My apologies for not paying closer attention during the meeting today, do we now have a plan of action to address querying the MAC address? That is currently a gating item for the IPsec application patch.
From: [email protected] [mailto:[email protected]] On Behalf Of Savolainen, Petri (NSN - FI/Espoo) Sent: Tuesday, August 19, 2014 8:24 AM To: ext Bala Manoharan; Alexandru Badicioiu Cc: [email protected] Subject: Re: [lng-odp] [PATCH 1/1] API support to get mac address of an interface using odp_pktio_t Hi, As Bala mentioned, the intention is that user refers to an interface (by name) only once and that’s in odp_pktio_open(). The resulting handle is used in other packet IO calls after that. We may need to specify two different handle types / APIs: one for managing interface level configuration (cf. physical function, one per interface), and one for packet IO / configuration read access (cf. virtual function, many per interface). The first would be used e.g. to set interface MAC address or promiscuous mode. The second could be used to read current MAC address, etc - in addition to packet IO. -Petri From: [email protected]<mailto:[email protected]> [mailto:[email protected]] On Behalf Of ext Bala Manoharan Sent: Tuesday, August 19, 2014 12:14 PM To: Alexandru Badicioiu Cc: [email protected]<mailto:[email protected]> Subject: Re: [lng-odp] [PATCH 1/1] API support to get mac address of an interface using odp_pktio_t Hi, If we add this additional API odp_get_mac_addr(const char* dev, char* mac_addr) then we need to have a limitation that this API cannot be called in case of bare-metal implementation before calling odp_pktio_open(). The reason I am sceptical about this API is that this provides a means to access the mac_addr of an interface without using the odp_pktio_t handle which defeats the purpose of abstraction. But if there is a consensus on this new API then I can add the same and provide a patch. We can discuss this topic in detail during todays ODP call. Regards, Bala On 19 August 2014 14:19, Alexandru Badicioiu <[email protected]<mailto:[email protected]>> wrote: I understand this bare metal use case where open() call initializes the port , but having the MAC address derived exclusively from a pktio assumes that this is meaningful for any implementation and the implementation has a way to do this. This is why I think there should be also an API call for getting the MAC address from a specific name. Alex On 19 August 2014 11:30, Bala Manoharan <[email protected]<mailto:[email protected]>> wrote: Hi, I agree with your point "const char* dev" need not point to a device, it can point to the location of the physical port which can be defined by the platform. In case of a bare-metal implementation the physical port will have to be initialized and mac address needs to be assigned to the port during initialization. Hence odp_pktio_open() call can be used to initialize the physical port and once the port gets initialized the mac_address of the port can be read using pktio handle. Pls let me know if I am missing some points from your description. Regards, Bala On 19 August 2014 13:37, Alexandru Badicioiu <[email protected]<mailto:[email protected]>> wrote: On 19 August 2014 10:13, Bala Manoharan <[email protected]<mailto:[email protected]>> wrote: Hi Alex, I had the same thought while implementing this API. The reason I decided to go ahead with getting mac address only though odp_pktio_t is because if we have this API to get mac address directly with the device name then the user can try to get mac address of a device even without creating the PKTIO instance for the device. Which could have an issue in the systems where some initialization is done during PKTIO creation. Also the check for validating the device name could be kept in odp_pktio_open() call alone. Maybe I'm not aware of all the cases, but in Linux usually you don't need to bring up an interface (which triggers the netdev_ops open() call) to get the MAC address of the Ethernet port. However, there is no restriction that *dev argument of pktio_open() should denote an Ethernet interface or to be possible to derive a single Ethernet port from it. I think in this case a (const char *dev, char *mac_addr) function is mandatory. Alex Regards, Bala On 19 August 2014 12:37, Alexandru Badicioiu <[email protected]<mailto:[email protected]>> wrote: Hi, from the implementation of the proposed API call it looks like the intent of the API function is to get the MAC address of the device used to open the pktio. From the documentation it looks like the *dev is the actual IO device: /** * Open an ODP packet IO instance * * @param dev Packet IO device * @param pool Pool to use for packet IO * @param params Set of parameters to pass to the arch dependent implementation * * @return ODP packet IO handle or ODP_PKTIO_INVALID on error */ odp_pktio_t odp_pktio_open(const char *dev, odp_buffer_pool_t pool, odp_pktio_params_t *params); Should we have also an API call : int odp_get_mac_addr(const char *dev, unsigned char* mac_addr) ? Thanks, Alex On 19 August 2014 08:56, Balasubramanian Manoharan <[email protected]<mailto:[email protected]>> wrote: Regards, Bala Signed-off-by: Balasubramanian Manoharan <[email protected]<mailto:[email protected]>> --- include/odp_packet_io.h | 8 +++++++ platform/linux-generic/include/odp_packet_netmap.h | 1 + platform/linux-generic/odp_packet_io.c | 28 +++++++++++++++++++++- platform/linux-generic/odp_packet_netmap.c | 4 +++- platform/linux-generic/odp_packet_socket.c | 8 +++---- 5 files changed, 43 insertions(+), 6 deletions(-) diff --git a/include/odp_packet_io.h b/include/odp_packet_io.h index cfefac0..6c06520 100644 --- a/include/odp_packet_io.h +++ b/include/odp_packet_io.h @@ -129,6 +129,14 @@ void odp_pktio_set_input(odp_packet_t pkt, odp_pktio_t id); */ odp_pktio_t odp_pktio_get_input(odp_packet_t pkt); +/** + * Get mac address of the interface + * + * @param id ODP packet IO handle + * @param mac_addr Storage for Mac address of the packet IO interface (filled by function) + * @return 0 on success or -1 on error +**/ +int odp_pktio_get_mac_addr(odp_pktio_t id, unsigned char* mac_addr); #ifdef __cplusplus } #endif diff --git a/platform/linux-generic/include/odp_packet_netmap.h b/platform/linux-generic/include/odp_packet_netmap.h index 57d9f2c..3693416 100644 --- a/platform/linux-generic/include/odp_packet_netmap.h +++ b/platform/linux-generic/include/odp_packet_netmap.h @@ -40,6 +40,7 @@ typedef struct { odp_queue_t tx_access; /* Used for exclusive access to send packets */ uint32_t if_flags; char ifname[32]; + unsigned char if_mac[ETH_ALEN]; } pkt_netmap_t; /** diff --git a/platform/linux-generic/odp_packet_io.c b/platform/linux-generic/odp_packet_io.c index 33ade10..069b0e1 100644 --- a/platform/linux-generic/odp_packet_io.c +++ b/platform/linux-generic/odp_packet_io.c @@ -245,7 +245,7 @@ odp_pktio_t odp_pktio_open(const char *dev, odp_buffer_pool_t pool, ODP_ERR("This type of I/O is not supported. Please recompile.\n"); break; } - + pktio_entry->s.params.type = params->type; unlock_entry(pktio_entry); return id; } @@ -535,3 +535,29 @@ int pktin_deq_multi(queue_entry_t *qentry, odp_buffer_hdr_t *buf_hdr[], int num) return nbr; } +int odp_pktio_get_mac_addr(odp_pktio_t pkt, unsigned char* mac_addr){ + + pktio_entry_t * pktio_entry = get_entry(pkt); + if(!pktio_entry){ + printf("Invalid odp_pktio_t value\n"); + return ODP_PKTIO_INVALID; + } + switch(pktio_entry->s.params.type){ + case ODP_PKTIO_TYPE_SOCKET_BASIC: + case ODP_PKTIO_TYPE_SOCKET_MMSG: + memcpy(mac_addr, pktio_entry->s.pkt_sock.if_mac, ETH_ALEN); + break; + case ODP_PKTIO_TYPE_SOCKET_MMAP: + memcpy(mac_addr, pktio_entry->s.pkt_sock_mmap.if_mac, ETH_ALEN); + break; +#ifdef ODP_HAVE_NETMAP + case ODP_PKTIO_TYPE_NETMAP: + memcpy(mac_addr, pktio_entry->s.pkt_nm.if_mac, ETH_ALEN); + break; +#endif + default: + ODP_ERR("Invalid pktio type: %02x\n", pktio_entry->s.params.type); + return ODP_PKTIO_INVALID; + } + return 0; +} diff --git a/platform/linux-generic/odp_packet_netmap.c b/platform/linux-generic/odp_packet_netmap.c index e2215ab..e9e9f56 100644 --- a/platform/linux-generic/odp_packet_netmap.c +++ b/platform/linux-generic/odp_packet_netmap.c @@ -222,9 +222,11 @@ int setup_pkt_netmap(pkt_netmap_t * const pkt_nm, const char *netdev, ODP_ERR("Error: token creation failed\n"); return -1; } + if( 0 != socket_store_hw_addr(pkt_nm->if_mac, netdev) + return -1; odp_queue_enq(pkt_nm->tx_access, token); - + ODP_DBG("Wait for link to come up\n"); sleep(WAITLINK_TMO); ODP_DBG("Done\n"); diff --git a/platform/linux-generic/odp_packet_socket.c b/platform/linux-generic/odp_packet_socket.c index d44c333..d96aa8e 100644 --- a/platform/linux-generic/odp_packet_socket.c +++ b/platform/linux-generic/odp_packet_socket.c @@ -735,7 +735,7 @@ static int mmap_bind_sock(pkt_sock_mmap_t *pkt_sock, const char *netdev) return 0; } -static int mmap_store_hw_addr(pkt_sock_mmap_t * const pkt_sock, +static int socket_store_hw_addr(int sockfd,unsigned char * if_mac, const char *netdev) { struct ifreq ethreq; @@ -744,13 +744,13 @@ static int mmap_store_hw_addr(pkt_sock_mmap_t * const pkt_sock, /* get MAC address */ memset(ðreq, 0, sizeof(ethreq)); strncpy(ethreq.ifr_name, netdev, IFNAMSIZ); - ret = ioctl(pkt_sock->sockfd, SIOCGIFHWADDR, ðreq); + ret = ioctl(sockfd, SIOCGIFHWADDR, ðreq); if (ret != 0) { perror("store_hw_addr() - ioctl(SIOCGIFHWADDR)"); return -1; } - ethaddr_copy(pkt_sock->if_mac, + ethaddr_copy(if_mac, (unsigned char *)ethreq.ifr_ifru.ifru_hwaddr.sa_data); return 0; @@ -805,7 +805,7 @@ int setup_pkt_sock_mmap(pkt_sock_mmap_t * const pkt_sock, const char *netdev, if (ret != 0) return -1; - ret = mmap_store_hw_addr(pkt_sock, netdev); + ret = socket_store_hw_addr(pkt_sock->sockfd, pkt_sock->if_mac, netdev); if (ret != 0) return -1; -- 2.0.1.472.g6f92e5f _______________________________________________ lng-odp mailing list [email protected]<mailto:[email protected]> http://lists.linaro.org/mailman/listinfo/lng-odp
_______________________________________________ lng-odp mailing list [email protected] http://lists.linaro.org/mailman/listinfo/lng-odp
