_queues[i].rx_pkts.cnt = 0;
-	for (i = 0; i < internal->nb_tx_queues; i++) {
+	for (i = 0; i < internal->nb_tx_queues &&
+			i < RTE_DIM(internal->tx_null_queues); i++) {
 		internal->tx_null_queues[i].tx_pkts.cnt = 0;
 		internal->tx_null_queues[i].err_pkts.cnt = 0;
 	}
-- 
2.4.1

[dpdk-dev] [PATCH v2 0/4] fix compilation on Fedora 22

[Bruce Richardson]

V2 Changes: 
* use RTE_MIN instead of additional comparison checks
  in some loops, as suggested by Neil.
* For stats reset in null PMD, just use the size of the
  array [which is 1] in place of the used queue counts.

GCC version 5.1 included with Fedora 22 is running stricter array 
bounds checks which are throwing up errors in a number of components 
in the DPDK code. This patchset fixes these errors to allow compilation
with GCC on Fedora 22.

Example error:
== Build lib/librte_eal/linuxapp/eal
  CC eal_memory.o
/home/bruce/dpdk.org/lib/librte_eal/linuxapp/eal/eal_memory.c: In function 'rte_eal_hugepage_init':
/home/bruce/dpdk.org/lib/librte_eal/linuxapp/eal/eal_memory.c:1193:35: error: array subscript is above array bounds [-Werror=array-bounds]
      internal_config.hugepage_info[j].hugepage_sz) {
                                   ^
compilation terminated due to -Wfatal-errors.

Bruce Richardson (4):
  eal: fix compile on Fedora 22 (GCC 5.1)
  ip_frag: fix compile on Fedora 22 (GCC 5.1)
  i40e: fix compile on Fedora 22 (GCC 5.1)
  null: fix compile on Fedora 22 (GCC 5.1)

 drivers/net/i40e/i40e_fdir.c             |  2 +-
 drivers/net/null/rte_eth_null.c          | 12 ++++++++----
 lib/librte_eal/linuxapp/eal/eal_memory.c |  4 +++-
 lib/librte_ip_frag/ip_frag_common.h      |  2 +-
 4 files changed, 13 insertions(+), 7 deletions(-)

-- 
2.4.1

[dpdk-dev] [PATCH v2 1/4] eal: fix compile on Fedora 22 (GCC 5.1)

[Bruce Richardson]

On Fedora 22, with GCC 5.1, errors are reported due to array accesses
being potentially out of bounds. This commit fixes this by ensuring the
bounds check in the loop takes account of the array size.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 lib/librte_eal/linuxapp/eal/eal_memory.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lib/librte_eal/linuxapp/eal/eal_memory.c b/lib/librte_eal/linuxapp/eal/eal_memory.c
index 5f9f92e..9b8d946 100644
--- a/lib/librte_eal/linuxapp/eal/eal_memory.c
+++ b/lib/librte_eal/linuxapp/eal/eal_memory.c
@@ -1188,7 +1188,9 @@ rte_eal_hugepage_init(void)
 		int socket = tmp_hp[i].socket_id;
 
 		/* find a hugepage info with right size and increment num_pages */
-		for (j = 0; j < (int) internal_config.num_hugepage_sizes; j++) {
+		const int nb_hpsizes = RTE_MIN(MAX_HUGEPAGE_SIZES,
+				(int)internal_config.num_hugepage_sizes);
+		for (j = 0; j < nb_hpsizes; j++) {
 			if (tmp_hp[i].size ==
 					internal_config.hugepage_info[j].hugepage_sz) {
 #ifdef RTE_EAL_SINGLE_FILE_SEGMENTS
-- 
2.4.1

[dpdk-dev] [PATCH v2 2/4] ip_frag: fix compile on Fedora 22 (GCC 5.1)

[Bruce Richardson]

On Fedora 22, with GCC 5.1, errors are reported due to array accesses
being potentially out of bounds. This commit fixes this by adding in an
extra bounds check to the loop counter.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 lib/librte_ip_frag/ip_frag_common.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/librte_ip_frag/ip_frag_common.h b/lib/librte_ip_frag/ip_frag_common.h
index 210f409..6b2acee 100644
--- a/lib/librte_ip_frag/ip_frag_common.h
+++ b/lib/librte_ip_frag/ip_frag_common.h
@@ -90,7 +90,7 @@ static inline int
 ip_frag_key_is_empty(const struct ip_frag_key * key)
 {
 	uint32_t i;
-	for (i = 0; i < key->key_len; i++)
+	for (i = 0; i < RTE_MIN(key->key_len, RTE_DIM(key->src_dst)); i++)
 		if (key->src_dst[i] != 0)
 			return 0;
 	return 1;
-- 
2.4.1

[dpdk-dev] [PATCH v2 3/4] i40e: fix compile on Fedora 22 (GCC 5.1)

[Bruce Richardson]

On Fedora 22, with GCC 5.1, errors are reported due to array accesses
being potentially out of bounds. This commit fixes this by adding in an
extra bounds check to the loop counter.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 drivers/net/i40e/i40e_fdir.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/i40e/i40e_fdir.c b/drivers/net/i40e/i40e_fdir.c
index e688b4f..4bf98d0 100644
--- a/drivers/net/i40e/i40e_fdir.c
+++ b/drivers/net/i40e/i40e_fdir.c
@@ -551,7 +551,7 @@ i40e_set_flx_pld_cfg(struct i40e_pf *pf,
 	memset(flex_pit, 0, sizeof(flex_pit));
 	num = i40e_srcoff_to_flx_pit(cfg->src_offset, flex_pit);
 
-	for (i = 0; i < num; i++) {
+	for (i = 0; i < RTE_MIN(num, RTE_DIM(flex_pit)); i++) {
 		field_idx = layer_idx * I40E_MAX_FLXPLD_FIED + i;
 		/* record the info in fdir structure */
 		pf->fdir.flex_set[field_idx].src_offset =
-- 
2.4.1

[dpdk-dev] [PATCH v2 4/4] null: fix compile on Fedora 22 (GCC 5.1)

[Bruce Richardson]

On Fedora 22, with GCC 5.1, errors are reported due to array accesses
being potentially out of bounds. This commit fixes this by adding in an
extra bounds check to the loop counters, or, in the case of stats reset,
by blindly zeroing the whole array, rather than just the part that is in
use.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 drivers/net/null/rte_eth_null.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/drivers/net/null/rte_eth_null.c b/drivers/net/null/rte_eth_null.c
index 5895065..7792315 100644
--- a/drivers/net/null/rte_eth_null.c
+++ b/drivers/net/null/rte_eth_null.c
@@ -298,7 +298,8 @@ eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *igb_stats)
 	internal = dev->data->dev_private;
 	memset(igb_stats, 0, sizeof(*igb_stats));
 	num_stats = RTE_MIN((unsigned)RTE_ETHDEV_QUEUE_STAT_CNTRS,
-					internal->nb_rx_queues);
+			RTE_MIN(internal->nb_rx_queues,
+				RTE_DIM(internal->rx_null_queues)));
 	for (i = 0; i < num_stats; i++) {
 		igb_stats->q_ipackets[i] =
 			internal->rx_null_queues[i].rx_pkts.cnt;
@@ -306,7 +307,8 @@ eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *igb_stats)
 	}
 
 	num_stats = RTE_MIN((unsigned)RTE_ETHDEV_QUEUE_STAT_CNTRS,
-					internal->nb_tx_queues);
+			RTE_MIN(internal->nb_tx_queues,
+				RTE_DIM(internal->tx_null_queues)));
 	for (i = 0; i < num_stats; i++) {
 		igb_stats->q_opackets[i] =
 			internal->tx_null_queues[i].tx_pkts.cnt;
@@ -331,9 +333,9 @@ eth_stats_reset(struct rte_eth_dev *dev)
 		return;
 
 	internal = dev->data->dev_private;
-	for (i = 0; i < internal->nb_rx_queues; i++)
+	for (i = 0; i < RTE_DIM(internal->rx_null_queues); i++)
 		internal->rx_null_queues[i].rx_pkts.cnt = 0;
-	for (i = 0; i < internal->nb_tx_queues; i++) {
+	for (i = 0; i < RTE_DIM(internal->tx_null_queues); i++) {
 		internal->tx_null_queues[i].tx_pkts.cnt = 0;
 		internal->tx_null_queues[i].err_pkts.cnt = 0;
 	}
-- 
2.4.1

Re: [dpdk-dev] [PATCH v2 0/4] fix compilation on Fedora 22

[Neil Horman]

On Fri, May 29, 2015 at 03:34:12PM +0100, Bruce Richardson wrote:
> V2 Changes: 
> * use RTE_MIN instead of additional comparison checks
>   in some loops, as suggested by Neil.
> * For stats reset in null PMD, just use the size of the
>   array [which is 1] in place of the used queue counts.
> 
> GCC version 5.1 included with Fedora 22 is running stricter array 
> bounds checks which are throwing up errors in a number of components 
> in the DPDK code. This patchset fixes these errors to allow compilation
> with GCC on Fedora 22.
> 
> Example error:
> == Build lib/librte_eal/linuxapp/eal
>   CC eal_memory.o
> /home/bruce/dpdk.org/lib/librte_eal/linuxapp/eal/eal_memory.c: In function 'rte_eal_hugepage_init':
> /home/bruce/dpdk.org/lib/librte_eal/linuxapp/eal/eal_memory.c:1193:35: error: array subscript is above array bounds [-Werror=array-bounds]
>       internal_config.hugepage_info[j].hugepage_sz) {
>                                    ^
> compilation terminated due to -Wfatal-errors.
> 
> Bruce Richardson (4):
>   eal: fix compile on Fedora 22 (GCC 5.1)
>   ip_frag: fix compile on Fedora 22 (GCC 5.1)
>   i40e: fix compile on Fedora 22 (GCC 5.1)
>   null: fix compile on Fedora 22 (GCC 5.1)
> 
>  drivers/net/i40e/i40e_fdir.c             |  2 +-
>  drivers/net/null/rte_eth_null.c          | 12 ++++++++----
>  lib/librte_eal/linuxapp/eal/eal_memory.c |  4 +++-
>  lib/librte_ip_frag/ip_frag_common.h      |  2 +-
>  4 files changed, 13 insertions(+), 7 deletions(-)
> 
> -- 
> 2.4.1
> 
> 

Series
Acked-by: Neil Horman <nhorman@tuxdriver.com>

Re: [dpdk-dev] [PATCH v2 0/4] fix compilation on Fedora 22

[Thomas Monjalon]

2015-05-29 12:57, Neil Horman:
> On Fri, May 29, 2015 at 03:34:12PM +0100, Bruce Richardson wrote:
> > V2 Changes: 
> > * use RTE_MIN instead of additional comparison checks
> >   in some loops, as suggested by Neil.
> > * For stats reset in null PMD, just use the size of the
> >   array [which is 1] in place of the used queue counts.
> > 
> > GCC version 5.1 included with Fedora 22 is running stricter array 
> > bounds checks which are throwing up errors in a number of components 
> > in the DPDK code. This patchset fixes these errors to allow compilation
> > with GCC on Fedora 22.
> > 
> > Example error:
> > == Build lib/librte_eal/linuxapp/eal
> >   CC eal_memory.o
> > /home/bruce/dpdk.org/lib/librte_eal/linuxapp/eal/eal_memory.c: In function 'rte_eal_hugepage_init':
> > /home/bruce/dpdk.org/lib/librte_eal/linuxapp/eal/eal_memory.c:1193:35: error: array subscript is above array bounds [-Werror=array-bounds]
> >       internal_config.hugepage_info[j].hugepage_sz) {
> >                                    ^
> > compilation terminated due to -Wfatal-errors.
> > 
> > Bruce Richardson (4):
> >   eal: fix compile on Fedora 22 (GCC 5.1)
> >   ip_frag: fix compile on Fedora 22 (GCC 5.1)
> >   i40e: fix compile on Fedora 22 (GCC 5.1)
> >   null: fix compile on Fedora 22 (GCC 5.1)
> 
> Series
> Acked-by: Neil Horman <nhorman@tuxdriver.com>

Applied, thanks