* [PATCH net-next 2/5] bnx2x: Fix stop-on-error
From: Yuval Mintz @ 2014-08-25 14:48 UTC (permalink / raw)
To: davem, netdev; +Cc: Ariel.Elior, Yuval Mintz
In-Reply-To: <1408978113-27902-1-git-send-email-Yuval.Mintz@qlogic.com>
When STOP_ON_ERROR is set driver will not compile. Even if it did,
traffic will not pass without this patch as several fields which are
verified by FW/HW on the Tx path are not properly set.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: Ariel Elior <Ariel.Elior@qlogic.com>
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 25 +++++++++++++++++++-----
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 2 +-
2 files changed, 21 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index 2388144..2f210ba 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -3873,12 +3873,16 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
/* when transmitting in a vf, start bd must hold the ethertype
* for fw to enforce it
*/
+#ifndef BNX2X_STOP_ON_ERROR
if (IS_VF(bp))
+#endif
tx_start_bd->vlan_or_ethertype =
cpu_to_le16(ntohs(eth->h_proto));
+#ifndef BNX2X_STOP_ON_ERROR
else
/* used by FW for packet accounting */
tx_start_bd->vlan_or_ethertype = cpu_to_le16(pkt_prod);
+#endif
}
nbd = 2; /* start_bd + pbd + frags (updated when pages are mapped) */
@@ -3951,11 +3955,22 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
&pbd_e2->data.mac_addr.dst_mid,
&pbd_e2->data.mac_addr.dst_lo,
eth->h_dest);
- } else if (bp->flags & TX_SWITCHING) {
- bnx2x_set_fw_mac_addr(&pbd_e2->data.mac_addr.dst_hi,
- &pbd_e2->data.mac_addr.dst_mid,
- &pbd_e2->data.mac_addr.dst_lo,
- eth->h_dest);
+ } else {
+ if (bp->flags & TX_SWITCHING)
+ bnx2x_set_fw_mac_addr(
+ &pbd_e2->data.mac_addr.dst_hi,
+ &pbd_e2->data.mac_addr.dst_mid,
+ &pbd_e2->data.mac_addr.dst_lo,
+ eth->h_dest);
+#ifdef BNX2X_STOP_ON_ERROR
+ /* Enforce security is always set in Stop on Error -
+ * source mac should be present in the parsing BD
+ */
+ bnx2x_set_fw_mac_addr(&pbd_e2->data.mac_addr.src_hi,
+ &pbd_e2->data.mac_addr.src_mid,
+ &pbd_e2->data.mac_addr.src_lo,
+ eth->h_source);
+#endif
}
SET_FLAG(pbd_e2_parsing_data,
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index a008f48..cf5b3b4 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -1142,7 +1142,7 @@ void bnx2x_panic_dump(struct bnx2x *bp, bool disable_int)
if (!fp->txdata_ptr)
break;
- if (!txdata.tx_cons_sb)
+ if (!txdata->tx_cons_sb)
continue;
start = TX_BD(le16_to_cpu(*txdata->tx_cons_sb) - 10);
--
1.8.3.1
^ permalink raw reply related
* [PATCH net-next 0/5] bnx2x: `fixes' patch-series
From: Yuval Mintz @ 2014-08-25 14:48 UTC (permalink / raw)
To: davem, netdev; +Cc: Ariel.Elior, Yuval Mintz
Hi,
This series contains mostly bug fixes, but never the less is intended
for `net-next' and not `net', as:
- Some of the fixes are quite insignificant [`VF clean statistics',
`ethtool -d might cause timeout in log'].
- Some only recently were submitted to `net-next' [`Fix timesync endianity'].
- Some are not usually compiled as part of the kernel [`Fix stop-on-error'].
Dave - please consider applying this series to `net-next'; If you prefer,
I can break this series into 2 parts [one for `net' and the other for
`net-next'] - but personally I don't see much benefit in it.
Thanks,
Yuval Mintz
Dmitry Kravkov (2):
bnx2x: ethtool -d might cause timeout in log
bnx2x: Be more forgiving toward SW GRO
Michal Kalderon (1):
bnx2x: Fix timesync endianity
Yuval Mintz (2):
bnx2x: Fix stop-on-error
bnx2x: VF clean statistics
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 35 +++++++++++++++++++----
drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h | 14 ++++-----
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 2 +-
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c | 6 ++--
drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.c | 5 ++++
5 files changed, 44 insertions(+), 18 deletions(-)
--
1.8.3.1
^ permalink raw reply
* [PATCH net-next 1/5] bnx2x: ethtool -d might cause timeout in log
From: Yuval Mintz @ 2014-08-25 14:48 UTC (permalink / raw)
To: davem, netdev; +Cc: Ariel.Elior, Dmitry Kravkov, Yuval Mintz
In-Reply-To: <1408978113-27902-1-git-send-email-Yuval.Mintz@qlogic.com>
From: Dmitry Kravkov <Dmitry.Kravkov@qlogic.com>
This changes slightly the set of registers read during `ethtool -d'.
Without this change, it's possible the HW will generate a grc Attention which
will be logged into system logs as `grc timeout'.
Signed-off-by: Dmitry Kravkov <Dmitry.Kravkov@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: Ariel Elior <Ariel.Elior@qlogic.com>
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h | 14 +++++---------
1 file changed, 5 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h
index 12eb4ba..741aa13 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h
@@ -40,7 +40,7 @@ struct dump_header {
u32 dump_meta_data; /* OR of CHIP and PATH. */
};
-#define BNX2X_DUMP_VERSION 0x50acff01
+#define BNX2X_DUMP_VERSION 0x61111111
struct reg_addr {
u32 addr;
u32 size;
@@ -1464,7 +1464,6 @@ static const struct reg_addr reg_addrs[] = {
{ 0x180398, 1, 0x1c, 0x924},
{ 0x1803a0, 5, 0x1c, 0x924},
{ 0x1803b4, 2, 0x18, 0x924},
- { 0x180400, 256, 0x3, 0xfff},
{ 0x181000, 4, 0x1f, 0x93c},
{ 0x181010, 1020, 0x1f, 0x38},
{ 0x182000, 4, 0x18, 0x924},
@@ -1576,7 +1575,6 @@ static const struct reg_addr reg_addrs[] = {
{ 0x200398, 1, 0x1c, 0x924},
{ 0x2003a0, 1, 0x1c, 0x924},
{ 0x2003a8, 2, 0x1c, 0x924},
- { 0x200400, 256, 0x3, 0xfff},
{ 0x202000, 4, 0x1f, 0x1927},
{ 0x202010, 2044, 0x1f, 0x1007},
{ 0x204000, 4, 0x18, 0x924},
@@ -1688,7 +1686,6 @@ static const struct reg_addr reg_addrs[] = {
{ 0x280398, 1, 0x1c, 0x924},
{ 0x2803a0, 1, 0x1c, 0x924},
{ 0x2803a8, 2, 0x1c, 0x924},
- { 0x280400, 256, 0x3, 0xfff},
{ 0x282000, 4, 0x1f, 0x9e4},
{ 0x282010, 2044, 0x1f, 0x1c0},
{ 0x284000, 4, 0x18, 0x924},
@@ -1800,7 +1797,6 @@ static const struct reg_addr reg_addrs[] = {
{ 0x300398, 1, 0x1c, 0x924},
{ 0x3003a0, 1, 0x1c, 0x924},
{ 0x3003a8, 2, 0x1c, 0x924},
- { 0x300400, 256, 0x3, 0xfff},
{ 0x302000, 4, 0x1f, 0xf24},
{ 0x302010, 2044, 0x1f, 0xe00},
{ 0x304000, 4, 0x18, 0x924},
@@ -2206,10 +2202,10 @@ static const struct wreg_addr wreg_addr_e3b0 = {
0x1b0c00, 128, 2, read_reg_e3b0, 0x1f, 0x1fff};
static const unsigned int dump_num_registers[NUM_CHIPS][NUM_PRESETS] = {
- {20782, 18567, 27975, 19729, 18311, 27719, 20836, 32391, 41799, 20812,
- 26247, 35655, 19074},
- {32774, 19297, 33277, 31721, 19041, 33021, 32828, 33121, 47101, 32804,
- 26977, 40957, 35895},
+ {19758, 17543, 26951, 18705, 17287, 26695, 19812, 31367, 40775, 19788,
+ 25223, 34631, 19074},
+ {31750, 18273, 32253, 30697, 18017, 31997, 31804, 32097, 46077, 31780,
+ 25953, 39933, 35895},
{36527, 17928, 33697, 35474, 18700, 34466, 36581, 31752, 47521, 36557,
25608, 41377, 43903},
{45239, 17936, 34387, 44186, 18708, 35156, 45293, 31760, 48211, 45269,
--
1.8.3.1
^ permalink raw reply related
* [PATCH v2 8/8] qlge: Fix TSO for non-accelerated vlan traffic
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev
Cc: Vladislav Yasevich, Shahed Shaikh, Jitendra Kalsaria, Ron Mercer,
linux-driver
In-Reply-To: <1408977295-9162-1-git-send-email-vyasevic@redhat.com>
This device claims TSO support for vlans. It also allows a user to
control vlan acceleration offloading. As such, it is possible to turn
off vlan acceleration and configure a vlan which will continue to send
TSO traffic.
In such situation the packet passed down the the device will contain
a vlan header and skb->protocol will be set to ETH_P_8021Q.
The device assumes that skb->protocol contains network protocol
value and uses that value to set up TSO information.
This results in corrupted frames sent on the wire.
This patch extracts the protocol value correctly by using a
vlan_get_protocol() helper and corrects corrupt TSO frames.
CC: Shahed Shaikh <shahed.shaikh@qlogic.com>
CC: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
CC: Ron Mercer <ron.mercer@qlogic.com>
CC: linux-driver@qlogic.com
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
drivers/net/ethernet/qlogic/qlge/qlge_main.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/qlogic/qlge/qlge_main.c b/drivers/net/ethernet/qlogic/qlge/qlge_main.c
index 188626e..3e96f26 100644
--- a/drivers/net/ethernet/qlogic/qlge/qlge_main.c
+++ b/drivers/net/ethernet/qlogic/qlge/qlge_main.c
@@ -2556,6 +2556,7 @@ static int ql_tso(struct sk_buff *skb, struct ob_mac_tso_iocb_req *mac_iocb_ptr)
if (skb_is_gso(skb)) {
int err;
+ __be16 l3_proto = vlan_get_protocol(skb);
err = skb_cow_head(skb, 0);
if (err < 0)
@@ -2572,7 +2573,7 @@ static int ql_tso(struct sk_buff *skb, struct ob_mac_tso_iocb_req *mac_iocb_ptr)
<< OB_MAC_TRANSPORT_HDR_SHIFT);
mac_iocb_ptr->mss = cpu_to_le16(skb_shinfo(skb)->gso_size);
mac_iocb_ptr->flags2 |= OB_MAC_TSO_IOCB_LSO;
- if (likely(skb->protocol == htons(ETH_P_IP))) {
+ if (likely(l3_proto == htons(ETH_P_IP))) {
struct iphdr *iph = ip_hdr(skb);
iph->check = 0;
mac_iocb_ptr->flags1 |= OB_MAC_TSO_IOCB_IP4;
@@ -2580,7 +2581,7 @@ static int ql_tso(struct sk_buff *skb, struct ob_mac_tso_iocb_req *mac_iocb_ptr)
iph->daddr, 0,
IPPROTO_TCP,
0);
- } else if (skb->protocol == htons(ETH_P_IPV6)) {
+ } else if (l3_proto == htons(ETH_P_IPV6)) {
mac_iocb_ptr->flags1 |= OB_MAC_TSO_IOCB_IP6;
tcp_hdr(skb)->check =
~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
--
1.9.3
^ permalink raw reply related
* [PATCH v2 7/8] mvneta: Fix TSO and checksum for non-acceleration vlan traffic
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev; +Cc: Vladislav Yasevich, Thomas Petazzoni
In-Reply-To: <1408977295-9162-1-git-send-email-vyasevic@redhat.com>
This driver doesn't appear to support vlan acceleration at
all. However, it does claim to support TSO and IP checksums
for vlan devices. Thus any configured vlan device would
end up passing down partial checksums or TSO frames.
The driver also uses the value from skb->protocol to
determine TSO and checksum offload information, but assumes
that skb->protocol holds the l3 protocol information.
As a result, vlan traffic with partial checksums or TSO
will fail those checks and TSO will not happen.
Fix this by using vlan_get_protocol() helper.
CC: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
drivers/net/ethernet/marvell/mvneta.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index c9f1d1b..133f8c6 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -1371,15 +1371,16 @@ static u32 mvneta_skb_tx_csum(struct mvneta_port *pp, struct sk_buff *skb)
{
if (skb->ip_summed == CHECKSUM_PARTIAL) {
int ip_hdr_len = 0;
+ __be16 l3_proto = vlan_get_protocol(skb);
u8 l4_proto;
- if (skb->protocol == htons(ETH_P_IP)) {
+ if (l3_proto == htons(ETH_P_IP)) {
struct iphdr *ip4h = ip_hdr(skb);
/* Calculate IPv4 checksum and L4 checksum */
ip_hdr_len = ip4h->ihl;
l4_proto = ip4h->protocol;
- } else if (skb->protocol == htons(ETH_P_IPV6)) {
+ } else if (l3_proto == htons(ETH_P_IPV6)) {
struct ipv6hdr *ip6h = ipv6_hdr(skb);
/* Read l4_protocol from one of IPv6 extra headers */
@@ -1390,7 +1391,7 @@ static u32 mvneta_skb_tx_csum(struct mvneta_port *pp, struct sk_buff *skb)
return MVNETA_TX_L4_CSUM_NOT;
return mvneta_txq_desc_csum(skb_network_offset(skb),
- skb->protocol, ip_hdr_len, l4_proto);
+ l3_proto, ip_hdr_len, l4_proto);
}
return MVNETA_TX_L4_CSUM_NOT;
--
1.9.3
^ permalink raw reply related
* [PATCH v2 4/8] ehea: Fix TSO and hw checksums with non-accelerated vlan packets.
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev; +Cc: Vladislav Yasevich, Thadeu Lima de Souza Cascardo
In-Reply-To: <1408977295-9162-1-git-send-email-vyasevic@redhat.com>
The driver claims that it can do TSO and IP checksums on vlan
devices and also allows user to control vlan acceleration offloading.
This makes it possible to push traffic to this driver that has TSO or
partial checksums set, but also have a non-accelearted vlan
header. In this case, the driver will fail to correctly
identify such traffic and will not correctly perform
segmentation and checksum calculation.
Fix this by using vlan_get_protocol() helper instead of
assuming skb->protocol always has this information.
CC: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
drivers/net/ethernet/ibm/ehea/ehea_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/ibm/ehea/ehea_main.c b/drivers/net/ethernet/ibm/ehea/ehea_main.c
index a0b418e..566b17d 100644
--- a/drivers/net/ethernet/ibm/ehea/ehea_main.c
+++ b/drivers/net/ethernet/ibm/ehea/ehea_main.c
@@ -1994,7 +1994,7 @@ static void xmit_common(struct sk_buff *skb, struct ehea_swqe *swqe)
{
swqe->tx_control |= EHEA_SWQE_IMM_DATA_PRESENT | EHEA_SWQE_CRC;
- if (skb->protocol != htons(ETH_P_IP))
+ if (vlan_get_protocol(skb) != htons(ETH_P_IP))
return;
if (skb->ip_summed == CHECKSUM_PARTIAL)
--
1.9.3
^ permalink raw reply related
* [PATCH v2 3/8] bna: Support TSO and partial checksum with non-accelerated vlans.
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev; +Cc: Vladislav Yasevich, Rasesh Mody
In-Reply-To: <1408977295-9162-1-git-send-email-vyasevic@redhat.com>
This device claims TSO and checksum support for vlans. It also
allows a user to control vlan acceleration offloading. As such,
it is possible to turn off vlan acceleration and configure a vlan
which will continue to support TSO.
In such situation the packet passed down the the device will contain
a vlan header and skb->protocol will be set to ETH_P_8021Q.
The device assumes that skb->protocol contains network protocol
value and uses that value to set up TSO information. This results
in corrupted frames sent on the wire.
This patch extract the protocol value correctly and corrects TSO
and checksums for non-accelerated traffic.
CC: Rasesh Mody <rmody@brocade.com>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
drivers/net/ethernet/brocade/bna/bnad.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/brocade/bna/bnad.c b/drivers/net/ethernet/brocade/bna/bnad.c
index ff8cae5..ffc92a4 100644
--- a/drivers/net/ethernet/brocade/bna/bnad.c
+++ b/drivers/net/ethernet/brocade/bna/bnad.c
@@ -2506,7 +2506,7 @@ bnad_tso_prepare(struct bnad *bnad, struct sk_buff *skb)
* For TSO, the TCP checksum field is seeded with pseudo-header sum
* excluding the length field.
*/
- if (skb->protocol == htons(ETH_P_IP)) {
+ if (vlan_get_protocol(skb) == htons(ETH_P_IP)) {
struct iphdr *iph = ip_hdr(skb);
/* Do we really need these? */
@@ -2870,12 +2870,13 @@ bnad_txq_wi_prepare(struct bnad *bnad, struct bna_tcb *tcb,
}
if (skb->ip_summed == CHECKSUM_PARTIAL) {
+ __be16 net_proto = vlan_get_protocol(skb);
u8 proto = 0;
- if (skb->protocol == htons(ETH_P_IP))
+ if (net_proto == htons(ETH_P_IP))
proto = ip_hdr(skb)->protocol;
#ifdef NETIF_F_IPV6_CSUM
- else if (skb->protocol == htons(ETH_P_IPV6)) {
+ else if (net_proto == htons(ETH_P_IPV6)) {
/* nexthdr may not be TCP immediately. */
proto = ipv6_hdr(skb)->nexthdr;
}
--
1.9.3
^ permalink raw reply related
* [PATCH v2 0/8] Fix TSO and checksum issues with non-accelerated vlan traffic.
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev
Cc: Vladislav Yasevich, Jeff Kirsher, Jesse Brandeburg, Bruce Allan,
Carolyn Wyborny, Don Skidmore, Greg Rose, Alex Duyck,
John Ronciak, Mitch Williams, Linux NICS, e1000-devel,
Rasesh Mody, Thadeu Lima de Souza Cascardo, Thomas Petazzoni,
Shahed Shaikh, Jitendra Kalsaria, Ron Mercer, linux-driver,
Chris Metcalf
I've recently ran across something rather interesting when testing vlans
from inside VMs. In some scenarios I was getting awfull thruput.
Some debugging uncovered a very scary packet corruption. I was
seeing packets that had original TSO length as IP total length
and their ip checksum was 0. This was with e1000e driver.
A bit more debugging uncovered an assumption made by that driver
that skb->protocol will contain l3 protocol information. This
was not the case in my setup since I manually turned off vlan
tx acceleration for the device. This caused the driver to not
initialize the tso information correctly and resulted in
corrupt TSO frames on the wire.
I decided to do some auditing of the usage of skb->protocols
in the drivers. Out of all the drivers, the included 8 appear
to be effected. They all allow user to control vlan acceleration
settings, all support TSO on vlan devices, and all use
skb->protocol to decide how to encode TSO information. Some
also have similar problems when initializing hw checksum information.
On such device, it is simple enough to reproduce the issue.
Simply turn off TX VLAN acceleration on the device, create a vlan,
and run you favorite network performance tool.
There is 1 driver I ran across that I belive will trigger a BUG
in the system when used with vlans. That driver is tile/tilepro.c
I have not changed it in this patch set and would hope that
the maintainer has time to look at it.
V2: Fix i40ev using the wrong function name. Full build.
Thanks
-vlad
Vladislav Yasevich (8):
e1000e: Fix TSO with non-accelerated vlans
e1000: Fix TSO for non-accelerated vlan traffic
bna: Support TSO and partial checksum with non-accelerated vlans.
ehea: Fix TSO and hw checksums with non-accelerated vlan packets.
i40e: Fix TSO and hw checksums for non-accelerated vlan packets.
i40evf: Fix TSO and hw checksums for non-accelerated vlan packets.
mvneta: Fix TSO and checksum for non-acceleration vlan traffic
qlge: Fix TSO for non-accelerated vlan traffic
drivers/net/ethernet/brocade/bna/bnad.c | 7 ++++---
drivers/net/ethernet/ibm/ehea/ehea_main.c | 2 +-
drivers/net/ethernet/intel/e1000/e1000_main.c | 19 +++++++++++--------
drivers/net/ethernet/intel/e1000e/netdev.c | 21 +++++++++------------
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 2 +-
drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 2 +-
drivers/net/ethernet/marvell/mvneta.c | 7 ++++---
drivers/net/ethernet/qlogic/qlge/qlge_main.c | 5 +++--
8 files changed, 34 insertions(+), 31 deletions(-)
--
1.9.3
^ permalink raw reply
* [PATCH v2 6/8] i40evf: Fix TSO and hw checksums for non-accelerated vlan packets.
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev
Cc: Vladislav Yasevich, Linux NICS, e1000-devel, Bruce Allan,
Jesse Brandeburg, John Ronciak
In-Reply-To: <1408977295-9162-1-git-send-email-vyasevic@redhat.com>
This device claims TSO and checksum support for vlans. It also
allows a user to control vlan acceleration offloading. As such,
it is possible to turn off vlan acceleration and configure a vlan
which will continue to support TSO and hw checksums.
In such situation the packet passed down the the device will contain
a vlan header and skb->protocol will be set to ETH_P_8021Q.
The device assumes that skb->protocol contains network protocol
value and uses that value to set up TSO and checksum information.
This results in corrupted frames sent on the wire.
This patch extract the protocol value correctly and corrects TSO
and checksums for non-accelerated traffic.
Fix this by using vlan_get_protocol() helper.
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
CC: Jesse Brandeburg <jesse.brandeburg@intel.com>
CC: Bruce Allan <bruce.w.allan@intel.com>
CC: Carolyn Wyborny <carolyn.wyborny@intel.com>
CC: Don Skidmore <donald.c.skidmore@intel.com>
CC: Greg Rose <gregory.v.rose@intel.com>
CC: Alex Duyck <alexander.h.duyck@intel.com>
CC: John Ronciak <john.ronciak@intel.com>
CC: Mitch Williams <mitch.a.williams@intel.com>
CC: Linux NICS <linux.nics@intel.com>
CC: e1000-devel@lists.sourceforge.net
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index 79bf96c..95a3ec2 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -1597,7 +1597,7 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb,
goto out_drop;
/* obtain protocol of skb */
- protocol = skb->protocol;
+ protocol = vlan_get_protocol(skb);
/* record the location of the first descriptor for this packet */
first = &tx_ring->tx_bi[tx_ring->next_to_use];
--
1.9.3
------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
^ permalink raw reply related
* [PATCH v2 5/8] i40e: Fix TSO and hw checksums for non-accelerated vlan packets.
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev
Cc: Vladislav Yasevich, Linux NICS, e1000-devel, Bruce Allan,
Jesse Brandeburg, John Ronciak
In-Reply-To: <1408977295-9162-1-git-send-email-vyasevic@redhat.com>
This device claims TSO and checksum support for vlans. It also
allows a user to control vlan acceleration offloading. As such,
it is possible to turn off vlan acceleration and configure a vlan
which will continue to support TSO and hw checksums.
In such situation the packet passed down the the device will contain
a vlan header and skb->protocol will be set to ETH_P_8021Q.
The device assumes that skb->protocol contains network protocol
value and uses that value to set up TSO and checksum information.
This results in corrupted frames sent on the wire.
This patch extract the protocol value correctly and corrects TSO
and checksums for non-accelerated traffic.
Fix this by using vlan_get_protocol() helper.
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
CC: Jesse Brandeburg <jesse.brandeburg@intel.com>
CC: Bruce Allan <bruce.w.allan@intel.com>
CC: Carolyn Wyborny <carolyn.wyborny@intel.com>
CC: Don Skidmore <donald.c.skidmore@intel.com>
CC: Greg Rose <gregory.v.rose@intel.com>
CC: Alex Duyck <alexander.h.duyck@intel.com>
CC: John Ronciak <john.ronciak@intel.com>
CC: Mitch Williams <mitch.a.williams@intel.com>
CC: Linux NICS <linux.nics@intel.com>
CC: e1000-devel@lists.sourceforge.net
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index a51aa37..369848e 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -2295,7 +2295,7 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb,
goto out_drop;
/* obtain protocol of skb */
- protocol = skb->protocol;
+ protocol = vlan_get_protocol(skb);
/* record the location of the first descriptor for this packet */
first = &tx_ring->tx_bi[tx_ring->next_to_use];
--
1.9.3
------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
^ permalink raw reply related
* [PATCH v2 2/8] e1000: Fix TSO for non-accelerated vlan traffic
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev
Cc: Vladislav Yasevich, Linux NICS, e1000-devel, Bruce Allan,
Jesse Brandeburg, John Ronciak
In-Reply-To: <1408977295-9162-1-git-send-email-vyasevic@redhat.com>
This device claims TSO and checksum support for vlans. It also
allows a user to control vlan acceleration offloading. As such,
it is possible to turn off vlan acceleration and configure a vlan
which will continue to support TSO.
In such situation the packet passed down the the device will contain
a vlan header and skb->protocol will be set to ETH_P_8021Q.
The device assumes that skb->protocol contains network protocol
value and uses that value to set up TSO and checksum information.
This will results in corrupted frames sent on the wire.
This patch extract the protocol value correctly and corrects TSO
for non-accelerated traffic.
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
CC: Jesse Brandeburg <jesse.brandeburg@intel.com>
CC: Bruce Allan <bruce.w.allan@intel.com>
CC: Carolyn Wyborny <carolyn.wyborny@intel.com>
CC: Don Skidmore <donald.c.skidmore@intel.com>
CC: Greg Rose <gregory.v.rose@intel.com>
CC: Alex Duyck <alexander.h.duyck@intel.com>
CC: John Ronciak <john.ronciak@intel.com>
CC: Mitch Williams <mitch.a.williams@intel.com>
CC: Linux NICS <linux.nics@intel.com>
CC: e1000-devel@lists.sourceforge.net
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
drivers/net/ethernet/intel/e1000/e1000_main.c | 19 +++++++++++--------
1 file changed, 11 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c
index cbc330b..ad3d5d1 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_main.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_main.c
@@ -2674,7 +2674,8 @@ set_itr_now:
#define E1000_TX_FLAGS_VLAN_SHIFT 16
static int e1000_tso(struct e1000_adapter *adapter,
- struct e1000_tx_ring *tx_ring, struct sk_buff *skb)
+ struct e1000_tx_ring *tx_ring, struct sk_buff *skb,
+ __be16 protocol)
{
struct e1000_context_desc *context_desc;
struct e1000_buffer *buffer_info;
@@ -2692,7 +2693,7 @@ static int e1000_tso(struct e1000_adapter *adapter,
hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
mss = skb_shinfo(skb)->gso_size;
- if (skb->protocol == htons(ETH_P_IP)) {
+ if (protocol == htons(ETH_P_IP)) {
struct iphdr *iph = ip_hdr(skb);
iph->tot_len = 0;
iph->check = 0;
@@ -2702,7 +2703,7 @@ static int e1000_tso(struct e1000_adapter *adapter,
0);
cmd_length = E1000_TXD_CMD_IP;
ipcse = skb_transport_offset(skb) - 1;
- } else if (skb->protocol == htons(ETH_P_IPV6)) {
+ } else if (skb_is_gso_v6(skb)) {
ipv6_hdr(skb)->payload_len = 0;
tcp_hdr(skb)->check =
~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
@@ -2745,7 +2746,8 @@ static int e1000_tso(struct e1000_adapter *adapter,
}
static bool e1000_tx_csum(struct e1000_adapter *adapter,
- struct e1000_tx_ring *tx_ring, struct sk_buff *skb)
+ struct e1000_tx_ring *tx_ring, struct sk_buff *skb,
+ __be16 protocol)
{
struct e1000_context_desc *context_desc;
struct e1000_buffer *buffer_info;
@@ -2756,7 +2758,7 @@ static bool e1000_tx_csum(struct e1000_adapter *adapter,
if (skb->ip_summed != CHECKSUM_PARTIAL)
return false;
- switch (skb->protocol) {
+ switch (protocol) {
case cpu_to_be16(ETH_P_IP):
if (ip_hdr(skb)->protocol == IPPROTO_TCP)
cmd_len |= E1000_TXD_CMD_TCP;
@@ -3097,6 +3099,7 @@ static netdev_tx_t e1000_xmit_frame(struct sk_buff *skb,
int count = 0;
int tso;
unsigned int f;
+ __be16 protocol = vlan_get_protocol(skb);
/* This goes back to the question of how to logically map a Tx queue
* to a flow. Right now, performance is impacted slightly negatively
@@ -3210,7 +3213,7 @@ static netdev_tx_t e1000_xmit_frame(struct sk_buff *skb,
first = tx_ring->next_to_use;
- tso = e1000_tso(adapter, tx_ring, skb);
+ tso = e1000_tso(adapter, tx_ring, skb, protocol);
if (tso < 0) {
dev_kfree_skb_any(skb);
return NETDEV_TX_OK;
@@ -3220,10 +3223,10 @@ static netdev_tx_t e1000_xmit_frame(struct sk_buff *skb,
if (likely(hw->mac_type != e1000_82544))
tx_ring->last_tx_tso = true;
tx_flags |= E1000_TX_FLAGS_TSO;
- } else if (likely(e1000_tx_csum(adapter, tx_ring, skb)))
+ } else if (likely(e1000_tx_csum(adapter, tx_ring, skb, protocol)))
tx_flags |= E1000_TX_FLAGS_CSUM;
- if (likely(skb->protocol == htons(ETH_P_IP)))
+ if (protocol == htons(ETH_P_IP))
tx_flags |= E1000_TX_FLAGS_IPV4;
if (unlikely(skb->no_fcs))
--
1.9.3
------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
^ permalink raw reply related
* [PATCH v2 1/8] e1000e: Fix TSO with non-accelerated vlans
From: Vladislav Yasevich @ 2014-08-25 14:34 UTC (permalink / raw)
To: netdev
Cc: Vladislav Yasevich, Linux NICS, e1000-devel, Bruce Allan,
Jesse Brandeburg, John Ronciak
In-Reply-To: <1408977295-9162-1-git-send-email-vyasevic@redhat.com>
This device claims TSO support for vlans. It also allows a
user to control vlan acceleration offloading. As such, it is
possible to turn off vlan acceleration and configure a vlan
which will continue to support TSO.
In such situation the packet passed down the the device will contain
a vlan header and skb->protocol will be set to ETH_P_8021Q.
The device assumes that skb->protocol contains network protocol
value and uses that value to set up TSO information. This results
in corrupted frames sent on the wire. Corruptions include
incorrect IP total length and invalid IP checksum.
This patch extract the protocol value correctly and corrects TSO
for non-accelerated traffic.
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
CC: Jesse Brandeburg <jesse.brandeburg@intel.com>
CC: Bruce Allan <bruce.w.allan@intel.com>
CC: Carolyn Wyborny <carolyn.wyborny@intel.com>
CC: Don Skidmore <donald.c.skidmore@intel.com>
CC: Greg Rose <gregory.v.rose@intel.com>
CC: Alex Duyck <alexander.h.duyck@intel.com>
CC: John Ronciak <john.ronciak@intel.com>
CC: Mitch Williams <mitch.a.williams@intel.com>
CC: Linux NICS <linux.nics@intel.com>
CC: e1000-devel@lists.sourceforge.net
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
---
drivers/net/ethernet/intel/e1000e/netdev.c | 21 +++++++++------------
1 file changed, 9 insertions(+), 12 deletions(-)
diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c
index 65c3aef..247335d 100644
--- a/drivers/net/ethernet/intel/e1000e/netdev.c
+++ b/drivers/net/ethernet/intel/e1000e/netdev.c
@@ -5164,7 +5164,8 @@ link_up:
#define E1000_TX_FLAGS_VLAN_MASK 0xffff0000
#define E1000_TX_FLAGS_VLAN_SHIFT 16
-static int e1000_tso(struct e1000_ring *tx_ring, struct sk_buff *skb)
+static int e1000_tso(struct e1000_ring *tx_ring, struct sk_buff *skb,
+ __be16 protocol)
{
struct e1000_context_desc *context_desc;
struct e1000_buffer *buffer_info;
@@ -5183,7 +5184,7 @@ static int e1000_tso(struct e1000_ring *tx_ring, struct sk_buff *skb)
hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
mss = skb_shinfo(skb)->gso_size;
- if (skb->protocol == htons(ETH_P_IP)) {
+ if (protocol == htons(ETH_P_IP)) {
struct iphdr *iph = ip_hdr(skb);
iph->tot_len = 0;
iph->check = 0;
@@ -5231,7 +5232,8 @@ static int e1000_tso(struct e1000_ring *tx_ring, struct sk_buff *skb)
return 1;
}
-static bool e1000_tx_csum(struct e1000_ring *tx_ring, struct sk_buff *skb)
+static bool e1000_tx_csum(struct e1000_ring *tx_ring, struct sk_buff *skb,
+ __be16 protocol)
{
struct e1000_adapter *adapter = tx_ring->adapter;
struct e1000_context_desc *context_desc;
@@ -5239,16 +5241,10 @@ static bool e1000_tx_csum(struct e1000_ring *tx_ring, struct sk_buff *skb)
unsigned int i;
u8 css;
u32 cmd_len = E1000_TXD_CMD_DEXT;
- __be16 protocol;
if (skb->ip_summed != CHECKSUM_PARTIAL)
return false;
- if (skb->protocol == cpu_to_be16(ETH_P_8021Q))
- protocol = vlan_eth_hdr(skb)->h_vlan_encapsulated_proto;
- else
- protocol = skb->protocol;
-
switch (protocol) {
case cpu_to_be16(ETH_P_IP):
if (ip_hdr(skb)->protocol == IPPROTO_TCP)
@@ -5546,6 +5542,7 @@ static netdev_tx_t e1000_xmit_frame(struct sk_buff *skb,
int count = 0;
int tso;
unsigned int f;
+ __be16 protocol = vlan_get_protocol(skb);
if (test_bit(__E1000_DOWN, &adapter->state)) {
dev_kfree_skb_any(skb);
@@ -5620,7 +5617,7 @@ static netdev_tx_t e1000_xmit_frame(struct sk_buff *skb,
first = tx_ring->next_to_use;
- tso = e1000_tso(tx_ring, skb);
+ tso = e1000_tso(tx_ring, skb, protocol);
if (tso < 0) {
dev_kfree_skb_any(skb);
return NETDEV_TX_OK;
@@ -5628,14 +5625,14 @@ static netdev_tx_t e1000_xmit_frame(struct sk_buff *skb,
if (tso)
tx_flags |= E1000_TX_FLAGS_TSO;
- else if (e1000_tx_csum(tx_ring, skb))
+ else if (e1000_tx_csum(tx_ring, skb, protocol))
tx_flags |= E1000_TX_FLAGS_CSUM;
/* Old method was to assume IPv4 packet by default if TSO was enabled.
* 82571 hardware supports TSO capabilities for IPv6 as well...
* no longer assume, we must.
*/
- if (skb->protocol == htons(ETH_P_IP))
+ if (protocol == htons(ETH_P_IP))
tx_flags |= E1000_TX_FLAGS_IPV4;
if (unlikely(skb->no_fcs))
--
1.9.3
------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
^ permalink raw reply related
* Re: [GIT PULL] Networking
From: John W. Linville @ 2014-08-25 14:31 UTC (permalink / raw)
To: Linus Torvalds
Cc: David Miller, Andrew Morton, Network Development,
Linux Kernel Mailing List
In-Reply-To: <CA+55aFy+2WiFDRDZwb9B3bh9HS_5Q8K=2PpTHkzi7keSzwc67g@mail.gmail.com>
On Fri, Aug 22, 2014 at 02:47:36PM -0700, Linus Torvalds wrote:
> On Fri, Aug 22, 2014 at 2:40 PM, David Miller <davem@davemloft.net> wrote:
> >
> > I think you can apply it directly and none of us will mind, and GIT
> > will sort it all out during the next merge.
>
> Ok, I'll apply it directly to my tree, since it affected my machine.
> Now that I'm back home I probably won't see the channel switch
> requests that caused the problem, but I just feel uncomfortable
> running a known-broken kernel.
ACK, thanks!
--
John W. Linville Someday the world will need a hero, and you
linville@tuxdriver.com might be all we have. Be ready.
^ permalink raw reply
* [PATCH iproute2] nstat: 64bit support on 32bit arches
From: Eric Dumazet @ 2014-08-25 14:27 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: netdev
From: Eric Dumazet <edumazet@google.com>
SNMP counters can be provided as 64bit numbers.
nstat needs to cope with this even if running in 32bit mode.
Signed-off-by: Eric Dumazet <edumazet@google.com>
---
misc/nstat.c | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/misc/nstat.c b/misc/nstat.c
index d3f8621..e54b3ae 100644
--- a/misc/nstat.c
+++ b/misc/nstat.c
@@ -77,7 +77,6 @@ struct nstat_ent
struct nstat_ent *next;
char *id;
unsigned long long val;
- unsigned long ival;
double rate;
};
@@ -143,7 +142,6 @@ static void load_good_table(FILE *fp)
if ((n = malloc(sizeof(*n))) == NULL)
abort();
n->id = strdup(idbuf);
- n->ival = (unsigned long)val;
n->val = val;
n->rate = rate;
n->next = db;
@@ -206,9 +204,8 @@ static void load_ugly_table(FILE *fp)
if (!p)
abort();
*p = 0;
- if (sscanf(p+1, "%lu", &n->ival) != 1)
+ if (sscanf(p+1, "%llu", &n->val) != 1)
abort();
- n->val = n->ival;
/* Trick to skip "dummy" trailing ICMP MIB in 2.4 */
if (strcmp(idbuf, "IcmpOutAddrMaskReps") == 0)
idbuf[5] = 0;
@@ -365,10 +362,10 @@ static void update_db(int interval)
for (h1 = h; h1; h1 = h1->next) {
if (strcmp(h1->id, n->id) == 0) {
double sample;
- unsigned long incr = h1->ival - n->ival;
- n->val += incr;
- n->ival = h1->ival;
- sample = (double)(incr*1000)/interval;
+ unsigned long long incr = h1->val - n->val;
+
+ n->val = h1->val;
+ sample = (double)incr * 1000.0 / interval;
if (interval >= scan_interval) {
n->rate += W*(sample-n->rate);
} else if (interval >= 1000) {
^ permalink raw reply related
* Re: skb_warn_bad_offload warnings with FreeBSD guests
From: Vlad Yasevich @ 2014-08-25 14:25 UTC (permalink / raw)
To: Brian Rak, netdev
In-Reply-To: <53F76D7E.4030307@gameservers.com>
On 08/22/2014 12:19 PM, Brian Rak wrote:
> We have a number of machines running qemu with bridged networking. We have noticed that
> *sometimes* FreeBSD guests cause this warning to flood the host "WARNING: CPU: 5 PID: 3705
> at net/core/dev.c:2238 skb_warn_bad_offload+0xc3/0xd0()". I haven't been able to come up
> with any sort of reproduction steps, it just seems to happen to some FreeBSD guests, but
> not others.
>
> A full stack trace looks like this:
>
> ------------[ cut here ]------------
> WARNING: CPU: 1 PID: 7147 at net/core/dev.c:2233 skb_warn_bad_offload+0xc3/0xd0()
> igb: caps=(0x0000000190114bb3, 0x0000000000000000) len=2962 data_len=0 gso_size=1448
> gso_type=5 ip_summed=0
> Modules linked in: dm_snapshot dm_bufio ipmi_devintf xt_physdev ebt_arp ebt_ip ebtable_nat
> ebtables cls_fw sch_sfq sch_htb tun kvm_intel kvm 8021q garp nfnetlink_queue nfnetlink_log
> nfnetlink bluetooth rfkill bridge stp llc xt_CHECKSUM iptable_mangle ipt_REJECT
> iptable_filter ip
> _tables ip6t_REJECT ip6table_filter ip6_tables ipv6 iTCO_wdt iTCO_vendor_support ipmi_si
> ipmi_msghandler microcode pcspkr i2c_i801 joydev sg lpc_ich shpchp igb dca ptp pps_core
> hwmon ext4 jbd2 mbcache sd_mod crc_t10dif crct10dif_common video ahci libahci xhci_hcd ast
> ttm drm_kms
> _helper sysimgblt sysfillrect syscopyarea dm_mirror dm_region_hash dm_log dm_mod
> CPU: 1 PID: 7147 Comm: qemu-kvm Tainted: G W 3.15.5-1.el6.elrepo.x86_64 #1
> Hardware name: Supermicro X10SLE-F/HF/X10SLE, BIOS 1.1 07/19/2013
> 00000000000008b9 ffff88081fc435d8 ffffffff8163ba90 00000000000008b9
> ffff88081fc43628 ffff88081fc43618 ffffffff8106c30c ffffc90007a06e30
> 0000000000000000 ffff8807f2b64000 ffff8807f2b64000 0000000000000000
> Call Trace:
> <IRQ> [<ffffffff8163ba90>] dump_stack+0x49/0x61
> [<ffffffff8106c30c>] warn_slowpath_common+0x8c/0xc0
> [<ffffffff8106c3f6>] warn_slowpath_fmt+0x46/0x50
> [<ffffffff8156ce93>] skb_warn_bad_offload+0xc3/0xd0
> [<ffffffff81574a29>] ? dev_hard_start_xmit+0x339/0x640
> [<ffffffff81574699>] __skb_gso_segment+0x89/0xe0
> [<ffffffff81574876>] dev_hard_start_xmit+0x186/0x640
> [<ffffffff81594f5a>] sch_direct_xmit+0xfa/0x1d0
> [<ffffffff81574f2f>] __dev_queue_xmit+0x1ff/0x4f0
> [<ffffffff81575240>] dev_queue_xmit+0x10/0x20
> [<ffffffffa02e6612>] br_dev_queue_push_xmit+0x82/0xb0 [bridge]
> [<ffffffffa02ee680>] br_nf_dev_queue_xmit+0x20/0x90 [bridge]
> [<ffffffffa02ef4b8>] br_nf_post_routing+0x2d8/0x300 [bridge]
> [<ffffffffa02e6590>] ? deliver_clone+0x60/0x60 [bridge]
> [<ffffffff815a357e>] nf_iterate+0x8e/0xc0
> [<ffffffffa02e6590>] ? deliver_clone+0x60/0x60 [bridge]
> [<ffffffff815a37ad>] nf_hook_slow+0x7d/0x150
> [<ffffffffa02e6590>] ? deliver_clone+0x60/0x60 [bridge]
> [<ffffffffa02ee6f0>] ? br_nf_dev_queue_xmit+0x90/0x90 [bridge]
> [<ffffffffa02e6b43>] br_forward_finish+0x43/0x60 [bridge]
> [<ffffffffa02ee8a8>] br_nf_forward_finish+0x1b8/0x1d0 [bridge]
> [<ffffffffa02ef178>] br_nf_forward_ip+0x3a8/0x410 [bridge]
> [<ffffffffa02e6b00>] ? br_flood_deliver+0x20/0x20 [bridge]
> [<ffffffff815a357e>] nf_iterate+0x8e/0xc0
> [<ffffffffa02e6b00>] ? br_flood_deliver+0x20/0x20 [bridge]
> [<ffffffff815a37ad>] nf_hook_slow+0x7d/0x150
> [<ffffffffa02e6b00>] ? br_flood_deliver+0x20/0x20 [bridge]
> [<ffffffffa02e66e4>] __br_forward+0xa4/0x100 [bridge]
> [<ffffffffa02e7800>] ? NF_HOOK.clone.0+0x70/0x70 [bridge]
> [<ffffffffa02e67d6>] br_forward+0x96/0xb0 [bridge]
> [<ffffffffa02e7800>] ? NF_HOOK.clone.0+0x70/0x70 [bridge]
> [<ffffffffa02e7997>] br_handle_frame_finish+0x197/0x3f0 [bridge]
> [<ffffffffa02e7800>] ? NF_HOOK.clone.0+0x70/0x70 [bridge]
> [<ffffffffa02ef790>] br_nf_pre_routing_finish+0x2b0/0x370 [bridge]
> [<ffffffffa02ef4e0>] ? br_nf_post_routing+0x300/0x300 [bridge]
> [<ffffffffa02ed986>] NF_HOOK_THRESH+0x56/0x60 [bridge]
> [<ffffffffa02eed2b>] br_nf_pre_routing+0x2fb/0x3a0 [bridge]
> [<ffffffff815a357e>] nf_iterate+0x8e/0xc0
> [<ffffffffa02e7800>] ? NF_HOOK.clone.0+0x70/0x70 [bridge]
> [<ffffffff815a37ad>] nf_hook_slow+0x7d/0x150
> [<ffffffffa02e7800>] ? NF_HOOK.clone.0+0x70/0x70 [bridge]
> [<ffffffffa02e7d8c>] br_handle_frame+0x19c/0x240 [bridge]
> [<ffffffffa02e7bf0>] ? br_handle_frame_finish+0x3f0/0x3f0 [bridge]
> [<ffffffff81572fa5>] __netif_receive_skb_core+0x1e5/0x620
> [<ffffffff81573407>] __netif_receive_skb+0x27/0x70
> [<ffffffff81573553>] process_backlog+0x103/0x200
> [<ffffffff81573d62>] net_rx_action+0x112/0x2a0
> [<ffffffff8107111c>] __do_softirq+0xfc/0x2b0
> [<ffffffff810713cd>] ? irq_exit+0xad/0xd0
> [<ffffffff8164a81c>] do_softirq_own_stack+0x1c/0x30
> <EOI> [<ffffffff81070e75>] do_softirq+0x55/0x60
> [<ffffffff81571e19>] netif_rx_ni+0x39/0x70
> [<ffffffffa03e84e0>] tun_get_user+0x310/0x6c0 [tun]
> [<ffffffffa03e8995>] tun_chr_aio_write+0x85/0xa0 [tun]
> [<ffffffff811beb9d>] do_sync_readv_writev+0x4d/0x80
> [<ffffffff811c0128>] do_readv_writev+0xc8/0x2c0
> [<ffffffff811bebd0>] ? do_sync_readv_writev+0x80/0x80
> [<ffffffff811d2c45>] ? poll_select_set_timeout+0x95/0xb0
> [<ffffffff811c0357>] vfs_writev+0x37/0x50
> [<ffffffff811c0496>] SyS_writev+0x56/0xf0
> [<ffffffff81648ee9>] system_call_fastpath+0x16/0x1b
> ---[ end trace d26e70ba037ab631 ]---
>
>
> gso_type=5 and ip_summed=0 are always the same (though len, data_len, and gso_size vary).
>
> What is causing this?
The reason that the warning is triggered is ip_summed = 0 which means there is not
checksum already in the packet and it needs to be calculated. If the packet is GSO,
then it needs to have partial checksum set (ip_summed == 3).
You might try using systemtap or instrumenting tun and bridge to see what the
ip_summed value is when this happens.
-vlad
> I've tried kernels as new as 3.15.5-1, which do not appear to help.
>
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" 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
* Re: [PATCH v5 1/2] net: moxa: clear TX descriptor length bits
From: Jonas Jensen @ 2014-08-25 14:23 UTC (permalink / raw)
To: Eric Dumazet, davem
Cc: netdev, linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org, Florian Fainelli,
Michał Mirosław
In-Reply-To: <1408554610.5604.11.camel@edumazet-glaptop2.roam.corp.google.com>
Thanks for giving feedback.
On 20 August 2014 19:10, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> Wouldnt it be faster to not use readl() at all here ?
On 22 August 2014 06:39, David Miller <davem@davemloft.net> wrote:
> Like others I wonder why the existing descriptor value is being read
> at all.
It makes sense and I have verified, transfers continue to work
clearing the whole thing, see update in v6.
Regards,
Jonas
^ permalink raw reply
* Re: [PATCH v4 2/2] net: moxa: replace build_skb() with netdev_alloc_skb_ip_align() / memcpy()
From: Jonas Jensen @ 2014-08-25 14:23 UTC (permalink / raw)
To: Michał Mirosław
Cc: netdev, Florian Fainelli, Linux Kernel, David Miller,
linux-arm-kernel@lists.infradead.org, Eric Dumazet
In-Reply-To: <CAHXqBFJ3bWW8NVYcjxLnkk=6qxcRv75iCuBLrHuF91wxcL=61g@mail.gmail.com>
Thanks for giving feedback.
On 21 August 2014 23:43, Michał Mirosław <mirqus@gmail.com> wrote:
> This has implicit: if (!skb) BUG(); There should probably be a return
> or continue inside the if (!skb).
Fixed, see v6 update (broken out to separate patch) which now includes
increment to RX head counter
> dma_sync_single_for_device() is not needed here as CPU does not and
> should not write to the DMA_FROM_DEVICE mapping.
Fixed, this was also broken out, dma_sync_single_for_device() moved to TX path.
Maybe someone can verify this is the correct thing to do.
Regards,
Jonas
^ permalink raw reply
* [PATCH v6 4/4] net: moxa: continue loop on skb allocation failure
From: Jonas Jensen @ 2014-08-25 14:22 UTC (permalink / raw)
To: netdev
Cc: davem, linux-arm-kernel, linux-kernel, f.fainelli, eric.dumazet,
mirqus, Jonas Jensen
If netdev_alloc_skb_ip_align() fails, subsequent code will
try to dereference an invalid pointer.
Continue to next descriptor on error.
While we're at it,
1. eliminate the chance of an endless loop, replace the main
loop with while(rx < budget)
2. use napi_complete() and remove the explicit napi_gro_flush()
Signed-off-by: Jonas Jensen <jonas.jensen@gmail.com>
---
Notes:
Applies to next-20140825
drivers/net/ethernet/moxa/moxart_ether.c | 12 +++++-------
1 file changed, 5 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/moxa/moxart_ether.c b/drivers/net/ethernet/moxa/moxart_ether.c
index 983d019..2f12c88 100644
--- a/drivers/net/ethernet/moxa/moxart_ether.c
+++ b/drivers/net/ethernet/moxa/moxart_ether.c
@@ -206,7 +206,7 @@ static int moxart_rx_poll(struct napi_struct *napi, int budget)
int rx_head = priv->rx_head;
int rx = 0;
- while (1) {
+ while (rx < budget) {
desc = priv->rx_desc_base + (RX_REG_DESC_SIZE * rx_head);
desc0 = readl(desc + RX_REG_OFFSET_DESC0);
@@ -218,7 +218,7 @@ static int moxart_rx_poll(struct napi_struct *napi, int budget)
net_dbg_ratelimited("packet error\n");
priv->stats.rx_dropped++;
priv->stats.rx_errors++;
- continue;
+ goto rx_next;
}
len = desc0 & RX_DESC0_FRAME_LEN_MASK;
@@ -235,6 +235,7 @@ static int moxart_rx_poll(struct napi_struct *napi, int budget)
net_dbg_ratelimited("netdev_alloc_skb_ip_align failed\n");
priv->stats.rx_dropped++;
priv->stats.rx_errors++;
+ goto rx_next;
}
memcpy(skb->data, priv->rx_buf[rx_head], len);
@@ -249,18 +250,15 @@ static int moxart_rx_poll(struct napi_struct *napi, int budget)
if (desc0 & RX_DESC0_MULTICAST)
priv->stats.multicast++;
+rx_next:
writel(RX_DESC0_DMA_OWN, desc + RX_REG_OFFSET_DESC0);
rx_head = RX_NEXT(rx_head);
priv->rx_head = rx_head;
-
- if (rx >= budget)
- break;
}
if (rx < budget) {
- napi_gro_flush(napi, false);
- __napi_complete(napi);
+ napi_complete(napi);
}
priv->reg_imr |= RPKT_FINISH_M;
--
1.8.2.1
^ permalink raw reply related
* [PATCH v6 3/4] net: moxa: synchronize DMA memory
From: Jonas Jensen @ 2014-08-25 14:22 UTC (permalink / raw)
To: netdev
Cc: davem, linux-arm-kernel, linux-kernel, f.fainelli, eric.dumazet,
mirqus, Jonas Jensen
DMA memory should be synchronized before data is passed
to/from controller.
Add dma_sync_single_for_cpu(.., DMA_FROM_DEVICE) to RX path
and dma_sync_single_for_device(.., DMA_TO_DEVICE) to TX path.
Signed-off-by: Jonas Jensen <jonas.jensen@gmail.com>
---
Notes:
Applies to next-20140825
drivers/net/ethernet/moxa/moxart_ether.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/net/ethernet/moxa/moxart_ether.c b/drivers/net/ethernet/moxa/moxart_ether.c
index d66058d..983d019 100644
--- a/drivers/net/ethernet/moxa/moxart_ether.c
+++ b/drivers/net/ethernet/moxa/moxart_ether.c
@@ -226,6 +226,9 @@ static int moxart_rx_poll(struct napi_struct *napi, int budget)
if (len > RX_BUF_SIZE)
len = RX_BUF_SIZE;
+ dma_sync_single_for_cpu(&ndev->dev,
+ priv->rx_mapping[rx_head],
+ priv->rx_buf_size, DMA_FROM_DEVICE);
skb = netdev_alloc_skb_ip_align(ndev, len);
if (unlikely(!skb)) {
@@ -348,6 +351,9 @@ static int moxart_mac_start_xmit(struct sk_buff *skb, struct net_device *ndev)
len = ETH_ZLEN;
}
+ dma_sync_single_for_device(&ndev->dev, priv->tx_mapping[tx_head],
+ priv->tx_buf_size, DMA_TO_DEVICE);
+
txdes1 = TX_DESC1_LTS | TX_DESC1_FTS | (len & TX_DESC1_BUF_SIZE_MASK);
if (tx_head == TX_DESC_NUM_MASK)
txdes1 |= TX_DESC1_END;
--
1.8.2.1
^ permalink raw reply related
* [PATCH v6 2/4] net: moxa: replace build_skb() with netdev_alloc_skb_ip_align() / memcpy()
From: Jonas Jensen @ 2014-08-25 14:22 UTC (permalink / raw)
To: netdev
Cc: davem, linux-arm-kernel, linux-kernel, f.fainelli, eric.dumazet,
mirqus, Jonas Jensen
In-Reply-To: <1408544342-32058-1-git-send-email-jonas.jensen@gmail.com>
build_skb() is used to make skbs out of existing RX ring memory
which is bad because the RX ring is allocated only once, on probe.
Memory corruption occur because said memory is reclaimed, i.e.
__kfree_skb() (and eventually put_page()).
Replace build_skb() with netdev_alloc_skb_ip_align() and use memcpy().
Remove SKB_DATA_ALIGN() from RX buffer size while we're at it.
Addresses https://bugzilla.kernel.org/show_bug.cgi?id=69041
Signed-off-by: Jonas Jensen <jonas.jensen@gmail.com>
---
Notes:
Changes since v5:
1. broke out DMA synchronization to separate patch
Applies to next-20140825
drivers/net/ethernet/moxa/moxart_ether.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/moxa/moxart_ether.c b/drivers/net/ethernet/moxa/moxart_ether.c
index eed70d9..d66058d 100644
--- a/drivers/net/ethernet/moxa/moxart_ether.c
+++ b/drivers/net/ethernet/moxa/moxart_ether.c
@@ -226,13 +226,15 @@ static int moxart_rx_poll(struct napi_struct *napi, int budget)
if (len > RX_BUF_SIZE)
len = RX_BUF_SIZE;
- skb = build_skb(priv->rx_buf[rx_head], priv->rx_buf_size);
+ skb = netdev_alloc_skb_ip_align(ndev, len);
+
if (unlikely(!skb)) {
- net_dbg_ratelimited("build_skb failed\n");
+ net_dbg_ratelimited("netdev_alloc_skb_ip_align failed\n");
priv->stats.rx_dropped++;
priv->stats.rx_errors++;
}
+ memcpy(skb->data, priv->rx_buf[rx_head], len);
skb_put(skb, len);
skb->protocol = eth_type_trans(skb, ndev);
napi_gro_receive(&priv->napi, skb);
@@ -464,8 +466,7 @@ static int moxart_mac_probe(struct platform_device *pdev)
spin_lock_init(&priv->txlock);
priv->tx_buf_size = TX_BUF_SIZE;
- priv->rx_buf_size = RX_BUF_SIZE +
- SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
+ priv->rx_buf_size = RX_BUF_SIZE;
priv->tx_desc_base = dma_alloc_coherent(NULL, TX_REG_DESC_SIZE *
TX_DESC_NUM, &priv->tx_base,
--
1.8.2.1
^ permalink raw reply related
* [PATCH v6 1/4] net: moxa: clear DESC1 on ndo_start_xmit()
From: Jonas Jensen @ 2014-08-25 14:22 UTC (permalink / raw)
To: netdev
Cc: davem, linux-arm-kernel, linux-kernel, f.fainelli, eric.dumazet,
mirqus, Jonas Jensen
In-Reply-To: <1408544322-32023-1-git-send-email-jonas.jensen@gmail.com>
TX buffer length is not cleared on ndo_start_xmit().
Failing to do so can bug/hang the controller and
cause TX interrupts to stop altogether.
Remove the readl() and compute a new value for DESC1.
Addresses https://bugzilla.kernel.org/show_bug.cgi?id=69031
Signed-off-by: Jonas Jensen <jonas.jensen@gmail.com>
---
Notes:
Changes since v5:
1. remove readl(desc + TX_REG_OFFSET_DESC1) in moxart_mac_start_xmit()
2. compute entirely new values for DESC1
3. set TX_DESC1_END if (tx_head == TX_DESC_NUM_MASK)
Applies to next-20140825
drivers/net/ethernet/moxa/moxart_ether.c | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/moxa/moxart_ether.c b/drivers/net/ethernet/moxa/moxart_ether.c
index 5020fd4..eed70d9 100644
--- a/drivers/net/ethernet/moxa/moxart_ether.c
+++ b/drivers/net/ethernet/moxa/moxart_ether.c
@@ -346,10 +346,9 @@ static int moxart_mac_start_xmit(struct sk_buff *skb, struct net_device *ndev)
len = ETH_ZLEN;
}
- txdes1 = readl(desc + TX_REG_OFFSET_DESC1);
- txdes1 |= TX_DESC1_LTS | TX_DESC1_FTS;
- txdes1 &= ~(TX_DESC1_FIFO_COMPLETE | TX_DESC1_INTR_COMPLETE);
- txdes1 |= (len & TX_DESC1_BUF_SIZE_MASK);
+ txdes1 = TX_DESC1_LTS | TX_DESC1_FTS | (len & TX_DESC1_BUF_SIZE_MASK);
+ if (tx_head == TX_DESC_NUM_MASK)
+ txdes1 |= TX_DESC1_END;
writel(txdes1, desc + TX_REG_OFFSET_DESC1);
writel(TX_DESC0_DMA_OWN, desc + TX_REG_OFFSET_DESC0);
--
1.8.2.1
^ permalink raw reply related
* Re: [patch net-next RFC 10/12] openvswitch: add support for datapath hardware offload
From: Thomas Graf @ 2014-08-25 14:17 UTC (permalink / raw)
To: Jamal Hadi Salim
Cc: John Fastabend, Scott Feldman, Jiri Pirko, netdev, David Miller,
Neil Horman, Andy Gospodarek, dborkman, ogerlitz, jesse, pshelar,
azhou, ben, stephen, jeffrey.t.kirsher, vyasevic, xiyou.wangcong,
john.r.fastabend, edumazet, f.fainelli, roopa, linville, dev,
jasowang, ebiederm, nicolas.dichtel, ryazanov.s.a, buytenh,
aviadr, nbd, alexei.starovoitov, Neil.Jerram, ronye
In-Reply-To: <53FB3FD5.2030905@mojatatu.com>
On 08/25/14 at 09:53am, Jamal Hadi Salim wrote:
> On 08/24/14 22:42, John Fastabend wrote:
>
> >In the L2 case we already have the fdb_add and fdb_del semantics that
> >are being used today by NICs with embedded switches. And we have a DSA
> >patch we could dig out of patchwork for those drivers.
>
> Indeed. That is an excellent starting point of something that is proven
> to work. I was hoping L3 would follow this path. For L2,
> there is the mess of claiming unicast NIC addresses as part of
> the fdb, but that can almost be ignored.
> Caveat: fdb_XXX works well for NICs as well as different larger
> ASICs - but some quark handling is going to be needed for the tinier
> openwrt type devices.
> For larger switches we are going to need more for L2:
> Really, the bridge already covers everthing large switches
> offer (Vlan filtering, bridge port setting etc). We just need
> to offload that...
fdb_add() *is* flow based. At least in my understanding, the whole
point here is to extend the idea of fdb_add() and make it understand
L2-L4 in a more generic way for the most common protocols.
The reason fdb_add() is not reused is because it is Netlink specific
and only suitable for User -> HW offload. Kernel -> HW offload is
technically possible but not clean.
The only reason swdev is needed at all is to represent the port model
and to allow for non flow based models built on top of the same
hardware abstraction. I see now reason why br_fdb cannot be represented
through swdev as soon as the code is stable.
> >So I think it makes more sense to use the explicit interface rather
> >than put another shim layer in the kernel. Its simpler and more to the
> >point IMO. I suspect the resulting code will be smaller and easier to
> >read. I'm the squemish one in the audience here.
> >
>
> L2/3 should be left alone and go this path. My concern on excessive
> NDOs was more on the flow view of the world.
The point I was trying to make earlier is that it is very hard to
program both protocol aware and generic filtering hardware through
a single NDO. It will make the driver specific part complex.
If you are saying we need yet another classifier model in the kernel
then I'm not sure that is needed in the presence of cls/act, iptables,
and nftables. They seem suitable to represent non flow based models
and I see nothing that would prevent an offload through swdev for them.
^ permalink raw reply
* Re: [patch net-next 1/3] net: propagate sock pointer through netfilter hooks
From: Jiri Pirko @ 2014-08-25 14:05 UTC (permalink / raw)
To: David Miller
Cc: netdev, kuznet, jmorris, yoshfuji, stephen, cwang, pshelar,
nicolas.dichtel, therbert, dborkman, edumazet
In-Reply-To: <20140821.173339.1409243624518012670.davem@davemloft.net>
Fri, Aug 22, 2014 at 02:33:39AM CEST, davem@davemloft.net wrote:
>From: Jiri Pirko <jiri@resnulli.us>
>Date: Fri, 15 Aug 2014 20:32:54 +0200
>
>> When output function (ip6_finish_output2 for example) needs to be called
>> with sock pointer, we need to push sock pointer through the netfilter
>> hooks. This patch does that.
>>
>> Signed-off-by: Jiri Pirko <jiri@resnulli.us>
>
>Ok, I'm going to admit that I am having second thoughts about this
>approach. This is a quite large set of churn to fix this bug.
>
>However, in the same breath, I can't come up with a simpler way to
>propagate this information without the really unacceptable overhead of
>adding another sk_buff member.
>
>And even if we found some simple way to deal with that sk_mc_loop()
>test, ipv6 has other demons in this area.
>
>For example, look at what ip6_fragment() does, it also assumes skb->sk
>is an inet6 socket.
>
> struct ipv6_pinfo *np = skb->sk ? inet6_sk(skb->sk) : NULL;
> ...
> if (np && np->frag_size < mtu) {
> if (np->frag_size)
> mtu = np->frag_size;
> }
>
>The rest of the skb->sk usage in these place is fine, as they are
>simply propagating socket ownership from one packet to another, rather
>than doing protocol specific things with them.
Dave, I see that the patchset is in state "Changes requested". I do not
think I understand what is needed to be done at this point. Would you
please tell me? Thanks.
^ permalink raw reply
* Re: [PATCH V4 3/8] namespaces: expose ns instance serial numbers in proc
From: Andy Lutomirski @ 2014-08-25 14:04 UTC (permalink / raw)
To: nicolas.dichtel-pdR9zngts4EAvxtiuMwx3w
Cc: Linux API, Linux Containers,
linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-audit-H+wXaHxf7aLQT0dZR+AlfA, Eric W. Biederman, netdev
In-Reply-To: <53FB3A86.2060203-pdR9zngts4EAvxtiuMwx3w@public.gmane.org>
On Aug 25, 2014 6:30 AM, "Nicolas Dichtel" <nicolas.dichtel-pdR9zngts4EAvxtiuMwx3w@public.gmane.org> wrote:
>> CRIU wants to save the complete state of a namespace and then restore
>> it. For that to work, any information exposed to things in the
>> namespace *cannot* be globally unique or unique per boot, since CRIU
>> needs to arrange for that information to match whatever it was when
>> CRIU saved it.
>
> How are ifindex of network devices managed? These ifindexes are unique per boot,
> thus can change depending on the order in which netdev are created.
> These ifindexes are unique per boot and exposed to userspace ...
>
This does not appear to be true.
$ sudo unshare --net
# ip link add veth0 type veth peer name veth1
# ip link
1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN mode DEFAULT group default
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: veth1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode
DEFAULT group default qlen 1000
link/ether 06:0d:59:c7:a6:a8 brd ff:ff:ff:ff:ff:ff
3: veth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode
DEFAULT group default qlen 1000
link/ether b2:5c:8b:f2:12:28 brd ff:ff:ff:ff:ff:ff
# logout
$ ip link
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
3: em1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast
state DOWN qlen 1000
>
>>
>> Also, I think that code running in a namespace has no business even
>> knowing a unique identity of that namespace from the perspective of
>> the host.
>
> Another scenario is when you have virtual network devices across two netns. You
> need to identify the peer netns to have a netlink message which is fully interpretable by the userspace.
Let me try again, with emphasis in the right place.
I think that *code running in a namespace* has no business even
knowing a unique identity of *that namespace* from the perspective of
the host.
In your example, if there's a veth device between netns A and netns B,
then code *in netns A* has no business knowing the identity of its
veth peer if its peer (B) is a sibling or ancestor. It also IMO has
no business knowing the identity of its own netns (A) other than as
"my netns".
If A and B are siblings, then their parent needs to know where that
veth device goes, but I think this is already the case to a sufficient
extent today.
I feel like this discussion is falling into a common trap of new API
discussions. Can one of you who wants this API please articulate,
with a reasonably precise example, what it is that you want to do, why
you can't easily do it already, and how this API helps? I currently
understand how the API creates problems, but I don't understand how it
solves any problems, and I will NAK it (and I suspect that Eric will,
too, which is pretty much fatal) unless that changes.
Thanks,
Andy
>
>
> Regards,
> Nicolas
^ permalink raw reply
* [PATCH] Driver for IBM System i/p VNIC protocol
From: Santiago Leon @ 2014-08-25 13:52 UTC (permalink / raw)
To: netdev; +Cc: David Miller
This is a new device driver for a high performance SR-IOV assisted virtual
network for IBM System p and IBM System i systems. The SR-IOV VF will be
attached to the VIOS partition and mapped to the Linux client via the
hypervisor's VNIC protocol that this driver implements.
This driver is able to perform basic tx and rx, new features and improvements
will be added as they are being developed and tested.
Signed-off-by: Santiago Leon <santil@linux.vnet.ibm.com>
---
arch/powerpc/include/asm/hvcall.h | 20
drivers/net/ethernet/ibm/Kconfig | 10
drivers/net/ethernet/ibm/Makefile | 1
drivers/net/ethernet/ibm/ibmvnic.c | 3642 +++++++++++++++++++++++++++++++++++++
drivers/net/ethernet/ibm/ibmvnic.h | 962 +++++++++
5 files changed, 4635 insertions(+)
--- a/drivers/net/ethernet/ibm/Kconfig 2014-08-16 11:40:26.000000000 -0500
+++ b/drivers/net/ethernet/ibm/Kconfig 2014-08-25 07:39:00.000000000 -0500
@@ -39,4 +39,14 @@ config EHEA
To compile the driver as a module, choose M here. The module
will be called ehea.
+config IBMVNIC
+ tristate "IBM Virtual NIC support"
+ depends on PPC_PSERIES
+ ---help---
+ This driver supports Virtual NIC adapters on IBM i and IBM System p
+ systems.
+
+ To compile this driver as a module, choose M here. The module will
+ be called ibmvnic.
+
endif # NET_VENDOR_IBM
--- a/drivers/net/ethernet/ibm/Makefile 2014-08-16 11:40:26.000000000 -0500
+++ b/drivers/net/ethernet/ibm/Makefile 2014-08-25 07:39:00.000000000 -0500
@@ -3,5 +3,6 @@
#
obj-$(CONFIG_IBMVETH) += ibmveth.o
+obj-$(CONFIG_IBMVNIC) += ibmvnic.o
obj-$(CONFIG_IBM_EMAC) += emac/
obj-$(CONFIG_EHEA) += ehea/
--- a/drivers/net/ethernet/ibm/ibmvnic.h 1969-12-31 18:00:00.000000000 -0600
+++ b/drivers/net/ethernet/ibm/ibmvnic.h 2014-08-25 07:39:00.000000000 -0500
@@ -0,0 +1,962 @@
+/**************************************************************************/
+/* */
+/* IBM System i and System p Virtual NIC Device Driver */
+/* Copyright (C) 2014 IBM Corp. */
+/* Santiago Leon (santil@linux.vnet.ibm.com) */
+/* */
+/* This program is free software; you can redistribute it and/or modify */
+/* it under the terms of the GNU General Public License as published by */
+/* the Free Software Foundation; either version 2 of the License, or */
+/* (at your option) any later version. */
+/* */
+/* This program is distributed in the hope that it will be useful, */
+/* but WITHOUT ANY WARRANTY; without even the implied warranty of */
+/* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the */
+/* GNU General Public License for more details. */
+/* */
+/* You should have received a copy of the GNU General Public License */
+/* along with this program; if not, write to the Free Software */
+/* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 */
+/* USA */
+/* */
+/* This module contains the implementation of a virtual ethernet device */
+/* for use with IBM i/pSeries LPAR Linux. It utilizes the logical LAN */
+/* option of the RS/6000 Platform Architechture to interface with virtual */
+/* ethernet NICs that are presented to the partition by the hypervisor. */
+/* */
+/**************************************************************************/
+
+#define IBMVNIC_NAME "ibmvnic"
+#define IBMVNIC_DRIVER_VERSION "1.0"
+#define IBMVNIC_INVALID_MAP -1
+#define IBMVNIC_STATS_TIMEOUT 1
+#define IBMVNIC_IO_ENTITLEMENT_DEFAULT 610305 /* basic structures plus
+ * 100 2k buffers */
+
+/* Initial module_parameters */
+#define IBMVNIC_RX_WEIGHT 16
+#define IBMVNIC_BUFFS_PER_POOL 100 /* when changing this, update
+ * IBMVNIC_IO_ENTITLEMENT_DEFAULT */
+#define IBMVNIC_MAX_TX_QUEUES 4
+
+struct ibmvnic_login_buffer {
+ __be32 len;
+ __be32 version;
+#define INITIAL_VERSION_LB 1
+ __be32 num_txcomp_subcrqs;
+ __be32 off_txcomp_subcrqs;
+ __be32 num_rxcomp_subcrqs;
+ __be32 off_rxcomp_subcrqs;
+ __be32 login_rsp_ioba;
+ __be32 login_rsp_len;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_login_rsp_buffer {
+ __be32 len;
+ __be32 version;
+#define INITIAL_VERSION_LRB 1
+ __be32 num_txsubm_subcrqs;
+ __be32 off_txsubm_subcrqs;
+ __be32 num_rxadd_subcrqs;
+ __be32 off_rxadd_subcrqs;
+ __be32 off_rxadd_buff_size;
+ __be32 num_supp_tx_desc;
+ __be32 off_supp_tx_desc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_query_ip_offload_buffer {
+ __be32 len;
+ __be32 version;
+#define INITIAL_VERSION_IOB 1
+ u8 ipv4_chksum;
+ u8 ipv6_chksum;
+ u8 tcp_ipv4_chksum;
+ u8 tcp_ipv6_chksum;
+ u8 udp_ipv4_chksum;
+ u8 udp_ipv6_chksum;
+ u8 large_tx_ipv4;
+ u8 large_tx_ipv6;
+ u8 large_rx_ipv4;
+ u8 large_rx_ipv6;
+ u8 reserved1[14];
+ __be16 max_ipv4_header_size;
+ __be16 max_ipv6_header_size;
+ __be16 max_tcp_header_size;
+ __be16 max_udp_header_size;
+ __be32 max_large_tx_size;
+ __be32 max_large_rx_size;
+ u8 reserved2[16];
+ u8 ipv6_extension_header;
+#define IPV6_EH_NOT_SUPPORTED 0x00
+#define IPV6_EH_SUPPORTED_LIM 0x01
+#define IPV6_EH_SUPPORTED 0xFF
+ u8 tcp_pseudosum_req;
+#define TCP_PS_NOT_REQUIRED 0x00
+#define TCP_PS_REQUIRED 0x01
+ u8 reserved3[30];
+ __be16 num_ipv6_ext_headers;
+ __be32 off_ipv6_ext_headers;
+ u8 reserved4[154];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_control_ip_offload_buffer {
+ __be32 len;
+ __be32 version;
+#define INITIAL_VERSION_IOB 1
+ u8 ipv4_chksum;
+ u8 ipv6_chksum;
+ u8 tcp_ipv4_chksum;
+ u8 tcp_ipv6_chksum;
+ u8 udp_ipv4_chksum;
+ u8 udp_ipv6_chksum;
+ u8 large_tx_ipv4;
+ u8 large_tx_ipv6;
+ u8 bad_packet_rx;
+ u8 large_rx_ipv4;
+ u8 large_rx_ipv6;
+ u8 reserved4[111];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_fw_component {
+ u8 name[48];
+ __be32 trace_buff_size;
+ u8 correlator;
+ u8 trace_level;
+ u8 parent_correlator;
+ u8 error_check_level;
+ u8 trace_on;
+ u8 reserved[7];
+ u8 description[192];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_fw_trace_entry {
+ __be32 trace_id;
+ u8 num_valid_data;
+ u8 reserved[3];
+ __be64 pmc_registers;
+ __be64 timebase;
+ __be64 trace_data[5];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_statistics {
+ __be32 version;
+ __be32 promiscuous;
+ __be64 rx_packets;
+ __be64 rx_bytes;
+ __be64 tx_packets;
+ __be64 tx_bytes;
+ __be64 ucast_tx_packets;
+ __be64 ucast_rx_packets;
+ __be64 mcast_tx_packets;
+ __be64 mcast_rx_packets;
+ __be64 bcast_tx_packets;
+ __be64 bcast_rx_packets;
+ __be64 align_errors;
+ __be64 fcs_errors;
+ __be64 single_collision_frames;
+ __be64 multi_collision_frames;
+ __be64 sqe_test_errors;
+ __be64 deferred_tx;
+ __be64 late_collisions;
+ __be64 excess_collisions;
+ __be64 internal_mac_tx_errors;
+ __be64 carrier_sense;
+ __be64 too_long_frames;
+ __be64 internal_mac_rx_errors;
+ u8 reserved[72];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_acl_buffer {
+ __be32 len;
+ __be32 version;
+#define INITIAL_VERSION_IOB 1
+ u8 mac_acls_restrict;
+ u8 vlan_acls_restrict;
+ u8 reserved1[22];
+ __be32 num_mac_addrs;
+ __be32 offset_mac_addrs;
+ __be32 num_vlan_ids;
+ __be32 offset_vlan_ids;
+ u8 reserved2[80];
+} __attribute__((packed, aligned(8)));
+
+#define IBMVNIC_TX_DESC_VERSIONS 3
+struct ibmvnic_tx_desc_v0 {
+ u8 first;
+ u8 type;
+#define IBMVNIC_TX_DESC_V0 0
+ u8 flags1;
+#define IBMVNIC_LARGE_TX 0x80
+#define IBMVNIC_IP_CHKSUM 0x40
+#define IBMVNIC_TCP_CHKSUM 0x20
+#define IBMVNIC_INS_VLANH 0x10
+#define IBMVNIC_UDP_CHKSUM 0x08
+#define IBMVNIC_MULTI_TX_DESC 0x04
+#define IBMVNIC_LAST_FRAG 0x02
+#define IBMVNIC_TX_COMP_EVENT 0x01
+ u8 flags2;
+#define IPV6 0x80
+ __be16 tcp_udp_h_off;
+ __be16 vlan_h;
+ u8 reserved;
+ u8 mss_msb;
+ __be16 mss;
+ __be32 correlator;
+ __be32 ioba1;
+ __be32 len1;
+ __be32 ioba2;
+ __be32 len2;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_tx_comp_desc {
+ u8 first;
+ u8 num_comps;
+ __be16 rcs[5];
+ __be32 correlators[5];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_tx_desc_v1 {
+ u8 first;
+ u8 type;
+#define IBMVNIC_TX_DESC_V1 1
+ u8 flags1; /* same as V0 */
+ u8 flags2; /* same as V0 */
+ __be16 tcp_udp_h_off;
+ __be16 vlan_h;
+ u8 reserved;
+ u8 mss[3];
+ __be32 correlator;
+ __be32 ioba1;
+ __be32 len1;
+ u8 dest_mac[6];
+ u8 ethertype;
+} __attribute__((packed, aligned(8)));
+
+#define IBMVNIC_MAX_FRAGS_PER_CRQ 3
+struct ibmvnic_tx_desc_v2 {
+ u8 first;
+ u8 type;
+#define IBMVNIC_TX_DESC_V2 2
+ u8 n_crq_elem;
+ u8 n_sge;
+ u8 flags1;
+#define IBMVNIC_TXV2_COMP_NEEDED 0x80
+#define IBMVNIC_TXV2_CHKSUM_OFFLOAD 0x40
+#define IBMVNIC_TXV2_LSO 0x20
+#define IBMVNIC_TXV2_PROT_TCP 0x10
+#define IBMVNIC_TXV2_PROT_UDP 0x08
+#define IBMVNIC_TXV2_PROT_IPV4 0x04
+#define IBMVNIC_TXV2_PROT_IPV6 0x02
+#define IBMVNIC_TXV2_VLAN_PRESENT 0x01
+ u8 flags2;
+#define IBMVNIC_TXV2_VLAN_INSERT 0x80
+#define IBMVNIC_TXV2_SEND_HDRS_IMMED 0x40
+ __be16 mss;
+ __be64 correlator;
+ __be16 vlan_id;
+ __be16 dma_reg;
+ __be32 sge_len;
+ __be64 ioba;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_sge_desc {
+ u8 first;
+ u8 type;
+#define IBMVNIC_SGE_DESC 0x30
+ __be16 sge1_dma_reg;
+ __be32 sge1_len;
+ __be64 sge1_ioba;
+ __be16 reserved;
+ __be16 sge2_dma_reg;
+ __be32 sge2_len;
+ __be64 sge2_ioba;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_rx_desc {
+ u8 first;
+ u8 flags;
+#define IBMVNIC_IP_CHKSUM_GOOD 0x80
+#define IBMVNIC_TCP_UDP_CHKSUM_GOOD 0x40
+#define IBMVNIC_END_FRAME 0x20
+#define IBMVNIC_EXACT_MC 0x10
+#define IBMVNIC_CSUM_FIELD 0x08
+ __be16 off_frame_data;
+ __be32 len;
+ __be64 correlator;
+ __be16 csum;
+ u8 reserved[14];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_generic_scrq {
+ u8 first;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_rx_buff_add_desc {
+ u8 first;
+ u8 reserved1[7];
+ __be64 correlator;
+ __be32 ioba;
+ __be32 len;
+ u8 reserved2[8];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_rc {
+ volatile u8 code; /* one of enum ibmvnic_rc_codes */
+ volatile u8 detailed_data[3];
+} __attribute__((packed, aligned(4)));
+
+struct ibmvnic_generic_crq {
+ u8 first;
+ u8 cmd;
+ u8 params[10];
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_version_exchange {
+ u8 first;
+ u8 cmd;
+ __be16 version;
+#define IBMVNIC_INITIAL_VERSION 1
+ u8 reserved[8];
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_capability {
+ u8 first;
+ u8 cmd;
+ __be16 capability; /* one of ibmvnic_capabilities */
+ struct ibmvnic_rc rc;
+ __be32 number; /*FIX: should be __be64, but I'm getting the least
+ * significant word first */
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_login {
+ u8 first;
+ u8 cmd;
+ u8 reserved[6];
+ __be32 ioba;
+ __be32 len;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_phys_parms {
+ u8 first;
+ u8 cmd;
+ u8 flags1;
+#define IBMVNIC_EXTERNAL_LOOPBACK 0x80
+#define IBMVNIC_INTERNAL_LOOPBACK 0x40
+#define IBMVNIC_PROMISC 0x20
+#define IBMVNIC_PHYS_LINK_ACTIVE 0x10
+#define IBMVNIC_AUTONEG_DUPLEX 0x08
+#define IBMVNIC_FULL_DUPLEX 0x04
+#define IBMVNIC_HALF_DUPLEX 0x02
+#define IBMVNIC_CAN_CHG_PHYS_PARMS 0x01
+ u8 flags2;
+#define IBMVNIC_LOGICAL_LNK_ACTIVE 0x80
+ __be32 speed;
+#define IBMVNIC_AUTONEG 0x80
+#define IBMVNIC_10MBPS 0x40
+#define IBMVNIC_100MBPS 0x20
+#define IBMVNIC_1GBPS 0x10
+#define IBMVNIC_10GBPS 0x08
+ __be32 mtu;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_logical_link_state {
+ u8 first;
+ u8 cmd;
+ u8 link_state;
+#define IBMVNIC_LOGICAL_LNK_DN 0x00
+#define IBMVNIC_LOGICAL_LNK_UP 0x01
+#define IBMVNIC_LOGICAL_LNK_QUERY 0xff
+ u8 reserved[9];
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_query_ip_offload {
+ u8 first;
+ u8 cmd;
+ u8 reserved[2];
+ __be32 len;
+ __be32 ioba;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_control_ip_offload {
+ u8 first;
+ u8 cmd;
+ u8 reserved[2];
+ __be32 ioba;
+ __be32 len;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_dump_size {
+ u8 first;
+ u8 cmd;
+ u8 reserved[6];
+ __be32 len;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_dump {
+ u8 first;
+ u8 cmd;
+ u8 reserved1[2];
+ __be32 ioba;
+ __be32 len;
+ u8 reserved2[4];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_dump_rsp {
+ u8 first;
+ u8 cmd;
+ u8 reserved[6];
+ __be32 dumped_len;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_ras_comp_num {
+ u8 first;
+ u8 cmd;
+ u8 reserved1[2];
+ __be32 num_components;
+ u8 reserved2[4];
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_ras_comps {
+ u8 first;
+ u8 cmd;
+ u8 reserved[2];
+ __be32 ioba;
+ __be32 len;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_control_ras {
+ u8 first;
+ u8 cmd;
+ u8 correlator;
+ u8 level;
+ u8 op;
+#define IBMVNIC_TRACE_LEVEL 1
+#define IBMVNIC_ERROR_LEVEL 2
+#define IBMVNIC_TRACE_PAUSE 3
+#define IBMVNIC_TRACE_RESUME 4
+#define IBMVNIC_TRACE_ON 5
+#define IBMVNIC_TRACE_OFF 6
+#define IBMVNIC_CHG_TRACE_BUFF_SZ 7
+ u8 trace_buff_sz[3];
+ u8 reserved[4];
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_collect_fw_trace {
+ u8 first;
+ u8 cmd;
+ u8 correlator;
+ u8 reserved;
+ __be32 ioba;
+ __be32 len;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_statistics {
+ u8 first;
+ u8 cmd;
+ u8 flags;
+#define IBMVNIC_PHYSICAL_PORT 0x80
+ u8 reserved1;
+ __be32 ioba;
+ __be32 len;
+ u8 reserved[4];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_debug_stats {
+ u8 first;
+ u8 cmd;
+ u8 reserved[2];
+ __be32 ioba;
+ __be32 len;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_error_indication {
+ u8 first;
+ u8 cmd;
+ u8 flags;
+#define IBMVNIC_FATAL_ERROR 0x80
+ u8 reserved1;
+ __be32 error_id;
+ __be32 detail_error_sz;
+ __be16 error_cause;
+ u8 reserved2[2];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_error_info {
+ u8 first;
+ u8 cmd;
+ u8 reserved[2];
+ __be32 ioba;
+ __be32 len;
+ __be32 error_id;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_request_error_rsp {
+ u8 first;
+ u8 cmd;
+ u8 reserved[2];
+ __be32 error_id;
+ __be32 len;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_link_state_indication {
+ u8 first;
+ u8 cmd;
+ u8 reserved1[2];
+ u8 phys_link_state;
+ u8 logical_link_state;
+ u8 reserved2[10];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_change_mac_addr {
+ u8 first;
+ u8 cmd;
+ u8 mac_addr[6];
+ struct ibmvnic_rc rc;
+ u8 reserved[4];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_multicast_ctrl {
+ u8 first;
+ u8 cmd;
+ u8 mac_addr[6];
+ u8 flags;
+#define IBMVNIC_ENABLE_MC 0x80
+#define IBMVNIC_DISABLE_MC 0x40
+#define IBMVNIC_ENABLE_ALL 0x20
+#define IBMVNIC_DISABLE_ALL 0x10
+ u8 reserved1;
+ __be16 reserved2; /* was num_enabled_mc_addr; */
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_get_vpd_size_rsp {
+ u8 first;
+ u8 cmd;
+ u8 reserved[2];
+ __be64 len;
+ struct ibmvnic_rc rc;
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_get_vpd {
+ u8 first;
+ u8 cmd;
+ u8 reserved1[2];
+ __be32 ioba;
+ __be32 len;
+ u8 reserved[4];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_acl_change_indication {
+ u8 first;
+ u8 cmd;
+ __be16 change_type;
+#define IBMVNIC_MAC_ACL 0;
+#define IBMVNIC_VLAN_ACL 1;
+ u8 reserved[12];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_acl_query {
+ u8 first;
+ u8 cmd;
+ u8 reserved1[2];
+ __be32 ioba;
+ __be32 len;
+ u8 reserved2[4];
+} __attribute__((packed, aligned(8)));
+
+struct ibmvnic_tune {
+ u8 first;
+ u8 cmd;
+ u8 reserved1[2];
+ __be32 ioba;
+ __be32 len;
+ u8 reserved2[4];
+} __attribute__((packed, aligned(8)));
+
+union ibmvnic_crq {
+ struct ibmvnic_generic_crq generic;
+ struct ibmvnic_version_exchange version_exchange;
+ struct ibmvnic_version_exchange version_exchange_rsp;
+ struct ibmvnic_capability query_capability;
+ struct ibmvnic_capability query_capability_rsp;
+ struct ibmvnic_capability request_capability;
+ struct ibmvnic_capability request_capability_rsp;
+ struct ibmvnic_login login;
+ struct ibmvnic_generic_crq login_rsp;
+ struct ibmvnic_phys_parms query_phys_parms;
+ struct ibmvnic_phys_parms query_phys_parms_rsp;
+ struct ibmvnic_phys_parms query_phys_capabilities;
+ struct ibmvnic_phys_parms query_phys_capabilities_rsp;
+ struct ibmvnic_phys_parms set_phys_parms;
+ struct ibmvnic_phys_parms set_phys_parms_rsp;
+ struct ibmvnic_logical_link_state logical_link_state;
+ struct ibmvnic_logical_link_state logical_link_state_rsp;
+ struct ibmvnic_query_ip_offload query_ip_offload;
+ struct ibmvnic_query_ip_offload query_ip_offload_rsp;
+ struct ibmvnic_control_ip_offload control_ip_offload;
+ struct ibmvnic_control_ip_offload control_ip_offload_rsp;
+ struct ibmvnic_request_dump_size request_dump_size;
+ struct ibmvnic_request_dump_size request_dump_size_rsp;
+ struct ibmvnic_request_dump request_dump;
+ struct ibmvnic_request_dump_rsp request_dump_rsp;
+ struct ibmvnic_request_ras_comp_num request_ras_comp_num;
+ struct ibmvnic_request_ras_comp_num request_ras_comp_num_rsp;
+ struct ibmvnic_request_ras_comps request_ras_comps;
+ struct ibmvnic_request_ras_comps request_ras_comps_rsp;
+ struct ibmvnic_control_ras control_ras;
+ struct ibmvnic_control_ras control_ras_rsp;
+ struct ibmvnic_collect_fw_trace collect_fw_trace;
+ struct ibmvnic_collect_fw_trace collect_fw_trace_rsp;
+ struct ibmvnic_request_statistics request_statistics;
+ struct ibmvnic_generic_crq request_statistics_rsp;
+ struct ibmvnic_request_debug_stats request_debug_stats;
+ struct ibmvnic_request_debug_stats request_debug_stats_rsp;
+ struct ibmvnic_error_indication error_indication;
+ struct ibmvnic_request_error_info request_error_info;
+ struct ibmvnic_request_error_rsp request_error_rsp;
+ struct ibmvnic_link_state_indication link_state_indication;
+ struct ibmvnic_change_mac_addr change_mac_addr;
+ struct ibmvnic_change_mac_addr change_mac_addr_rsp;
+ struct ibmvnic_multicast_ctrl multicast_ctrl;
+ struct ibmvnic_multicast_ctrl multicast_ctrl_rsp;
+ struct ibmvnic_generic_crq get_vpd_size;
+ struct ibmvnic_get_vpd_size_rsp get_vpd_size_rsp;
+ struct ibmvnic_get_vpd get_vpd;
+ struct ibmvnic_generic_crq get_vpd_rsp;
+ struct ibmvnic_acl_change_indication acl_change_indication;
+ struct ibmvnic_acl_query acl_query;
+ struct ibmvnic_generic_crq acl_query_rsp;
+ struct ibmvnic_tune tune;
+ struct ibmvnic_generic_crq tune_rsp;
+};
+
+enum ibmvnic_rc_codes {
+ SUCCESS = 0,
+ PARTIALSUCCESS = 1,
+ PERMISSION = 2,
+ NOMEMORY = 3,
+ PARAMETER = 4,
+ UNKNOWNCOMMAND = 5,
+ ABORTED = 6,
+ INVALIDSTATE = 7,
+ INVALIDIOBA = 8,
+ INVALIDLENGTH = 9,
+ UNSUPPORTEDOPTION = 10,
+};
+
+enum ibmvnic_capabilities {
+ MIN_TX_QUEUES = 1,
+ MIN_RX_QUEUES = 2,
+ MIN_RX_ADD_QUEUES = 3,
+ MAX_TX_QUEUES = 4,
+ MAX_RX_QUEUES = 5,
+ MAX_RX_ADD_QUEUES = 6,
+ REQ_TX_QUEUES = 7,
+ REQ_RX_QUEUES = 8,
+ REQ_RX_ADD_QUEUES = 9,
+ MIN_TX_ENTRIES_PER_SUBCRQ = 10,
+ MIN_RX_ADD_ENTRIES_PER_SUBCRQ = 11,
+ MAX_TX_ENTRIES_PER_SUBCRQ = 12,
+ MAX_RX_ADD_ENTRIES_PER_SUBCRQ = 13,
+ REQ_TX_ENTRIES_PER_SUBCRQ = 14,
+ REQ_RX_ADD_ENTRIES_PER_SUBCRQ = 15,
+ TCP_IP_OFFLOAD = 16,
+ PROMISC_REQUESTED = 17,
+ PROMISC_SUPPORTED = 18,
+ MIN_MTU = 19,
+ MAX_MTU = 20,
+ REQ_MTU = 21,
+ MAX_MULTICAST_FILTERS = 22,
+ VLAN_HEADER_INSERTION = 23,
+ MAX_TX_SG_ENTRIES = 25,
+ RX_SG_SUPPORTED = 26,
+ RX_SG_REQUESTED = 27,
+};
+
+enum ibmvnic_error_cause {
+ ADAPTER_PROBLEM = 0,
+ BUS_PROBLEM = 1,
+ FW_PROBLEM = 2,
+ DD_PROBLEM = 3,
+ EEH_RECOVERY = 4,
+ FW_UPDATED = 5,
+ LOW_MEMORY = 6,
+};
+
+enum ibmvnic_commands {
+ VERSION_EXCHANGE = 0x01,
+ VERSION_EXCHANGE_RSP = 0x81,
+ QUERY_CAPABILITY = 0x02,
+ QUERY_CAPABILITY_RSP = 0x82,
+ REQUEST_CAPABILITY = 0x03,
+ REQUEST_CAPABILITY_RSP = 0x83,
+ LOGIN = 0x04,
+ LOGIN_RSP = 0x84,
+ QUERY_PHYS_PARMS = 0x05,
+ QUERY_PHYS_PARMS_RSP = 0x85,
+ QUERY_PHYS_CAPABILITIES = 0x06,
+ QUERY_PHYS_CAPABILITIES_RSP = 0x86,
+ SET_PHYS_PARMS = 0x07,
+ SET_PHYS_PARMS_RSP = 0x87,
+ ERROR_INDICATION = 0x08,
+ REQUEST_ERROR_INFO = 0x09,
+ REQUEST_ERROR_RSP = 0x89,
+ REQUEST_DUMP_SIZE = 0x0A,
+ REQUEST_DUMP_SIZE_RSP = 0x8A,
+ REQUEST_DUMP = 0x0B,
+ REQUEST_DUMP_RSP = 0x8B,
+ LOGICAL_LINK_STATE = 0x0C,
+ LOGICAL_LINK_STATE_RSP = 0x8C,
+ REQUEST_STATISTICS = 0x0D,
+ REQUEST_STATISTICS_RSP = 0x8D,
+ REQUEST_RAS_COMP_NUM = 0x0E,
+ REQUEST_RAS_COMP_NUM_RSP = 0x8E,
+ REQUEST_RAS_COMPS = 0x0F,
+ REQUEST_RAS_COMPS_RSP = 0x8F,
+ CONTROL_RAS = 0x10,
+ CONTROL_RAS_RSP = 0x90,
+ COLLECT_FW_TRACE = 0x11,
+ COLLECT_FW_TRACE_RSP = 0x91,
+ LINK_STATE_INDICATION = 0x12,
+ CHANGE_MAC_ADDR = 0x13,
+ CHANGE_MAC_ADDR_RSP = 0x93,
+ MULTICAST_CTRL = 0x14,
+ MULTICAST_CTRL_RSP = 0x94,
+ GET_VPD_SIZE = 0x15,
+ GET_VPD_SIZE_RSP = 0x95,
+ GET_VPD = 0x16,
+ GET_VPD_RSP = 0x96,
+ TUNE = 0x17,
+ TUNE_RSP = 0x97,
+ QUERY_IP_OFFLOAD = 0x18,
+ QUERY_IP_OFFLOAD_RSP = 0x98,
+ CONTROL_IP_OFFLOAD = 0x19,
+ CONTROL_IP_OFFLOAD_RSP = 0x99,
+ ACL_CHANGE_INDICATION = 0x1A,
+ ACL_QUERY = 0x1B,
+ ACL_QUERY_RSP = 0x9B,
+ REQUEST_DEBUG_STATS = 0x1C,
+ REQUEST_DEBUG_STATS_RSP = 0x9C,
+};
+
+enum ibmvnic_crq_type {
+ IBMVNIC_CRQ_CMD = 0x80,
+ IBMVNIC_CRQ_CMD_RSP = 0x80,
+ IBMVNIC_CRQ_INIT_CMD = 0xC0,
+ IBMVNIC_CRQ_INIT_RSP = 0xC0,
+ IBMVNIC_CRQ_XPORT_EVENT = 0xFF,
+};
+
+enum ibmvfc_crq_format {
+ IBMVNIC_CRQ_INIT = 0x01,
+ IBMVNIC_CRQ_INIT_COMPLETE = 0x02,
+ IBMVNIC_PARTITION_MIGRATED = 0x06,
+};
+
+struct ibmvnic_crq_queue {
+ union ibmvnic_crq *msgs;
+ int size, cur;
+ dma_addr_t msg_token;
+ spinlock_t lock;
+};
+
+union sub_crq {
+ struct ibmvnic_generic_scrq generic;
+ struct ibmvnic_tx_comp_desc tx_comp;
+ struct ibmvnic_tx_desc_v0 v0;
+ struct ibmvnic_tx_desc_v1 v1;
+ struct ibmvnic_tx_desc_v2 v2;
+ struct ibmvnic_sge_desc sge;
+ struct ibmvnic_rx_desc rx_comp;
+ struct ibmvnic_rx_buff_add_desc rx_add;
+};
+
+struct ibmvnic_sub_crq_queue {
+ union sub_crq *msgs;
+ int size, cur;
+ dma_addr_t msg_token;
+ unsigned long crq_num;
+ unsigned long hw_irq;
+ unsigned int irq;
+ unsigned int pool_index;
+ int scrq_num;
+ spinlock_t lock;
+ struct sk_buff *rx_skb_top;
+ struct ibmvnic_adapter *adapter;
+};
+
+struct ibmvnic_tx_buff {
+ struct sk_buff *skb;
+ dma_addr_t data_dma[IBMVNIC_MAX_FRAGS_PER_CRQ];
+ unsigned int data_len[IBMVNIC_MAX_FRAGS_PER_CRQ];
+ int index;
+ int pool_index;
+ int last_frag;
+ int used_bounce;
+};
+
+struct ibmvnic_tx_pool {
+ struct ibmvnic_tx_buff *tx_buff;
+ int *free_map;
+ int consumer_index;
+ int producer_index;
+ wait_queue_head_t ibmvnic_tx_comp_q;
+ struct task_struct *work_thread;
+};
+
+struct ibmvnic_rx_buff {
+ struct sk_buff *skb;
+ dma_addr_t dma;
+ int size;
+ int pool_index;
+};
+
+struct ibmvnic_rx_pool {
+ struct ibmvnic_rx_buff *rx_buff;
+ int size;
+ int index;
+ int buff_size;
+ atomic_t available;
+ int *free_map;
+ int next_free;
+ int next_alloc;
+ int active;
+};
+
+struct ibmvnic_error_buff {
+ char *buff;
+ dma_addr_t dma;
+ int len;
+ struct list_head list;
+ __be32 error_id;
+};
+
+struct ibmvnic_fw_comp_internal {
+ struct ibmvnic_adapter *adapter;
+ int num;
+ struct debugfs_blob_wrapper desc_blob;
+ int paused;
+};
+
+struct ibmvnic_inflight_cmd {
+ union ibmvnic_crq crq;
+ struct list_head list;
+};
+
+struct ibmvnic_adapter {
+ struct vio_dev *vdev;
+ struct net_device *netdev;
+ struct ibmvnic_crq_queue crq;
+ u8 mac_addr[ETH_ALEN];
+ struct ibmvnic_query_ip_offload_buffer ip_offload_buf;
+ dma_addr_t ip_offload_tok;
+ struct ibmvnic_control_ip_offload_buffer ip_offload_ctrl;
+ dma_addr_t ip_offload_ctrl_tok;
+ int migrated;
+ u32 msg_enable;
+ void *bounce_buffer;
+ int bounce_buffer_size;
+ dma_addr_t bounce_buffer_dma;
+
+ /* Statistics */
+ struct net_device_stats net_stats;
+ struct ibmvnic_statistics stats;
+ dma_addr_t stats_token;
+ struct completion stats_done;
+ spinlock_t stats_lock;
+ int replenish_no_mem;
+ int replenish_add_buff_success;
+ int replenish_add_buff_failure;
+ int replenish_task_cycles;
+ int tx_send_failed;
+ int tx_map_failed;
+
+ int phys_link_state;
+ int logical_link_state;
+
+ /* login data */
+ struct ibmvnic_login_buffer *login_buf;
+ dma_addr_t login_buf_token;
+ int login_buf_sz;
+
+ struct ibmvnic_login_rsp_buffer *login_rsp_buf;
+ dma_addr_t login_rsp_buf_token;
+ int login_rsp_buf_sz;
+
+ atomic_t running_cap_queries;
+
+ struct ibmvnic_sub_crq_queue **tx_scrq;
+ struct ibmvnic_sub_crq_queue **rx_scrq;
+ int requested_caps;
+
+ /* rx structs */
+ struct napi_struct *napi;
+ struct ibmvnic_rx_pool *rx_pool;
+ u64 promisc;
+
+ struct ibmvnic_tx_pool *tx_pool;
+ int closing;
+ struct completion init_done;
+
+ struct list_head errors;
+ spinlock_t error_list_lock;
+
+ /* debugfs */
+ struct dentry *debugfs_dir;
+ struct dentry *debugfs_dump;
+ struct completion fw_done;
+ char *dump_data;
+ dma_addr_t dump_data_token;
+ int dump_data_size;
+ int ras_comp_num;
+ struct ibmvnic_fw_component *ras_comps;
+ struct ibmvnic_fw_comp_internal *ras_comp_int;
+ dma_addr_t ras_comps_tok;
+ struct dentry *ras_comps_ent;
+
+ /* in-flight commands that allocate and/or map memory*/
+ struct list_head inflight;
+ spinlock_t inflight_lock;
+
+ /* partner capabilities */
+ u64 min_tx_queues;
+ u64 min_rx_queues;
+ u64 min_rx_add_queues;
+ u64 max_tx_queues;
+ u64 max_rx_queues;
+ u64 max_rx_add_queues;
+ u64 req_tx_queues;
+ u64 req_rx_queues;
+ u64 req_rx_add_queues;
+ u64 min_tx_entries_per_subcrq;
+ u64 min_rx_add_entries_per_subcrq;
+ u64 max_tx_entries_per_subcrq;
+ u64 max_rx_add_entries_per_subcrq;
+ u64 req_tx_entries_per_subcrq;
+ u64 req_rx_add_entries_per_subcrq;
+ u64 tcp_ip_offload;
+ u64 promisc_requested;
+ u64 promisc_supported;
+ u64 min_mtu;
+ u64 max_mtu;
+ u64 req_mtu;
+ u64 max_multicast_filters;
+ u64 vlan_header_insertion;
+ u64 max_tx_sg_entries;
+ u64 rx_sg_supported;
+ u64 rx_sg_requested;
+};
--- a/drivers/net/ethernet/ibm/ibmvnic.c 1969-12-31 18:00:00.000000000 -0600
+++ b/drivers/net/ethernet/ibm/ibmvnic.c 2014-08-25 08:22:47.000000000 -0500
@@ -0,0 +1,3642 @@
+/**************************************************************************/
+/* */
+/* IBM System i and System p Virtual NIC Device Driver */
+/* Copyright (C) 2014 IBM Corp. */
+/* Santiago Leon (santil@linux.vnet.ibm.com) */
+/* */
+/* This program is free software; you can redistribute it and/or modify */
+/* it under the terms of the GNU General Public License as published by */
+/* the Free Software Foundation; either version 2 of the License, or */
+/* (at your option) any later version. */
+/* */
+/* This program is distributed in the hope that it will be useful, */
+/* but WITHOUT ANY WARRANTY; without even the implied warranty of */
+/* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the */
+/* GNU General Public License for more details. */
+/* */
+/* You should have received a copy of the GNU General Public License */
+/* along with this program; if not, write to the Free Software */
+/* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 */
+/* USA */
+/* */
+/* This module contains the implementation of a virtual ethernet device */
+/* for use with IBM i/pSeries LPAR Linux. It utilizes the logical LAN */
+/* option of the RS/6000 Platform Architecture to interface with virtual */
+/* ethernet NICs that are presented to the partition by the hypervisor. */
+/* */
+/**************************************************************************/
+/*
+ * TODO:
+*/
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/completion.h>
+#include <linux/ioport.h>
+#include <linux/dma-mapping.h>
+#include <linux/kernel.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/skbuff.h>
+#include <linux/mii.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/mm.h>
+#include <linux/ethtool.h>
+#include <linux/proc_fs.h>
+#include <linux/in.h>
+#include <linux/ip.h>
+#include <linux/irq.h>
+#include <linux/kthread.h>
+#include <linux/seq_file.h>
+#include <linux/debugfs.h>
+#include <linux/interrupt.h>
+#include <net/net_namespace.h>
+#include <asm/hvcall.h>
+#include <linux/atomic.h>
+#include <asm/vio.h>
+#include <asm/iommu.h>
+#include <linux/uaccess.h>
+#include <asm/firmware.h>
+#include <linux/seq_file.h>
+
+#include "ibmvnic.h"
+
+static const char ibmvnic_driver_name[] = "ibmvnic";
+static const char ibmvnic_driver_string[] = "IBM System i/p Virtual NIC Driver";
+
+MODULE_AUTHOR("Santiago Leon <santil@us.ibm.com>");
+MODULE_DESCRIPTION("IBM System i/p Virtual NIC Driver");
+MODULE_LICENSE("GPL");
+MODULE_VERSION(IBMVNIC_DRIVER_VERSION);
+
+static int ibmvnic_buffs_per_pool = IBMVNIC_BUFFS_PER_POOL;
+module_param(ibmvnic_buffs_per_pool, int, S_IRUGO);
+MODULE_PARM_DESC(ibmvnic_buffs_per_pool,
+ "IBMVNIC number of buffers per rx pool");
+
+static int ibmvnic_max_tx_queues = IBMVNIC_MAX_TX_QUEUES;
+module_param(ibmvnic_max_tx_queues, int, S_IRUGO);
+MODULE_PARM_DESC(ibmvnic_max_tx_queues, "IBMVNIC max number of tx queues");
+
+static int ibmvnic_req_mtu;
+module_param(ibmvnic_req_mtu, int, 0);
+MODULE_PARM_DESC(ibmvnic_req_mtu,
+ "IBMVNIC requested MTU, (0 = max allowed by hw)");
+
+static int ibmvnic_version = IBMVNIC_INITIAL_VERSION;
+static int ibmvnic_remove(struct vio_dev *);
+static void release_sub_crqs(struct ibmvnic_adapter *);
+static int ibmvnic_reset_crq(struct ibmvnic_adapter *);
+static int ibmvnic_send_crq_init(struct ibmvnic_adapter *);
+static int ibmvnic_reenable_crq_queue(struct ibmvnic_adapter *);
+static int ibmvnic_send_crq(struct ibmvnic_adapter *, union ibmvnic_crq *);
+static int send_subcrq(struct ibmvnic_adapter *adapter, u64 remote_handle,
+ union sub_crq *sub_crq);
+static irqreturn_t ibmvnic_interrupt_rx(int irq, void *instance);
+static int enable_scrq_irq(struct ibmvnic_adapter *,
+ struct ibmvnic_sub_crq_queue *);
+static int disable_scrq_irq(struct ibmvnic_adapter *,
+ struct ibmvnic_sub_crq_queue *);
+static int pending_scrq(struct ibmvnic_adapter *,
+ struct ibmvnic_sub_crq_queue *);
+static union sub_crq *ibmvnic_next_scrq(struct ibmvnic_adapter *,
+ struct ibmvnic_sub_crq_queue *);
+static int ibmvnic_tx_work(void *data);
+int ibmvnic_poll(struct napi_struct *, int);
+
+struct ibmvnic_stat {
+ char name[ETH_GSTRING_LEN];
+ int offset;
+};
+
+#define IBMVNIC_STAT_OFF(stat) (offsetof(struct ibmvnic_adapter, stats) + \
+ offsetof(struct ibmvnic_statistics, stat))
+#define IBMVNIC_GET_STAT(a, off) (*((u64 *)(((unsigned long)(a)) + off)))
+
+static const struct ibmvnic_stat ibmvnic_stats[] = {
+ {"rx_packets", IBMVNIC_STAT_OFF(rx_packets)},
+ {"rx_bytes", IBMVNIC_STAT_OFF(rx_bytes)},
+ {"tx_packets", IBMVNIC_STAT_OFF(tx_packets)},
+ {"tx_bytes", IBMVNIC_STAT_OFF(tx_bytes)},
+ {"ucast_tx_packets", IBMVNIC_STAT_OFF(ucast_tx_packets)},
+ {"ucast_rx_packets", IBMVNIC_STAT_OFF(ucast_rx_packets)},
+ {"mcast_tx_packets", IBMVNIC_STAT_OFF(mcast_tx_packets)},
+ {"mcast_rx_packets", IBMVNIC_STAT_OFF(mcast_rx_packets)},
+ {"bcast_tx_packets", IBMVNIC_STAT_OFF(bcast_tx_packets)},
+ {"bcast_rx_packets", IBMVNIC_STAT_OFF(bcast_rx_packets)},
+ {"align_errors", IBMVNIC_STAT_OFF(align_errors)},
+ {"fcs_errors", IBMVNIC_STAT_OFF(fcs_errors)},
+ {"single_collision_frames", IBMVNIC_STAT_OFF(single_collision_frames)},
+ {"multi_collision_frames", IBMVNIC_STAT_OFF(multi_collision_frames)},
+ {"sqe_test_errors", IBMVNIC_STAT_OFF(sqe_test_errors)},
+ {"deferred_tx", IBMVNIC_STAT_OFF(deferred_tx)},
+ {"late_collisions", IBMVNIC_STAT_OFF(late_collisions)},
+ {"excess_collisions", IBMVNIC_STAT_OFF(excess_collisions)},
+ {"internal_mac_tx_errors", IBMVNIC_STAT_OFF(internal_mac_tx_errors)},
+ {"carrier_sense", IBMVNIC_STAT_OFF(carrier_sense)},
+ {"too_long_frames", IBMVNIC_STAT_OFF(too_long_frames)},
+ {"internal_mac_rx_errors", IBMVNIC_STAT_OFF(internal_mac_rx_errors)},
+};
+
+static long h_reg_sub_crq(unsigned long unit_address, unsigned long token,
+ unsigned long length, unsigned long *number,
+ unsigned long *irq)
+{
+ long rc;
+ unsigned long retbuf[PLPAR_HCALL_BUFSIZE];
+
+ rc = plpar_hcall(H_REG_SUB_CRQ, retbuf, unit_address, token, length);
+ *number = retbuf[0];
+ *irq = retbuf[1];
+
+ return rc;
+}
+
+/* net_device_ops functions */
+static void init_rx_pool(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_rx_pool *rx_pool, int num, int index,
+ int buff_size, int active)
+{
+ netdev_dbg(adapter->netdev,
+ "Initializing rx_pool %d, %d buffs, %d bytes each\n",
+ index, num, buff_size);
+ rx_pool->size = num;
+ rx_pool->index = index;
+ rx_pool->buff_size = buff_size;
+ rx_pool->active = active;
+}
+
+static int alloc_rx_pool(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_rx_pool *pool)
+{
+ struct device *dev = &adapter->vdev->dev;
+ int i;
+
+ pool->free_map = kcalloc(pool->size, sizeof(int), GFP_KERNEL);
+ if (!pool->free_map)
+ return -ENOMEM;
+
+ pool->rx_buff = kcalloc(pool->size, sizeof(struct ibmvnic_rx_buff),
+ GFP_KERNEL);
+ if (!pool->rx_buff) {
+ dev_err(dev, "Couldn't alloc rx buffers\n");
+ kfree(pool->free_map);
+ return -ENOMEM;
+ }
+
+ for (i = 0; i < pool->size; ++i)
+ pool->free_map[i] = i;
+
+ atomic_set(&pool->available, 0);
+ pool->next_alloc = 0;
+ pool->next_free = 0;
+
+ return 0;
+}
+
+static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_rx_pool *pool)
+{
+ int i;
+ int count = pool->size - atomic_read(&pool->available);
+ int buffers_added = 0;
+ struct sk_buff *skb;
+ int index;
+ unsigned long lpar_rc;
+ dma_addr_t dma_addr;
+ struct device *dev = &adapter->vdev->dev;
+ union sub_crq sub_crq;
+ u64 *handle_array = (u64 *)((u8 *)(adapter->login_rsp_buf) +
+ be32_to_cpu(adapter->login_rsp_buf->
+ off_rxadd_subcrqs));
+
+ for (i = 0; i < count; ++i) {
+ skb = alloc_skb(pool->buff_size, GFP_ATOMIC);
+ if (!skb) {
+ dev_err(dev, "Couldn't replenish rx buff\n");
+ adapter->replenish_no_mem++;
+ break;
+ }
+
+ index = pool->free_map[pool->next_free];
+
+ if (pool->rx_buff[index].skb != NULL)
+ dev_err(dev, "Inconsistent free_map!\n");
+
+ dma_addr = dma_map_single(dev, skb->data, pool->buff_size,
+ DMA_FROM_DEVICE);
+ if (dma_mapping_error(&adapter->vdev->dev, dma_addr))
+ goto failure;
+
+ pool->free_map[pool->next_free] = IBMVNIC_INVALID_MAP;
+ pool->rx_buff[index].dma = dma_addr;
+ pool->rx_buff[index].skb = skb;
+ pool->rx_buff[index].pool_index = pool->index;
+ pool->rx_buff[index].size = pool->buff_size;
+
+ sub_crq.rx_add.first = IBMVNIC_CRQ_CMD;
+ sub_crq.rx_add.correlator =
+ cpu_to_be64((u64)&pool->rx_buff[index]);
+ sub_crq.rx_add.ioba = cpu_to_be32(dma_addr);
+ sub_crq.rx_add.len = cpu_to_be32(pool->buff_size);
+
+ lpar_rc = send_subcrq(adapter, handle_array[pool->index],
+ &sub_crq);
+ if (lpar_rc != H_SUCCESS)
+ goto failure;
+
+ buffers_added++;
+ adapter->replenish_add_buff_success++;
+ pool->next_free = (pool->next_free + 1) % pool->size;
+ }
+ atomic_add(buffers_added, &(pool->available));
+ return;
+
+failure:
+ pool->free_map[pool->next_free] = index;
+ pool->rx_buff[index].skb = NULL;
+ if (!dma_mapping_error(dev, dma_addr))
+ dma_unmap_single(dev, dma_addr, pool->buff_size,
+ DMA_FROM_DEVICE);
+
+ dev_kfree_skb_any(skb);
+ adapter->replenish_add_buff_failure++;
+ atomic_add(buffers_added, &(pool->available));
+
+}
+
+static void replenish_pools(struct ibmvnic_adapter *adapter)
+{
+ int i;
+
+ if (adapter->migrated)
+ return;
+
+ adapter->replenish_task_cycles++;
+ for (i = 0; i < be32_to_cpu(adapter->login_rsp_buf->num_rxadd_subcrqs);
+ i++) {
+ if (adapter->rx_pool[i].active)
+ replenish_rx_pool(adapter, &adapter->rx_pool[i]);
+ }
+}
+
+static void free_rx_pool(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_rx_pool *pool)
+{
+ struct device *dev = &adapter->vdev->dev;
+ int i;
+
+ kfree(pool->free_map);
+ pool->free_map = NULL;
+
+ if (!pool->rx_buff)
+ return;
+
+ for (i = 0; i < pool->size; i++) {
+ if (pool->rx_buff[i].skb) {
+ dma_unmap_single(dev, pool->rx_buff[i].dma,
+ pool->rx_buff[i].size,
+ DMA_FROM_DEVICE);
+ dev_kfree_skb_any(pool->rx_buff[i].skb);
+ pool->rx_buff[i].skb = NULL;
+ }
+ }
+ kfree(pool->rx_buff);
+ pool->rx_buff = NULL;
+}
+
+static int ibmvnic_open(struct net_device *netdev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ struct device *dev = &adapter->vdev->dev;
+ union ibmvnic_crq crq;
+ int i, j;
+ int rxadd_subcrqs =
+ be32_to_cpu(adapter->login_rsp_buf->num_rxadd_subcrqs);
+ int tx_subcrqs =
+ be32_to_cpu(adapter->login_rsp_buf->num_txsubm_subcrqs);
+ u64 *size_array = (u64 *)((u8 *)(adapter->login_rsp_buf) +
+ be32_to_cpu(adapter->login_rsp_buf->
+ off_rxadd_buff_size));
+
+ adapter->napi = kcalloc(adapter->req_rx_queues,
+ sizeof(struct napi_struct), GFP_KERNEL);
+ if (!adapter->napi)
+ goto alloc_napi_failed;
+
+ for (i = 0; i < adapter->req_rx_queues; i++) {
+ netif_napi_add(netdev, &adapter->napi[i], ibmvnic_poll,
+ NAPI_POLL_WEIGHT);
+ napi_enable(&adapter->napi[i]);
+ }
+
+ adapter->rx_pool =
+ kcalloc(rxadd_subcrqs, sizeof(struct ibmvnic_rx_pool), GFP_KERNEL);
+
+ if (!adapter->rx_pool)
+ goto rx_pool_arr_alloc_failed;
+
+ for (i = 0; i < rxadd_subcrqs; i++) {
+ if (!adapter->rx_pool->active)
+ continue;
+ init_rx_pool(adapter, &adapter->rx_pool[i],
+ ibmvnic_buffs_per_pool, i, size_array[i], 1);
+ if (alloc_rx_pool(adapter, &adapter->rx_pool[i])) {
+ dev_err(dev, "Couldn't alloc rx pool\n");
+ goto rx_pool_alloc_failed;
+ }
+ }
+
+ adapter->tx_pool =
+ kcalloc(rxadd_subcrqs, sizeof(struct ibmvnic_tx_pool), GFP_KERNEL);
+
+ if (!adapter->tx_pool)
+ goto tx_pool_arr_alloc_failed;
+
+ for (i = 0; i < tx_subcrqs; i++) {
+ adapter->tx_pool[i].tx_buff =
+ kcalloc(adapter->max_tx_entries_per_subcrq,
+ sizeof(struct ibmvnic_tx_buff), GFP_KERNEL);
+ if (!adapter->tx_pool[i].tx_buff)
+ goto tx_pool_alloc_failed;
+
+ adapter->tx_pool[i].free_map =
+ kcalloc(adapter->max_tx_entries_per_subcrq,
+ sizeof(int), GFP_KERNEL);
+ if (!adapter->tx_pool[i].free_map)
+ goto tx_fm_alloc_failed;
+
+ for (j = 0; j < adapter->max_tx_entries_per_subcrq; j++)
+ adapter->tx_pool[i].free_map[j] = j;
+
+ adapter->tx_pool[i].consumer_index = 0;
+ adapter->tx_pool[i].producer_index = 0;
+
+ init_waitqueue_head(&adapter->tx_pool[i].ibmvnic_tx_comp_q);
+ adapter->tx_pool[i].work_thread =
+ kthread_run(ibmvnic_tx_work, adapter->tx_scrq[i],
+ "%s_%s_%d",
+ IBMVNIC_NAME, adapter->netdev->name, i);
+ if (IS_ERR(adapter->tx_pool[i].work_thread)) {
+ dev_err(dev, "Couldn't create kernel thread: %ld\n",
+ PTR_ERR(adapter->tx_pool[i].work_thread));
+ goto thread_failed;
+ }
+ }
+
+ adapter->bounce_buffer_size =
+ (netdev->mtu + ETH_HLEN - 1) / PAGE_SIZE + 1;
+ adapter->bounce_buffer = kmalloc(adapter->bounce_buffer_size,
+ GFP_KERNEL);
+ if (!adapter->bounce_buffer)
+ goto bounce_alloc_failed;
+
+ adapter->bounce_buffer_dma = dma_map_single(dev, adapter->bounce_buffer,
+ adapter->bounce_buffer_size,
+ DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, adapter->bounce_buffer_dma)) {
+ dev_err(dev, "Couldn't map tx bounce buffer\n");
+ goto bounce_map_failed;
+ }
+
+ replenish_pools(adapter);
+
+ /* We're ready to receive frames, enable the sub-crq interrupts and
+ * set the logical link state to up
+ */
+ for (i = 0; i < adapter->req_rx_queues; i++)
+ enable_scrq_irq(adapter, adapter->rx_scrq[i]);
+
+ for (i = 0; i < adapter->req_tx_queues; i++)
+ enable_scrq_irq(adapter, adapter->tx_scrq[i]);
+
+ memset(&crq, 0, sizeof(crq));
+ crq.logical_link_state.first = IBMVNIC_CRQ_CMD;
+ crq.logical_link_state.cmd = LOGICAL_LINK_STATE;
+ crq.logical_link_state.link_state = IBMVNIC_LOGICAL_LNK_UP;
+ ibmvnic_send_crq(adapter, &crq);
+
+ netif_start_queue(netdev);
+ return 0;
+
+bounce_map_failed:
+ kfree(adapter->bounce_buffer);
+bounce_alloc_failed:
+ i = tx_subcrqs;
+thread_failed:
+ for (j = 0; j < i; j++)
+ kthread_stop(adapter->tx_pool[j].work_thread);
+tx_fm_alloc_failed:
+ i = tx_subcrqs;
+tx_pool_alloc_failed:
+ for (j = 0; j < i; j++)
+ kfree(adapter->tx_pool[j].tx_buff);
+ kfree(adapter->tx_pool);
+ adapter->tx_pool = NULL;
+tx_pool_arr_alloc_failed:
+ i = rxadd_subcrqs;
+rx_pool_alloc_failed:
+ for (j = 0; j < i; j++)
+ free_rx_pool(adapter, &adapter->rx_pool[j]);
+ kfree(adapter->rx_pool);
+ adapter->rx_pool = NULL;
+rx_pool_arr_alloc_failed:
+ for (i = 0; i < adapter->req_rx_queues; i++)
+ napi_enable(&adapter->napi[i]);
+alloc_napi_failed:
+ return -ENOMEM;
+}
+
+static int ibmvnic_close(struct net_device *netdev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ struct device *dev = &adapter->vdev->dev;
+ int i;
+
+ adapter->closing = 1;
+ for (i = 0; i < adapter->req_tx_queues; i++) {
+ wake_up(&adapter->tx_pool[i].ibmvnic_tx_comp_q);
+ kthread_stop(adapter->tx_pool[i].work_thread);
+ }
+
+ for (i = 0; i < adapter->req_rx_queues; i++)
+ napi_disable(&adapter->napi[i]);
+
+ netif_stop_queue(netdev);
+
+ if (adapter->bounce_buffer != NULL) {
+ if (!dma_mapping_error(dev, adapter->bounce_buffer_dma)) {
+ dma_unmap_single(&adapter->vdev->dev,
+ adapter->bounce_buffer_dma,
+ adapter->bounce_buffer_size,
+ DMA_BIDIRECTIONAL);
+ adapter->bounce_buffer_dma = DMA_ERROR_CODE;
+ }
+ kfree(adapter->bounce_buffer);
+ adapter->bounce_buffer = NULL;
+ }
+
+ for (i = 0; i < be32_to_cpu(adapter->login_rsp_buf->num_txsubm_subcrqs);
+ i++)
+ kfree(adapter->tx_pool[i].tx_buff);
+ kfree(adapter->tx_pool);
+ adapter->tx_pool = NULL;
+
+ for (i = 0; i < be32_to_cpu(adapter->login_rsp_buf->num_rxadd_subcrqs);
+ i++)
+ free_rx_pool(adapter, &adapter->rx_pool[i]);
+ kfree(adapter->rx_pool);
+ adapter->rx_pool = NULL;
+
+ adapter->closing = 0;
+
+ return 0;
+}
+
+static int ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ struct device *dev = &adapter->vdev->dev;
+ dma_addr_t data_dma_addr;
+ union sub_crq tx_crq;
+ unsigned long lpar_rc;
+ unsigned int tx_dropped = 0;
+ unsigned int tx_bytes = 0;
+ unsigned int tx_packets = 0;
+ unsigned int tx_send_failed = 0;
+ unsigned int tx_map_failed = 0;
+ unsigned int nr_frags;
+ int ret = 0;
+ int index = 0;
+ int i;
+ int used_bounce = 0;
+ struct skb_frag_struct *frag = NULL;
+ int queue_num = skb_get_queue_mapping(skb);
+ struct ibmvnic_tx_pool *tx_pool = &adapter->tx_pool[queue_num];
+ struct ibmvnic_control_ip_offload_buffer *offload =
+ &adapter->ip_offload_ctrl;
+ struct netdev_queue *txq =
+ netdev_get_tx_queue(netdev, skb_get_queue_mapping(skb));
+ u64 *handle_array = (u64 *)((u8 *)(adapter->login_rsp_buf) +
+ be32_to_cpu(adapter->login_rsp_buf->
+ off_txsubm_subcrqs));
+ u8 offload_flags = 0;
+
+ if (adapter->migrated) {
+ tx_send_failed++;
+ tx_dropped++;
+ ret = NETDEV_TX_BUSY;
+ goto out;
+ }
+
+ if (skb->ip_summed == CHECKSUM_PARTIAL) {
+ if (offload->tcp_ipv4_chksum &&
+ ip_hdr(skb)->protocol == IPPROTO_TCP)
+ offload_flags |= IBMVNIC_TCP_CHKSUM;
+ else if (offload->udp_ipv4_chksum &&
+ ip_hdr(skb)->protocol == IPPROTO_UDP)
+ offload_flags |= IBMVNIC_UDP_CHKSUM;
+ else if (skb_checksum_help(skb)) {
+ dev_err(dev, "tx: failed to checksum packet\n");
+ tx_dropped++;
+ goto out;
+ }
+ }
+
+ /* If the hardware can't handle the number of memory segments,
+ * try to linearize it
+ */
+ if (skb_shinfo(skb)->nr_frags > adapter->max_tx_sg_entries) {
+ if (skb_linearize(skb)) {
+ kfree_skb(skb);
+ tx_dropped++;
+ return NETDEV_TX_OK;
+ }
+ }
+
+ nr_frags = skb_shinfo(skb)->nr_frags;
+ if (!nr_frags) {
+ data_dma_addr = dma_map_single(dev, skb->data, skb->len,
+ DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, data_dma_addr)) {
+ if (!firmware_has_feature(FW_FEATURE_CMO))
+ dev_err(dev, "tx: unable to map xmit buffer\n");
+ skb_copy_from_linear_data(skb, adapter->bounce_buffer,
+ skb->len);
+ data_dma_addr = adapter->bounce_buffer_dma;
+ used_bounce = 1;
+ tx_map_failed++;
+ tx_dropped++;
+ ret = NETDEV_TX_BUSY;
+ goto out;
+ }
+
+ index = tx_pool->free_map[tx_pool->consumer_index];
+ tx_pool->consumer_index =
+ ++index % adapter->max_tx_entries_per_subcrq;
+
+ memset(&tx_crq, 0, sizeof(tx_crq));
+ tx_crq.v2.first = IBMVNIC_CRQ_CMD;
+ tx_crq.v2.type = IBMVNIC_TX_DESC_V2;
+ tx_crq.v2.n_crq_elem = 1;
+ tx_crq.v2.flags1 = IBMVNIC_TXV2_COMP_NEEDED;
+ tx_crq.v2.correlator = cpu_to_be32(index);
+ tx_crq.v2.dma_reg = cpu_to_be32(0);
+ tx_crq.v2.sge_len = cpu_to_be32(skb->len);
+ tx_crq.v2.ioba = cpu_to_be32(data_dma_addr);
+ if (adapter->vlan_header_insertion) {
+ tx_crq.v2.flags2 |= IBMVNIC_TXV2_VLAN_INSERT;
+ tx_crq.v2.vlan_id = cpu_to_be16(skb->vlan_tci);
+ }
+ if (skb_shinfo(skb)->gso_size) {
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_LSO;
+ tx_crq.v2.mss = cpu_to_be16(skb_shinfo(skb)->gso_size);
+ }
+ if (ip_hdr(skb)->version == 4)
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_PROT_IPV4;
+ else if (ip_hdr(skb)->version == 6)
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_PROT_IPV6;
+
+ if (ip_hdr(skb)->protocol != IPPROTO_TCP)
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_PROT_TCP;
+ else if (ip_hdr(skb)->protocol != IPPROTO_TCP)
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_PROT_UDP;
+
+ lpar_rc = send_subcrq(adapter, handle_array[0], &tx_crq);
+
+ if (lpar_rc != H_SUCCESS) {
+ dev_err(dev, "tx failed with code %ld\n", lpar_rc);
+ if (!used_bounce)
+ dma_unmap_single(dev, data_dma_addr, skb->len,
+ DMA_TO_DEVICE);
+
+ if (tx_pool->consumer_index == 0)
+ tx_pool->consumer_index =
+ adapter->max_tx_entries_per_subcrq - 1;
+ else
+ tx_pool->consumer_index--;
+
+ tx_send_failed++;
+ tx_dropped++;
+ ret = NETDEV_TX_BUSY;
+ goto out;
+ }
+ tx_packets++;
+ tx_bytes += skb->len;
+ txq->trans_start = jiffies;
+ ret = NETDEV_TX_OK;
+ goto out;
+ }
+
+ /* nr_frags > 0 */
+ for (i = 0; i < nr_frags; i++) {
+ void *data;
+
+ frag = &skb_shinfo(skb)->frags[i];
+ data = skb_frag_address(frag);
+ data_dma_addr = dma_map_single(dev, data, frag->size,
+ DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, data_dma_addr)) {
+ dev_err(dev, "tx: unable to map xmit buffer\n");
+ tx_map_failed++;
+ tx_dropped++;
+ ret = NETDEV_TX_BUSY;
+ goto out;
+ }
+
+ if (i == 0) { /* first frag */
+ memset(&tx_crq, 0, sizeof(tx_crq));
+ tx_crq.v2.first = IBMVNIC_CRQ_CMD;
+ tx_crq.v2.type = IBMVNIC_TX_DESC_V2;
+ tx_crq.v2.n_crq_elem = (nr_frags / 2) + 1;
+ tx_crq.v2.flags1 = IBMVNIC_TXV2_COMP_NEEDED;
+ tx_crq.v2.correlator = cpu_to_be32(index);
+ tx_crq.v2.dma_reg = cpu_to_be32(0);
+ tx_crq.v2.sge_len = cpu_to_be32(frag->size);
+ tx_crq.v2.ioba = cpu_to_be32(data_dma_addr);
+
+ if (adapter->vlan_header_insertion) {
+ tx_crq.v2.flags2 |= IBMVNIC_TXV2_VLAN_INSERT;
+ tx_crq.v2.vlan_id = cpu_to_be16(skb->vlan_tci);
+ }
+ if (skb_shinfo(skb)->gso_size) {
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_LSO;
+ tx_crq.v2.mss =
+ cpu_to_be16(skb_shinfo(skb)->gso_size);
+ }
+ if (ip_hdr(skb)->version == 4)
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_PROT_IPV4;
+ else if (ip_hdr(skb)->version == 6)
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_PROT_IPV6;
+
+ if (ip_hdr(skb)->protocol != IPPROTO_TCP)
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_PROT_TCP;
+ else if (ip_hdr(skb)->protocol != IPPROTO_TCP)
+ tx_crq.v2.flags1 |= IBMVNIC_TXV2_PROT_UDP;
+
+ lpar_rc =
+ send_subcrq(adapter, handle_array[0], &tx_crq);
+
+ if (lpar_rc != H_SUCCESS) {
+ dev_err(dev, "tx failed with code %ld\n",
+ lpar_rc);
+ if (!used_bounce)
+ dma_unmap_single(dev, data_dma_addr,
+ skb->len,
+ DMA_TO_DEVICE);
+
+ if (tx_pool->consumer_index == 0)
+ tx_pool->consumer_index =
+ adapter->max_tx_entries_per_subcrq -
+ 1;
+ else
+ tx_pool->consumer_index--;
+
+ tx_send_failed++;
+ tx_dropped++;
+ ret = NETDEV_TX_BUSY;
+ goto out;
+ }
+ tx_packets++;
+ tx_bytes += skb->len;
+ txq->trans_start = jiffies;
+ ret = NETDEV_TX_OK;
+ } else if (i % 2) { /* odd frags */
+ index = tx_pool->free_map[tx_pool->consumer_index];
+ tx_pool->consumer_index =
+ ++index % adapter->max_tx_entries_per_subcrq;
+
+ tx_crq.sge.first = IBMVNIC_CRQ_CMD;
+ tx_crq.sge.type = IBMVNIC_SGE_DESC;
+ tx_crq.sge.sge1_dma_reg = cpu_to_be32(0);
+ tx_crq.sge.sge1_len = cpu_to_be32(frag->size);
+ tx_crq.sge.sge1_ioba = cpu_to_be32(data_dma_addr);
+
+ /* Send CRQ only if it's the last frag */
+ if (i != nr_frags - 1)
+ continue;
+
+ lpar_rc =
+ send_subcrq(adapter, handle_array[0], &tx_crq);
+
+ if (lpar_rc != H_SUCCESS) {
+ dev_err(dev, "tx failed with code %ld\n",
+ lpar_rc);
+ if (!used_bounce)
+ dma_unmap_single(dev, data_dma_addr,
+ skb->len,
+ DMA_TO_DEVICE);
+
+ if (tx_pool->consumer_index == 0)
+ tx_pool->consumer_index =
+ adapter->max_tx_entries_per_subcrq -
+ 1;
+ else
+ tx_pool->consumer_index--;
+
+ tx_send_failed++;
+ tx_dropped++;
+ ret = NETDEV_TX_BUSY;
+ goto out;
+ }
+ tx_packets++;
+ tx_bytes += skb->len;
+ txq->trans_start = jiffies;
+ ret = NETDEV_TX_OK;
+ } else { /* even frags */
+ index = tx_pool->free_map[tx_pool->consumer_index];
+ tx_pool->consumer_index =
+ ++index % adapter->max_tx_entries_per_subcrq;
+
+ tx_crq.sge.sge2_dma_reg = cpu_to_be32(0);
+ tx_crq.sge.sge2_len = cpu_to_be32(frag->size);
+ tx_crq.sge.sge2_ioba = cpu_to_be32(data_dma_addr);
+
+ lpar_rc =
+ send_subcrq(adapter, handle_array[0], &tx_crq);
+
+ if (lpar_rc != H_SUCCESS) {
+ dev_err(dev, "tx failed with code %ld\n",
+ lpar_rc);
+ if (!used_bounce)
+ dma_unmap_single(dev, data_dma_addr,
+ skb->len,
+ DMA_TO_DEVICE);
+
+ if (tx_pool->consumer_index == 0)
+ tx_pool->consumer_index =
+ adapter->max_tx_entries_per_subcrq -
+ 1;
+ else
+ tx_pool->consumer_index--;
+
+ tx_send_failed++;
+ tx_dropped++;
+ ret = NETDEV_TX_BUSY;
+ goto out;
+ }
+ tx_packets++;
+ tx_bytes += skb->len;
+ txq->trans_start = jiffies;
+ ret = NETDEV_TX_OK;
+ }
+ }
+
+out:
+ adapter->net_stats.tx_dropped += tx_dropped;
+ adapter->net_stats.tx_bytes += tx_bytes;
+ adapter->net_stats.tx_packets += tx_packets;
+ adapter->tx_send_failed += tx_send_failed;
+ adapter->tx_map_failed += tx_map_failed;
+
+ return ret;
+}
+
+static void ibmvnic_set_multi(struct net_device *netdev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ union ibmvnic_crq crq;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.request_capability.first = IBMVNIC_CRQ_CMD;
+ crq.request_capability.cmd = REQUEST_CAPABILITY;
+
+ if (netdev->flags & IFF_PROMISC) {
+ if (!adapter->promisc_supported)
+ return;
+ crq.request_capability.capability =
+ cpu_to_be16(PROMISC_REQUESTED);
+ crq.request_capability.number = cpu_to_be32(1);
+ ibmvnic_send_crq(adapter, &crq);
+ } else {
+ crq.request_capability.capability =
+ cpu_to_be16(PROMISC_REQUESTED);
+ crq.request_capability.number = cpu_to_be32(0);
+ ibmvnic_send_crq(adapter, &crq);
+
+ if (netdev->flags & IFF_ALLMULTI) {
+ /* Accept all multicast */
+ memset(&crq, 0, sizeof(crq));
+ crq.multicast_ctrl.first = IBMVNIC_CRQ_CMD;
+ crq.multicast_ctrl.cmd = MULTICAST_CTRL;
+ crq.multicast_ctrl.flags = IBMVNIC_ENABLE_ALL;
+ ibmvnic_send_crq(adapter, &crq);
+ } else if (netdev_mc_empty(netdev)) {
+ /* Reject all multicast */
+ memset(&crq, 0, sizeof(crq));
+ crq.multicast_ctrl.first = IBMVNIC_CRQ_CMD;
+ crq.multicast_ctrl.cmd = MULTICAST_CTRL;
+ crq.multicast_ctrl.flags = IBMVNIC_DISABLE_ALL;
+ ibmvnic_send_crq(adapter, &crq);
+ } else {
+ /* Accept one or more multicast(s) */
+ int i;
+
+ for (i = 0; i < netdev_mc_count(netdev); i++) {
+ memset(&crq, 0, sizeof(crq));
+ crq.multicast_ctrl.first = IBMVNIC_CRQ_CMD;
+ crq.multicast_ctrl.cmd = MULTICAST_CTRL;
+ crq.multicast_ctrl.flags = IBMVNIC_ENABLE_MC;
+ ibmvnic_send_crq(adapter, &crq);
+ }
+ }
+ }
+}
+
+static int ibmvnic_set_mac(struct net_device *netdev, void *p)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ struct sockaddr *addr = p;
+ union ibmvnic_crq crq;
+
+ if (!is_valid_ether_addr(addr->sa_data))
+ return -EADDRNOTAVAIL;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.change_mac_addr.first = IBMVNIC_CRQ_CMD;
+ crq.change_mac_addr.cmd = CHANGE_MAC_ADDR;
+ memcpy(&crq.change_mac_addr.mac_addr[0], addr->sa_data, ETH_ALEN);
+ ibmvnic_send_crq(adapter, &crq);
+
+ return 0;
+}
+
+static int ibmvnic_mii_ioctl(struct net_device *netdev, struct ifreq *ifr,
+ int cmd)
+{
+ struct mii_ioctl_data *data = if_mii(ifr);
+
+ switch (cmd) {
+ case SIOCGMIIPHY:
+ break;
+ case SIOCGMIIREG:
+ switch (data->reg_num & 0x1F) {
+ case MII_BMCR:
+ break;
+ case MII_BMSR:
+ break;
+ case MII_PHYSID1:
+ break;
+ case MII_PHYSID2:
+ break;
+ case MII_ADVERTISE:
+ break;
+ case MII_LPA:
+ break;
+ case MII_EXPANSION:
+ break;
+ case MII_CTRL1000:
+ break;
+ case MII_STAT1000:
+ break;
+ case MII_ESTATUS:
+ break;
+ default:
+ return -EIO;
+ }
+ break;
+ case SIOCSMIIREG:
+ default:
+ return -EOPNOTSUPP;
+ }
+ return 0;
+}
+
+static int ibmvnic_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
+{
+ switch (cmd) {
+ case SIOCGMIIPHY:
+ case SIOCGMIIREG:
+ case SIOCSMIIREG:
+ return ibmvnic_mii_ioctl(netdev, ifr, cmd);
+ default:
+ return -EOPNOTSUPP;
+ }
+}
+
+static int ibmvnic_change_mtu(struct net_device *netdev, int new_mtu)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+
+ if (new_mtu > adapter->req_mtu)
+ return -EINVAL;
+
+ netdev->mtu = new_mtu;
+ return 0;
+}
+
+static void ibmvnic_tx_timeout(struct net_device *dev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(dev);
+ int rc;
+
+ /* Adapter timed out, resetting it */
+ release_sub_crqs(adapter);
+ rc = ibmvnic_reset_crq(adapter);
+ if (rc)
+ dev_err(&adapter->vdev->dev, "Adapter timeout, reset failed\n");
+ else
+ ibmvnic_send_crq_init(adapter);
+}
+
+static struct net_device_stats *ibmvnic_get_stats(struct net_device *dev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(dev);
+
+ /* only return the current stats */
+ return &adapter->net_stats;
+}
+
+static void remove_buff_from_pool(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_rx_buff *rx_buff)
+{
+ struct device *dev = &adapter->vdev->dev;
+ struct ibmvnic_rx_pool *pool = &adapter->rx_pool[rx_buff->pool_index];
+
+ rx_buff->skb = NULL;
+ dma_unmap_single(dev, rx_buff->dma, rx_buff->size, DMA_FROM_DEVICE);
+
+ pool->free_map[pool->next_alloc] = (int)(rx_buff - pool->rx_buff);
+ pool->next_alloc = (pool->next_alloc + 1) % pool->size;
+
+ atomic_dec(&(pool->available));
+}
+
+int ibmvnic_poll(struct napi_struct *napi, int budget)
+{
+ struct net_device *netdev = napi->dev;
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ int frames_processed = 0;
+ int scrq_num = (int)(napi - adapter->napi);
+
+restart_poll:
+ do {
+ struct sk_buff *skb;
+ struct ibmvnic_rx_buff *rx_buff;
+ union sub_crq *next;
+ int length;
+ int offset;
+ u16 csum;
+ u8 flags;
+ struct sk_buff *rx_top =
+ adapter->rx_scrq[scrq_num]->rx_skb_top;
+
+ if (!pending_scrq(adapter, adapter->rx_scrq[scrq_num]))
+ break;
+
+ next = ibmvnic_next_scrq(adapter, adapter->rx_scrq[scrq_num]);
+ rx_buff =
+ (struct ibmvnic_rx_buff *)be64_to_cpu(next->
+ rx_comp.correlator);
+ length = be32_to_cpu(next->rx_comp.len);
+ offset = be16_to_cpu(next->rx_comp.off_frame_data);
+ flags = next->rx_comp.flags;
+ csum = be16_to_cpu(next->rx_comp.csum);
+ skb = rx_buff->skb;
+
+ /* free the entry */
+ next->rx_comp.first = 0;
+ remove_buff_from_pool(adapter, rx_buff);
+
+ /* handle large rx offload */
+ if (!(flags & IBMVNIC_END_FRAME)) {
+ /* this is the first or middle frame */
+ if (!rx_top) {
+ /* first frame */
+ adapter->rx_scrq[scrq_num]->rx_skb_top = skb;
+ skb_add_rx_frag(skb, 0, virt_to_page(skb->data),
+ ((long)skb->data & ~PAGE_MASK) +
+ offset, length, length);
+ } else {
+ /* middle frame */
+ skb_add_rx_frag(skb,
+ skb_shinfo(rx_top)->nr_frags,
+ virt_to_page(skb->data),
+ ((long)skb->data & ~PAGE_MASK) +
+ offset, length, length);
+ /* get a reference to the data, free the skb */
+ atomic_inc(&(skb_shinfo(skb)->dataref));
+ dev_kfree_skb_any(skb);
+ }
+ } else {
+ if (rx_top) {
+ /* last frame */
+ skb_add_rx_frag(skb,
+ skb_shinfo(rx_top)->nr_frags,
+ virt_to_page(skb->data),
+ ((long)skb->data & ~PAGE_MASK) +
+ offset, length, length);
+ atomic_inc(&(skb_shinfo(skb)->dataref));
+ dev_kfree_skb_any(skb);
+ skb = rx_top;
+ adapter->rx_scrq[scrq_num]->rx_skb_top = NULL;
+ } else {
+ /* no multi-frame packet */
+ skb_reserve(skb, offset);
+ skb_put(skb, length);
+ }
+ }
+
+ skb->protocol = eth_type_trans(skb, netdev);
+
+ if (flags & IBMVNIC_IP_CHKSUM_GOOD &&
+ flags & IBMVNIC_TCP_UDP_CHKSUM_GOOD) {
+ if (flags & IBMVNIC_CSUM_FIELD) {
+ skb->csum = csum_unfold((__force __sum16) csum);
+ skb->ip_summed = CHECKSUM_COMPLETE;
+ } else {
+ skb->ip_summed = CHECKSUM_UNNECESSARY;
+ }
+ }
+
+ length = skb->len;
+ napi_gro_receive(napi, skb); /* send it up */
+
+ adapter->net_stats.rx_packets++;
+ adapter->net_stats.rx_bytes += length;
+ frames_processed++;
+ } while (frames_processed < budget);
+
+ replenish_pools(adapter);
+
+ if (frames_processed < budget) {
+ enable_scrq_irq(adapter, adapter->rx_scrq[scrq_num]);
+ napi_complete(napi);
+ if (pending_scrq(adapter, adapter->rx_scrq[scrq_num]) &&
+ napi_reschedule(napi)) {
+ disable_scrq_irq(adapter, adapter->rx_scrq[scrq_num]);
+ goto restart_poll;
+ }
+ }
+
+ return frames_processed;
+}
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+static void ibmvnic_netpoll_controller(struct net_device *dev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(dev);
+ int i;
+
+ replenish_pools(netdev_priv(dev));
+ for (i = 0; i < adapter->req_rx_queues; i++)
+ ibmvnic_interrupt_rx(adapter->rx_scrq[i]->irq,
+ adapter->rx_scrq[i]);
+}
+#endif
+
+static const struct net_device_ops ibmvnic_netdev_ops = {
+ .ndo_open = ibmvnic_open,
+ .ndo_stop = ibmvnic_close,
+ .ndo_start_xmit = ibmvnic_xmit,
+ .ndo_set_rx_mode = ibmvnic_set_multi,
+ .ndo_set_mac_address = ibmvnic_set_mac,
+ .ndo_validate_addr = eth_validate_addr,
+ .ndo_do_ioctl = ibmvnic_ioctl,
+ .ndo_change_mtu = ibmvnic_change_mtu,
+ .ndo_tx_timeout = ibmvnic_tx_timeout,
+ .ndo_get_stats = ibmvnic_get_stats,
+#ifdef CONFIG_NET_POLL_CONTROLLER
+ .ndo_poll_controller = ibmvnic_netpoll_controller,
+#endif
+};
+
+/* ethtool functions */
+static int ibmvnic_get_settings(struct net_device *netdev,
+ struct ethtool_cmd *cmd)
+{
+ cmd->supported = (SUPPORTED_1000baseT_Full | SUPPORTED_Autoneg |
+ SUPPORTED_FIBRE);
+ cmd->advertising = (ADVERTISED_1000baseT_Full | ADVERTISED_Autoneg |
+ ADVERTISED_FIBRE);
+ ethtool_cmd_speed_set(cmd, SPEED_1000);
+ cmd->duplex = DUPLEX_FULL;
+ cmd->port = PORT_FIBRE;
+ cmd->phy_address = 0;
+ cmd->transceiver = XCVR_INTERNAL;
+ cmd->autoneg = AUTONEG_ENABLE;
+ cmd->maxtxpkt = 0;
+ cmd->maxrxpkt = 1;
+ return 0;
+}
+
+static void ibmvnic_get_drvinfo(struct net_device *dev,
+ struct ethtool_drvinfo *info)
+{
+ strncpy(info->driver, ibmvnic_driver_name, sizeof(info->driver) - 1);
+ strncpy(info->version, IBMVNIC_DRIVER_VERSION,
+ sizeof(info->version) - 1);
+}
+
+static u32 ibmvnic_get_msglevel(struct net_device *netdev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ return adapter->msg_enable;
+}
+
+static void ibmvnic_set_msglevel(struct net_device *netdev, u32 data)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ adapter->msg_enable = data;
+}
+
+static u32 ibmvnic_get_link(struct net_device *netdev)
+{
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+
+ /* Don't need to send a query because we request a logical link up at
+ * init and then we wait for link state indications
+ */
+ return adapter->logical_link_state;
+}
+
+static void ibmvnic_get_ringparam(struct net_device *netdev,
+ struct ethtool_ringparam *ring)
+{
+ ring->rx_max_pending = 0;
+ ring->tx_max_pending = 0;
+ ring->rx_mini_max_pending = 0;
+ ring->rx_jumbo_max_pending = 0;
+ ring->rx_pending = 0;
+ ring->tx_pending = 0;
+ ring->rx_mini_pending = 0;
+ ring->rx_jumbo_pending = 0;
+}
+
+static void ibmvnic_get_strings(struct net_device *dev, u32 stringset, u8 *data)
+{
+ int i;
+
+ if (stringset != ETH_SS_STATS)
+ return;
+
+ for (i = 0; i < ARRAY_SIZE(ibmvnic_stats); i++, data += ETH_GSTRING_LEN)
+ memcpy(data, ibmvnic_stats[i].name, ETH_GSTRING_LEN);
+}
+
+static int ibmvnic_get_sset_count(struct net_device *dev, int sset)
+{
+ switch (sset) {
+ case ETH_SS_STATS:
+ return ARRAY_SIZE(ibmvnic_stats);
+ default:
+ return -EOPNOTSUPP;
+ }
+}
+
+static void ibmvnic_get_ethtool_stats(struct net_device *dev,
+ struct ethtool_stats *stats, u64 *data)
+{
+ int i;
+ struct ibmvnic_adapter *adapter = netdev_priv(dev);
+ union ibmvnic_crq crq;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.request_statistics.first = IBMVNIC_CRQ_CMD;
+ crq.request_statistics.cmd = REQUEST_STATISTICS;
+ crq.request_statistics.ioba = cpu_to_be32(adapter->stats_token);
+ crq.request_statistics.len =
+ cpu_to_be32(sizeof(struct ibmvnic_statistics));
+ ibmvnic_send_crq(adapter, &crq);
+
+ /* Wait for data to be written */
+ init_completion(&adapter->stats_done);
+ wait_for_completion(&adapter->stats_done);
+
+ for (i = 0; i < ARRAY_SIZE(ibmvnic_stats); i++)
+ data[i] = IBMVNIC_GET_STAT(adapter, ibmvnic_stats[i].offset);
+
+}
+
+static const struct ethtool_ops ibmvnic_ethtool_ops = {
+ .get_settings = ibmvnic_get_settings,
+ .get_drvinfo = ibmvnic_get_drvinfo,
+ .get_msglevel = ibmvnic_get_msglevel,
+ .set_msglevel = ibmvnic_set_msglevel,
+ .get_link = ibmvnic_get_link,
+ .get_ringparam = ibmvnic_get_ringparam,
+ .get_strings = ibmvnic_get_strings,
+ .get_sset_count = ibmvnic_get_sset_count,
+ .get_ethtool_stats = ibmvnic_get_ethtool_stats,
+};
+
+/* vio bus functions */
+
+static void release_sub_crq_queue(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_sub_crq_queue *scrq)
+{
+ struct device *dev = &adapter->vdev->dev;
+ long rc;
+
+ netdev_dbg(adapter->netdev, "Releasing sub-CRQ\n");
+
+ /* Close the sub-crqs */
+ do {
+ rc = plpar_hcall_norets(H_FREE_SUB_CRQ,
+ adapter->vdev->unit_address,
+ scrq->crq_num);
+ } while (rc == H_BUSY || H_IS_LONG_BUSY(rc));
+
+ dma_unmap_single(dev, scrq->msg_token, PAGE_SIZE, DMA_BIDIRECTIONAL);
+ free_page((unsigned long)scrq->msgs);
+ kfree(scrq);
+
+}
+
+static struct ibmvnic_sub_crq_queue *init_sub_crq_queue(struct ibmvnic_adapter
+ *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ struct ibmvnic_sub_crq_queue *scrq;
+ int rc;
+
+ scrq = kmalloc(sizeof(*scrq), GFP_ATOMIC);
+ if (!scrq)
+ return NULL;
+
+ scrq->msgs = (union sub_crq *)get_zeroed_page(GFP_KERNEL);
+ if (!scrq->msgs) {
+ dev_warn(dev, "Couldn't allocate crq queue messages page\n");
+ goto zero_page_failed;
+ }
+
+ scrq->msg_token = dma_map_single(dev, scrq->msgs, PAGE_SIZE,
+ DMA_BIDIRECTIONAL);
+ if (dma_mapping_error(dev, scrq->msg_token)) {
+ dev_warn(dev, "Couldn't map crq queue messages page\n");
+ goto map_failed;
+ }
+
+ rc = h_reg_sub_crq(adapter->vdev->unit_address, scrq->msg_token,
+ PAGE_SIZE, &scrq->crq_num, &scrq->hw_irq);
+
+ if (rc == H_RESOURCE)
+ rc = ibmvnic_reset_crq(adapter);
+
+ if (rc == H_CLOSED) {
+ dev_warn(dev, "Partner adapter not ready, waiting.\n");
+ } else if (rc) {
+ dev_warn(dev, "Error %d registering sub-crq\n", rc);
+ goto reg_failed;
+ }
+
+ scrq->irq = irq_create_mapping(NULL, scrq->hw_irq);
+ if (scrq->irq == NO_IRQ) {
+ dev_err(dev, "Error mapping irq\n");
+ goto map_irq_failed;
+ }
+
+ scrq->adapter = adapter;
+ scrq->size = PAGE_SIZE / sizeof(*scrq->msgs);
+ scrq->cur = 0;
+ scrq->rx_skb_top = NULL;
+ spin_lock_init(&scrq->lock);
+
+ netdev_dbg(adapter->netdev,
+ "sub-crq initialized, num %lx, hw_irq=%lx, irq=%x\n",
+ scrq->crq_num, scrq->hw_irq, scrq->irq);
+
+ return scrq;
+
+map_irq_failed:
+ do {
+ rc = plpar_hcall_norets(H_FREE_SUB_CRQ,
+ adapter->vdev->unit_address,
+ scrq->crq_num);
+ } while (rc == H_BUSY || H_IS_LONG_BUSY(rc));
+reg_failed:
+ dma_unmap_single(dev, scrq->msg_token, PAGE_SIZE, DMA_BIDIRECTIONAL);
+map_failed:
+ free_page((unsigned long)scrq->msgs);
+zero_page_failed:
+ kfree(scrq);
+
+ return NULL;
+}
+
+static void release_sub_crqs(struct ibmvnic_adapter *adapter)
+{
+ int i;
+
+ if (adapter->tx_scrq) {
+ for (i = 0; i < adapter->req_tx_queues; i++)
+ if (adapter->tx_scrq[i]) {
+ free_irq(adapter->tx_scrq[i]->irq,
+ adapter->tx_scrq[i]);
+ release_sub_crq_queue(adapter,
+ adapter->tx_scrq[i]);
+ }
+ adapter->tx_scrq = NULL;
+ }
+
+ if (adapter->rx_scrq) {
+ for (i = 0; i < adapter->req_rx_queues; i++)
+ if (adapter->rx_scrq[i]) {
+ free_irq(adapter->rx_scrq[i]->irq,
+ adapter->rx_scrq[i]);
+ release_sub_crq_queue(adapter,
+ adapter->rx_scrq[i]);
+ }
+ adapter->rx_scrq = NULL;
+ }
+
+ adapter->requested_caps = 0;
+}
+
+static int disable_scrq_irq(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_sub_crq_queue *scrq)
+{
+ struct device *dev = &adapter->vdev->dev;
+ unsigned long rc;
+
+ rc = plpar_hcall_norets(H_VIOCTL, adapter->vdev->unit_address,
+ DISABLE_VIO_INTERRUPT, scrq->hw_irq, 0, 0);
+ if (rc)
+ dev_err(dev, "Couldn't disable scrq irq 0x%lx. rc=%ld\n",
+ scrq->hw_irq, rc);
+ return rc;
+}
+
+static int enable_scrq_irq(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_sub_crq_queue *scrq)
+{
+ struct device *dev = &adapter->vdev->dev;
+ unsigned long rc;
+
+ if (scrq->hw_irq > 0x100000000ULL) {
+ dev_err(dev, "bad hw_irq = %lx\n", scrq->hw_irq);
+ return 1;
+ }
+
+ rc = plpar_hcall_norets(H_VIOCTL, adapter->vdev->unit_address,
+ ENABLE_VIO_INTERRUPT, scrq->hw_irq, 0, 0);
+ if (rc)
+ dev_err(dev, "Couldn't enable scrq irq 0x%lx. rc=%ld\n",
+ scrq->hw_irq, rc);
+ return rc;
+}
+
+static int ibmvnic_complete_tx(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_sub_crq_queue *scrq)
+{
+ struct device *dev = &adapter->vdev->dev;
+ union sub_crq *next;
+ struct ibmvnic_tx_buff *txbuff;
+ int index;
+ int i, j;
+
+restart_loop:
+ while (pending_scrq(adapter, scrq)) {
+ unsigned int pool = scrq->pool_index;
+
+ next = ibmvnic_next_scrq(adapter, scrq);
+ for (i = 0; i < next->tx_comp.num_comps; i++) {
+ if (next->tx_comp.rcs[i]) {
+ dev_err(dev, "tx error %x\n",
+ next->tx_comp.rcs[i]);
+ continue;
+ }
+ index = be32_to_cpu(next->tx_comp.correlators[i]);
+ txbuff = &adapter->tx_pool[pool].tx_buff[index];
+
+ for (j = 0; j < IBMVNIC_MAX_FRAGS_PER_CRQ; j++) {
+ if (!txbuff->data_dma[j])
+ continue;
+
+ if (txbuff->used_bounce)
+ dma_unmap_single(dev,
+ txbuff->data_dma[j],
+ txbuff->data_len[j],
+ DMA_TO_DEVICE);
+ txbuff->data_dma[j] = 0;
+ txbuff->used_bounce = 0;
+ }
+
+ if (txbuff->last_frag)
+ dev_kfree_skb(txbuff->skb);
+
+ adapter->tx_pool[pool].free_map[adapter->tx_pool[pool].
+ producer_index] = index;
+ adapter->tx_pool[pool].producer_index =
+ (adapter->tx_pool[pool].producer_index + 1) %
+ adapter->max_tx_entries_per_subcrq;
+ }
+ }
+
+ enable_scrq_irq(adapter, scrq);
+
+ if (pending_scrq(adapter, scrq)) {
+ disable_scrq_irq(adapter, scrq);
+ goto restart_loop;
+ }
+
+ return 0;
+}
+
+static int ibmvnic_tx_work(void *data)
+{
+ int rc;
+ struct ibmvnic_sub_crq_queue *scrq = data;
+ struct ibmvnic_adapter *adapter = scrq->adapter;
+
+ set_user_nice(current, -20);
+
+ while (1) {
+ rc = wait_event_interruptible(adapter->
+ tx_pool[scrq->pool_index].
+ ibmvnic_tx_comp_q,
+ pending_scrq(adapter, scrq));
+ BUG_ON(rc);
+
+ if (kthread_should_stop())
+ break;
+
+ ibmvnic_complete_tx(adapter, scrq);
+ }
+
+ return 0;
+}
+
+static irqreturn_t ibmvnic_interrupt_tx(int irq, void *instance)
+{
+ struct ibmvnic_sub_crq_queue *scrq = instance;
+ struct ibmvnic_adapter *adapter = scrq->adapter;
+
+ disable_scrq_irq(adapter, scrq);
+ wake_up(&adapter->tx_pool[scrq->pool_index].ibmvnic_tx_comp_q);
+
+ return IRQ_HANDLED;
+}
+
+static irqreturn_t ibmvnic_interrupt_rx(int irq, void *instance)
+{
+ struct ibmvnic_sub_crq_queue *scrq = instance;
+ struct ibmvnic_adapter *adapter = scrq->adapter;
+
+ if (napi_schedule_prep(&adapter->napi[scrq->scrq_num])) {
+ disable_scrq_irq(adapter, scrq);
+ __napi_schedule(&adapter->napi[scrq->scrq_num]);
+ }
+
+ return IRQ_HANDLED;
+}
+
+static void init_sub_crqs(struct ibmvnic_adapter *adapter, int retry)
+{
+ struct device *dev = &adapter->vdev->dev;
+ int i, j, more = 0, total_queues, registered_queues = 0;
+ struct ibmvnic_sub_crq_queue **allqueues;
+ union ibmvnic_crq crq;
+ int rc;
+
+ if (!retry) {
+ /* Sub-CRQ entries are 32 byte long */
+ int entries_page = PAGE_SIZE / (sizeof(u64) * 4);
+
+ if (adapter->min_tx_entries_per_subcrq > entries_page ||
+ adapter->min_rx_add_entries_per_subcrq > entries_page) {
+ dev_err(dev, "Fatal, invalid entries per sub-crq\n");
+ goto allqueues_failed;
+ }
+
+ /* Get the minimum between the queried max and the entries
+ * that fit in our PAGE_SIZE
+ */
+ adapter->req_tx_entries_per_subcrq =
+ adapter->max_tx_entries_per_subcrq > entries_page ?
+ entries_page : adapter->max_tx_entries_per_subcrq;
+ adapter->req_rx_add_entries_per_subcrq =
+ adapter->max_rx_add_entries_per_subcrq > entries_page ?
+ entries_page : adapter-> max_rx_add_entries_per_subcrq;
+
+ /* Choosing the maximum number of queues supported by firmware*/
+ adapter->req_tx_queues = adapter->min_tx_queues;
+ adapter->req_rx_queues = adapter->min_rx_queues;
+ adapter->req_rx_add_queues = adapter->min_rx_add_queues;
+
+ adapter->req_mtu = adapter->max_mtu;
+ if (ibmvnic_req_mtu) {
+ if (ibmvnic_req_mtu > adapter->max_mtu ||
+ ibmvnic_req_mtu < adapter->min_mtu)
+ dev_warn(dev, "Module param req_mtu out of "
+ "range, ignoring it\n");
+ else
+ adapter->req_mtu = ibmvnic_req_mtu;
+ }
+ }
+
+ total_queues = adapter->req_tx_queues + adapter->req_rx_queues;
+
+ allqueues = kcalloc(total_queues, sizeof(*allqueues), GFP_ATOMIC);
+ if (!allqueues)
+ goto allqueues_failed;
+
+ for (i = 0; i < total_queues; i++) {
+ allqueues[i] = init_sub_crq_queue(adapter);
+ if (!allqueues[i]) {
+ dev_warn(dev, "Couldn't allocate all sub-crqs\n");
+ break;
+ }
+ registered_queues++;
+ }
+
+ /* Make sure we were able to register the minimum number of queues */
+ if (registered_queues <
+ adapter->min_tx_queues + adapter->min_rx_queues) {
+ dev_err(dev, "Fatal: Couldn't init min number of sub-crqs\n");
+ goto tx_failed;
+ }
+
+ /* Distribute the failed allocated queues*/
+ for (i = 0; i < total_queues - registered_queues + more ; i++) {
+ netdev_dbg(adapter->netdev, "Reducing number of queues\n");
+ switch (i % 3) {
+ case 0:
+ if (adapter->req_rx_queues > adapter->min_rx_queues)
+ adapter->req_rx_queues--;
+ else
+ more++;
+ break;
+ case 1:
+ if (adapter->req_tx_queues > adapter->min_tx_queues)
+ adapter->req_tx_queues--;
+ else
+ more++;
+ break;
+ }
+ }
+
+ adapter->tx_scrq = kcalloc(adapter->req_tx_queues,
+ sizeof(*adapter->tx_scrq), GFP_ATOMIC);
+ if (!adapter->tx_scrq)
+ goto tx_failed;
+
+ for (i = 0; i < adapter->req_tx_queues; i++) {
+ adapter->tx_scrq[i] = allqueues[i];
+ adapter->tx_scrq[i]->pool_index = i;
+ rc = request_irq(adapter->tx_scrq[i]->irq, ibmvnic_interrupt_tx,
+ 0, "ibmvnic_tx", adapter->tx_scrq[i]);
+ if (rc) {
+ dev_err(dev, "Couldn't register tx irq 0x%x. rc=%d\n",
+ adapter->tx_scrq[i]->irq, rc);
+ goto req_tx_irq_failed;
+ }
+ }
+
+ adapter->rx_scrq = kcalloc(adapter->req_rx_queues,
+ sizeof(*adapter->rx_scrq), GFP_ATOMIC);
+ if (!adapter->rx_scrq)
+ goto rx_failed;
+
+ for (i = 0; i < adapter->req_rx_queues; i++) {
+ adapter->rx_scrq[i] = allqueues[i + adapter->req_tx_queues];
+ adapter->rx_scrq[i]->scrq_num = i;
+ rc = request_irq(adapter->rx_scrq[i]->irq, ibmvnic_interrupt_rx,
+ 0, "ibmvnic_rx", adapter->rx_scrq[i]);
+ if (rc) {
+ dev_err(dev, "Couldn't register rx irq 0x%x. rc=%d\n",
+ adapter->rx_scrq[i]->irq, rc);
+ goto req_rx_irq_failed;
+ }
+ }
+
+ memset(&crq, 0, sizeof(crq));
+ crq.request_capability.first = IBMVNIC_CRQ_CMD;
+ crq.request_capability.cmd = REQUEST_CAPABILITY;
+
+ crq.request_capability.capability = cpu_to_be16(REQ_TX_QUEUES);
+ crq.request_capability.number = cpu_to_be32(adapter->req_tx_queues);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.request_capability.capability = cpu_to_be16(REQ_RX_QUEUES);
+ crq.request_capability.number = cpu_to_be32(adapter->req_rx_queues);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.request_capability.capability = cpu_to_be16(REQ_RX_ADD_QUEUES);
+ crq.request_capability.number = cpu_to_be32(adapter->req_rx_add_queues);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.request_capability.capability =
+ cpu_to_be16(REQ_TX_ENTRIES_PER_SUBCRQ);
+ crq.request_capability.number =
+ cpu_to_be32(adapter->req_tx_entries_per_subcrq);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.request_capability.capability =
+ cpu_to_be16(REQ_RX_ADD_ENTRIES_PER_SUBCRQ);
+ crq.request_capability.number =
+ cpu_to_be32(adapter->req_rx_add_entries_per_subcrq);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.request_capability.capability = cpu_to_be16(REQ_MTU);
+ crq.request_capability.number = cpu_to_be32(adapter->req_mtu);
+ ibmvnic_send_crq(adapter, &crq);
+
+ kfree(allqueues);
+
+ return;
+
+req_rx_irq_failed:
+ for (j = 0; j < i; j++)
+ free_irq(adapter->rx_scrq[j]->irq, adapter->rx_scrq[j]);
+ i = adapter->req_tx_queues;
+req_tx_irq_failed:
+ for (j = 0; j < i; j++)
+ free_irq(adapter->tx_scrq[j]->irq, adapter->tx_scrq[j]);
+ kfree(adapter->rx_scrq);
+ adapter->rx_scrq = NULL;
+rx_failed:
+ kfree(adapter->tx_scrq);
+ adapter->tx_scrq = NULL;
+tx_failed:
+ for (i = 0; i < registered_queues; i++)
+ release_sub_crq_queue(adapter, allqueues[i]);
+ kfree(allqueues);
+allqueues_failed:
+ ibmvnic_remove(adapter->vdev);
+ return;
+}
+
+static int pending_scrq(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_sub_crq_queue *scrq)
+{
+ union sub_crq *entry = &scrq->msgs[scrq->cur];
+
+ if (entry->generic.first & IBMVNIC_CRQ_CMD_RSP || adapter->closing)
+ return 1;
+ else
+ return 0;
+}
+
+static union sub_crq *ibmvnic_next_scrq(struct ibmvnic_adapter *adapter,
+ struct ibmvnic_sub_crq_queue *scrq)
+{
+ union sub_crq *entry;
+ unsigned long flags;
+
+ spin_lock_irqsave(&scrq->lock, flags);
+ entry = &scrq->msgs[scrq->cur];
+ if (entry->generic.first & IBMVNIC_CRQ_CMD_RSP) {
+ if (++scrq->cur == scrq->size)
+ scrq->cur = 0;
+ } else {
+ entry = NULL;
+ }
+ spin_unlock_irqrestore(&scrq->lock, flags);
+
+ return entry;
+}
+
+static union ibmvnic_crq *ibmvnic_next_crq(struct ibmvnic_adapter *adapter)
+{
+ struct ibmvnic_crq_queue *queue = &adapter->crq;
+ union ibmvnic_crq *crq;
+ unsigned long flags;
+
+ spin_lock_irqsave(&queue->lock, flags);
+ crq = &queue->msgs[queue->cur];
+ if (crq->generic.first & IBMVNIC_CRQ_CMD_RSP) {
+ if (++queue->cur == queue->size)
+ queue->cur = 0;
+ } else {
+ crq = NULL;
+ }
+
+ spin_unlock_irqrestore(&queue->lock, flags);
+
+ return crq;
+}
+
+static int send_subcrq(struct ibmvnic_adapter *adapter, u64 remote_handle,
+ union sub_crq *sub_crq)
+{
+ unsigned int ua = adapter->vdev->unit_address;
+ struct device *dev = &adapter->vdev->dev;
+ int rc;
+
+ u64 *u64_crq = (u64 *) sub_crq;
+
+ netdev_dbg(adapter->netdev,
+ "Sending sCRQ %016lx: %016lx %016lx %016lx %016lx\n",
+ (long unsigned int)remote_handle,
+ (long unsigned int)u64_crq[0], (long unsigned int)u64_crq[1],
+ (long unsigned int)u64_crq[2],
+ (long unsigned int)u64_crq[3]);
+
+ /* Make sure the hypervisor sees the complete request */
+ mb();
+
+ rc = plpar_hcall_norets(H_SEND_SUB_CRQ, ua, remote_handle, u64_crq[0],
+ u64_crq[1], u64_crq[2], u64_crq[3]);
+ if (rc) {
+ if (rc == H_CLOSED)
+ dev_warn(dev, "CRQ Queue closed\n");
+ dev_err(dev, "Send error (rc=%d)\n", rc);
+ }
+
+ return rc;
+}
+
+static int ibmvnic_send_crq(struct ibmvnic_adapter *adapter,
+ union ibmvnic_crq *crq)
+{
+ unsigned int ua = adapter->vdev->unit_address;
+ struct device *dev = &adapter->vdev->dev;
+ int rc;
+
+ u64 *u64_crq = (u64 *)crq;
+ netdev_dbg(adapter->netdev, "Sending CRQ: %016lx %016lx\n",
+ (long unsigned int)u64_crq[0],
+ (long unsigned int)u64_crq[1]);
+
+ /* Make sure the hypervisor sees the complete request */
+ mb();
+
+ rc = plpar_hcall_norets(H_SEND_CRQ, ua, u64_crq[0], u64_crq[1]);
+ if (rc) {
+ if (rc == H_CLOSED)
+ dev_warn(dev, "CRQ Queue closed\n");
+ dev_warn(dev, "Send error (rc=%d)\n", rc);
+ }
+
+ return rc;
+}
+
+static int ibmvnic_send_crq_init(struct ibmvnic_adapter *adapter)
+{
+ union ibmvnic_crq crq;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.generic.first = IBMVNIC_CRQ_INIT_CMD;
+ crq.generic.cmd = IBMVNIC_CRQ_INIT;
+ netdev_dbg(adapter->netdev, "Sending CRQ init\n");
+
+ return ibmvnic_send_crq(adapter, &crq);
+}
+
+static int ibmvnic_send_crq_init_complete(struct ibmvnic_adapter *adapter)
+{
+ union ibmvnic_crq crq;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.generic.first = IBMVNIC_CRQ_INIT_CMD;
+ crq.generic.cmd = IBMVNIC_CRQ_INIT_COMPLETE;
+ netdev_dbg(adapter->netdev, "Sending CRQ init complete\n");
+
+ return ibmvnic_send_crq(adapter, &crq);
+}
+
+static int send_version_xchg(struct ibmvnic_adapter *adapter)
+{
+ union ibmvnic_crq crq;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.version_exchange.first = IBMVNIC_CRQ_CMD;
+ crq.version_exchange.cmd = VERSION_EXCHANGE;
+ crq.version_exchange.version = cpu_to_be16(ibmvnic_version);
+
+ return ibmvnic_send_crq(adapter, &crq);
+}
+
+static void send_login(struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ union ibmvnic_crq crq;
+ struct ibmvnic_login_buffer *login_buffer;
+ size_t buffer_size;
+ dma_addr_t buffer_token;
+ struct ibmvnic_login_rsp_buffer *login_rsp_buffer;
+ size_t rsp_buffer_size;
+ dma_addr_t rsp_buffer_token;
+ long unsigned int *tx_list_p;
+ long unsigned int *rx_list_p;
+ struct ibmvnic_inflight_cmd *inflight_cmd;
+ unsigned long flags;
+ int i;
+
+ buffer_size =
+ sizeof(struct ibmvnic_login_buffer) +
+ sizeof(u64) * (adapter->req_tx_queues + adapter->req_rx_queues);
+
+ login_buffer = kmalloc(buffer_size, GFP_ATOMIC);
+ if (!login_buffer)
+ goto buf_alloc_failed;
+
+ buffer_token = dma_map_single(dev, login_buffer, buffer_size,
+ DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, buffer_token)) {
+ dev_err(dev, "Couldn't map login buffer\n");
+ goto buf_map_failed;
+ }
+
+ rsp_buffer_size =
+ sizeof(struct ibmvnic_login_rsp_buffer) +
+ sizeof(u64) * (adapter->req_tx_queues +
+ adapter->req_rx_queues *
+ adapter->req_rx_add_queues + adapter->
+ req_rx_add_queues) +
+ sizeof(u8) * (IBMVNIC_TX_DESC_VERSIONS);
+
+ login_rsp_buffer = kmalloc(rsp_buffer_size, GFP_ATOMIC);
+ if (!login_rsp_buffer)
+ goto buf_rsp_alloc_failed;
+
+ rsp_buffer_token = dma_map_single(dev, login_rsp_buffer,
+ rsp_buffer_size, DMA_FROM_DEVICE);
+ if (dma_mapping_error(dev, rsp_buffer_token)) {
+ dev_err(dev, "Couldn't map login rsp buffer\n");
+ goto buf_rsp_map_failed;
+ }
+
+ inflight_cmd = kmalloc(sizeof(inflight_cmd), GFP_ATOMIC);
+ if (!inflight_cmd) {
+ dev_err(dev, "Couldn't allocate inflight_cmd\n");
+ goto inflight_alloc_failed;
+ }
+
+ adapter->login_buf = login_buffer;
+ adapter->login_buf_token = buffer_token;
+ adapter->login_buf_sz = buffer_size;
+ adapter->login_rsp_buf = login_rsp_buffer;
+ adapter->login_rsp_buf_token = rsp_buffer_token;
+ adapter->login_rsp_buf_sz = rsp_buffer_size;
+
+ login_buffer->len = cpu_to_be32(buffer_size);
+ login_buffer->version = cpu_to_be32(INITIAL_VERSION_LB);
+ login_buffer->num_txcomp_subcrqs = cpu_to_be32(adapter->req_tx_queues);
+ login_buffer->off_txcomp_subcrqs =
+ cpu_to_be32(sizeof(struct ibmvnic_login_buffer));
+ login_buffer->num_rxcomp_subcrqs = cpu_to_be32(adapter->req_rx_queues);
+ login_buffer->off_rxcomp_subcrqs =
+ cpu_to_be32(sizeof(struct ibmvnic_login_buffer) +
+ sizeof(u64) * adapter->req_tx_queues);
+ login_buffer->login_rsp_ioba = cpu_to_be32(rsp_buffer_token);
+ login_buffer->login_rsp_len = cpu_to_be32(rsp_buffer_size);
+
+ tx_list_p = (unsigned long *)((char *)login_buffer +
+ sizeof(struct ibmvnic_login_buffer));
+ rx_list_p = (unsigned long *)((char *)login_buffer +
+ sizeof(struct ibmvnic_login_buffer) +
+ sizeof(u64) * adapter->req_tx_queues);
+
+ for (i = 0; i < adapter->req_tx_queues; i++) {
+ if (adapter->tx_scrq[i])
+ tx_list_p[i] = adapter->tx_scrq[i]->crq_num;
+ }
+
+ for (i = 0; i < adapter->req_rx_queues; i++) {
+ if (adapter->rx_scrq[i])
+ rx_list_p[i] = adapter->rx_scrq[i]->crq_num;
+ }
+
+ netdev_dbg(adapter->netdev, "Login Buffer:\n");
+ for (i = 0; i < (adapter->login_buf_sz - 1) / 8 + 1; i++) {
+ netdev_dbg(adapter->netdev, "%016lx\n",
+ ((long unsigned int *)(adapter->login_buf))[i]);
+ }
+
+ memset(&crq, 0, sizeof(crq));
+ crq.login.first = IBMVNIC_CRQ_CMD;
+ crq.login.cmd = LOGIN;
+ crq.login.ioba = cpu_to_be32(buffer_token);
+ crq.login.len = cpu_to_be32(buffer_size);
+
+ memcpy(&inflight_cmd->crq, &crq, sizeof(crq));
+
+ spin_lock_irqsave(&adapter->inflight_lock, flags);
+ list_add_tail(&inflight_cmd->list, &adapter->inflight);
+ spin_unlock_irqrestore(&adapter->inflight_lock, flags);
+
+ ibmvnic_send_crq(adapter, &crq);
+
+ return;
+
+inflight_alloc_failed:
+ dma_unmap_single(dev, rsp_buffer_token, rsp_buffer_size,
+ DMA_FROM_DEVICE);
+buf_rsp_map_failed:
+ kfree(login_rsp_buffer);
+buf_rsp_alloc_failed:
+ dma_unmap_single(dev, buffer_token, buffer_size, DMA_TO_DEVICE);
+buf_map_failed:
+ kfree(login_buffer);
+buf_alloc_failed:
+ return;
+}
+
+static void send_cap_queries(struct ibmvnic_adapter *adapter)
+{
+ union ibmvnic_crq crq;
+
+ atomic_set(&adapter->running_cap_queries, 0);
+ memset(&crq, 0, sizeof(crq));
+ crq.query_capability.first = IBMVNIC_CRQ_CMD;
+ crq.query_capability.cmd = QUERY_CAPABILITY;
+
+ crq.query_capability.capability = cpu_to_be16(MIN_TX_QUEUES);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MIN_RX_QUEUES);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MIN_RX_ADD_QUEUES);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MAX_TX_QUEUES);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MAX_RX_QUEUES);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MAX_RX_ADD_QUEUES);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability =
+ cpu_to_be16(MIN_TX_ENTRIES_PER_SUBCRQ);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability =
+ cpu_to_be16(MIN_RX_ADD_ENTRIES_PER_SUBCRQ);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability =
+ cpu_to_be16(MAX_TX_ENTRIES_PER_SUBCRQ);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability =
+ cpu_to_be16(MAX_RX_ADD_ENTRIES_PER_SUBCRQ);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(TCP_IP_OFFLOAD);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(PROMISC_SUPPORTED);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MIN_MTU);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MAX_MTU);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MAX_MULTICAST_FILTERS);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(VLAN_HEADER_INSERTION);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(MAX_TX_SG_ENTRIES);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ crq.query_capability.capability = cpu_to_be16(RX_SG_SUPPORTED);
+ atomic_inc(&adapter->running_cap_queries);
+ ibmvnic_send_crq(adapter, &crq);
+
+ return;
+}
+
+static void handle_query_ip_offload_rsp(struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ struct ibmvnic_query_ip_offload_buffer *buf = &adapter->ip_offload_buf;
+ union ibmvnic_crq crq;
+ int i;
+
+ dma_unmap_single(dev, adapter->ip_offload_tok,
+ sizeof(adapter->ip_offload_buf), DMA_FROM_DEVICE);
+
+ netdev_dbg(adapter->netdev, "Query IP Offload Buffer:\n");
+ for (i = 0; i < (sizeof(adapter->ip_offload_buf) - 1) / 8 + 1; i++)
+ netdev_dbg(adapter->netdev, "%016lx\n",
+ ((long unsigned int *)(buf))[i]);
+
+ netdev_dbg(adapter->netdev, "ipv4_chksum = %d\n", buf->ipv4_chksum);
+ netdev_dbg(adapter->netdev, "ipv6_chksum = %d\n", buf->ipv6_chksum);
+ netdev_dbg(adapter->netdev, "tcp_ipv4_chksum = %d\n",
+ buf->tcp_ipv4_chksum);
+ netdev_dbg(adapter->netdev, "tcp_ipv6_chksum = %d\n",
+ buf->tcp_ipv6_chksum);
+ netdev_dbg(adapter->netdev, "udp_ipv4_chksum = %d\n",
+ buf->udp_ipv4_chksum);
+ netdev_dbg(adapter->netdev, "udp_ipv6_chksum = %d\n",
+ buf->udp_ipv6_chksum);
+ netdev_dbg(adapter->netdev, "large_tx_ipv4 = %d\n", buf->large_tx_ipv4);
+ netdev_dbg(adapter->netdev, "large_tx_ipv6 = %d\n", buf->large_tx_ipv6);
+ netdev_dbg(adapter->netdev, "large_rx_ipv4 = %d\n", buf->large_rx_ipv4);
+ netdev_dbg(adapter->netdev, "large_rx_ipv6 = %d\n", buf->large_rx_ipv6);
+ netdev_dbg(adapter->netdev, "max_ipv4_hdr_sz = %d\n",
+ buf->max_ipv4_header_size);
+ netdev_dbg(adapter->netdev, "max_ipv6_hdr_sz = %d\n",
+ buf->max_ipv6_header_size);
+ netdev_dbg(adapter->netdev, "max_tcp_hdr_size = %d\n",
+ buf->max_tcp_header_size);
+ netdev_dbg(adapter->netdev, "max_udp_hdr_size = %d\n",
+ buf->max_udp_header_size);
+ netdev_dbg(adapter->netdev, "max_large_tx_size = %d\n",
+ buf->max_large_tx_size);
+ netdev_dbg(adapter->netdev, "max_large_rx_size = %d\n",
+ buf->max_large_rx_size);
+ netdev_dbg(adapter->netdev, "ipv6_ext_hdr = %d\n",
+ buf->ipv6_extension_header);
+ netdev_dbg(adapter->netdev, "tcp_pseudosum_req = %d\n",
+ buf->tcp_pseudosum_req);
+ netdev_dbg(adapter->netdev, "num_ipv6_ext_hd = %d\n",
+ buf->num_ipv6_ext_headers);
+ netdev_dbg(adapter->netdev, "off_ipv6_ext_hd = %d\n",
+ buf->off_ipv6_ext_headers);
+
+ adapter->ip_offload_ctrl_tok =
+ dma_map_single(dev, &adapter->ip_offload_ctrl,
+ sizeof(adapter->ip_offload_ctrl), DMA_TO_DEVICE);
+
+ if (dma_mapping_error(dev, adapter->ip_offload_ctrl_tok)) {
+ dev_err(dev, "Couldn't map ip offload control buffer\n");
+ return;
+ }
+
+ adapter->ip_offload_ctrl.version = cpu_to_be32(INITIAL_VERSION_IOB);
+ adapter->ip_offload_ctrl.tcp_ipv4_chksum = buf->tcp_ipv4_chksum;
+ adapter->ip_offload_ctrl.udp_ipv4_chksum = buf->udp_ipv4_chksum;
+ adapter->ip_offload_ctrl.tcp_ipv6_chksum = buf->tcp_ipv6_chksum;
+ adapter->ip_offload_ctrl.udp_ipv6_chksum = buf->udp_ipv6_chksum;
+ adapter->ip_offload_ctrl.large_tx_ipv4 = buf->large_tx_ipv4;
+ adapter->ip_offload_ctrl.large_tx_ipv6 = buf->large_tx_ipv6;
+ adapter->ip_offload_ctrl.large_rx_ipv4 = buf->large_rx_ipv4;
+ adapter->ip_offload_ctrl.large_rx_ipv6 = buf->large_rx_ipv6;
+
+ adapter->netdev->features = NETIF_F_SG | NETIF_F_GSO;
+
+ if (buf->tcp_ipv4_chksum || buf->udp_ipv4_chksum)
+ adapter->netdev->features |= NETIF_F_SG | NETIF_F_IP_CSUM;
+
+ if (buf->tcp_ipv6_chksum || buf->udp_ipv6_chksum)
+ adapter->netdev->features |= NETIF_F_SG | NETIF_F_IPV6_CSUM;
+
+ if (buf->large_tx_ipv4)
+ adapter->netdev->features |= NETIF_F_TSO;
+
+ if (buf->large_tx_ipv6)
+ adapter->netdev->features |= NETIF_F_TSO6;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.control_ip_offload.first = IBMVNIC_CRQ_CMD;
+ crq.control_ip_offload.cmd = CONTROL_IP_OFFLOAD;
+ crq.control_ip_offload.len =
+ cpu_to_be32(sizeof(adapter->ip_offload_ctrl));
+ crq.control_ip_offload.ioba = cpu_to_be32(adapter->ip_offload_ctrl_tok);
+ ibmvnic_send_crq(adapter, &crq);
+}
+
+static void handle_error_info_rsp(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ struct ibmvnic_error_buff *error_buff;
+ int found = 0;
+ unsigned long flags;
+ int i;
+
+ spin_lock_irqsave(&adapter->error_list_lock, flags);
+ list_for_each_entry(error_buff, &adapter->errors, list)
+ if (error_buff->error_id == crq->request_error_rsp.error_id) {
+ found = 1;
+ list_del(&error_buff->list);
+ break;
+ }
+ spin_unlock_irqrestore(&adapter->error_list_lock, flags);
+
+ if (!found) {
+ dev_err(dev, "Couldn't find error id %x\n",
+ crq->request_error_rsp.error_id);
+ return;
+ }
+
+ dev_err(dev, "Detailed info for error id %x:",
+ crq->request_error_rsp.error_id);
+
+ for (i = 0; i < error_buff->len; i++) {
+ pr_cont("%02x", (int)error_buff->buff[i]);
+ if (i % 8 == 7)
+ pr_cont(" ");
+ }
+ pr_cont("\n");
+
+ dma_unmap_single(dev, error_buff->dma, error_buff->len,
+ DMA_FROM_DEVICE);
+ kfree(error_buff->buff);
+ kfree(error_buff);
+}
+
+static void handle_dump_size_rsp(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ union ibmvnic_crq newcrq;
+ struct device *dev = &adapter->vdev->dev;
+ int len = be32_to_cpu(crq->request_dump_size_rsp.len);
+ struct ibmvnic_inflight_cmd *inflight_cmd;
+ unsigned long flags;
+
+ /* allocate and map buffer */
+ adapter->dump_data = kmalloc(len, GFP_KERNEL);
+ if (!adapter->dump_data) {
+ complete(&adapter->fw_done);
+ return;
+ }
+
+ adapter->dump_data_token = dma_map_single(dev, adapter->dump_data, len,
+ DMA_FROM_DEVICE);
+
+ if (dma_mapping_error(dev, adapter->dump_data_token)) {
+ if (!firmware_has_feature(FW_FEATURE_CMO))
+ dev_err(dev, "Couldn't map dump data\n");
+ kfree(adapter->dump_data);
+ complete(&adapter->fw_done);
+ return;
+ }
+
+ inflight_cmd = kmalloc(sizeof(*inflight_cmd), GFP_ATOMIC);
+ if (!inflight_cmd) {
+ dma_unmap_single(dev, adapter->dump_data_token, len,
+ DMA_FROM_DEVICE);
+ kfree(adapter->dump_data);
+ complete(&adapter->fw_done);
+ return;
+ }
+
+ memset(&newcrq, 0, sizeof(newcrq));
+ newcrq.request_dump.first = IBMVNIC_CRQ_CMD;
+ newcrq.request_dump.cmd = REQUEST_DUMP;
+ newcrq.request_dump.ioba = cpu_to_be32(adapter->dump_data_token);
+ newcrq.request_dump.len = cpu_to_be32(adapter->dump_data_size);
+
+ memcpy(&inflight_cmd->crq, &newcrq, sizeof(newcrq));
+
+ spin_lock_irqsave(&adapter->inflight_lock, flags);
+ list_add_tail(&inflight_cmd->list, &adapter->inflight);
+ spin_unlock_irqrestore(&adapter->inflight_lock, flags);
+
+ ibmvnic_send_crq(adapter, &newcrq);
+}
+
+static void handle_error_indication(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ union ibmvnic_crq new_crq;
+ struct ibmvnic_error_buff *error_buff;
+ unsigned long flags;
+ int detail_len = be32_to_cpu(crq->error_indication.detail_error_sz);
+ struct ibmvnic_inflight_cmd *inflight_cmd;
+
+ dev_err(dev, "Firmware reports %serror id %x, cause %d\n",
+ crq->error_indication.
+ flags & IBMVNIC_FATAL_ERROR ? "FATAL " : "",
+ crq->error_indication.error_id,
+ crq->error_indication.error_cause);
+
+ error_buff = kmalloc(sizeof(*error_buff), GFP_ATOMIC);
+ if (!error_buff)
+ return;
+
+ error_buff->buff = kmalloc(detail_len, GFP_ATOMIC);
+ if (!error_buff->buff) {
+ kfree(error_buff);
+ return;
+ }
+
+ error_buff->dma = dma_map_single(dev, error_buff->buff, detail_len,
+ DMA_FROM_DEVICE);
+ if (dma_mapping_error(dev, error_buff->dma)) {
+ if (!firmware_has_feature(FW_FEATURE_CMO))
+ dev_err(dev, "Couldn't map error buffer\n");
+ kfree(error_buff->buff);
+ kfree(error_buff);
+ return;
+ }
+
+ inflight_cmd = kmalloc(sizeof(*inflight_cmd), GFP_ATOMIC);
+ if (!inflight_cmd) {
+ dma_unmap_single(dev, error_buff->dma, detail_len,
+ DMA_FROM_DEVICE);
+ kfree(error_buff->buff);
+ kfree(error_buff);
+ return;
+ }
+
+ error_buff->len = detail_len;
+ error_buff->error_id = crq->error_indication.error_id;
+
+ spin_lock_irqsave(&adapter->error_list_lock, flags);
+ list_add_tail(&error_buff->list, &adapter->errors);
+ spin_unlock_irqrestore(&adapter->error_list_lock, flags);
+
+ memset(&new_crq, 0, sizeof(new_crq));
+ new_crq.request_error_info.first = IBMVNIC_CRQ_CMD;
+ new_crq.request_error_info.cmd = REQUEST_ERROR_INFO;
+ new_crq.request_error_info.ioba = cpu_to_be32(error_buff->dma);
+ new_crq.request_error_info.len = cpu_to_be32(detail_len);
+ new_crq.request_error_info.error_id = crq->error_indication.error_id;
+
+ memcpy(&inflight_cmd->crq, &crq, sizeof(crq));
+
+ spin_lock_irqsave(&adapter->inflight_lock, flags);
+ list_add_tail(&inflight_cmd->list, &adapter->inflight);
+ spin_unlock_irqrestore(&adapter->inflight_lock, flags);
+
+ ibmvnic_send_crq(adapter, &new_crq);
+
+ return;
+}
+
+static void handle_change_mac_rsp(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ struct net_device *netdev = adapter->netdev;
+ long rc;
+
+ rc = crq->change_mac_addr_rsp.rc.code;
+ if (rc) {
+ dev_err(dev, "Error %ld in CHANGE_MAC_ADDR_RSP\n", rc);
+ return;
+ }
+ memcpy(netdev->dev_addr, &crq->change_mac_addr_rsp.mac_addr[0],
+ ETH_ALEN);
+}
+
+static void handle_request_cap_rsp(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ u64 *req_value;
+ char *name;
+
+ switch (be16_to_cpu(crq->request_capability_rsp.capability)) {
+ case REQ_TX_QUEUES:
+ req_value = &adapter->req_tx_queues;
+ name = "tx";
+ break;
+ case REQ_RX_QUEUES:
+ req_value = &adapter->req_rx_queues;
+ name = "rx";
+ break;
+ case REQ_RX_ADD_QUEUES:
+ req_value = &adapter->req_rx_add_queues;
+ name = "rx_add";
+ break;
+ case REQ_TX_ENTRIES_PER_SUBCRQ:
+ req_value = &adapter->req_tx_entries_per_subcrq;
+ name = "tx_entries_per_subcrq";
+ break;
+ case REQ_RX_ADD_ENTRIES_PER_SUBCRQ:
+ req_value = &adapter->req_rx_add_entries_per_subcrq;
+ name = "rx_add_entries_per_subcrq";
+ break;
+ case REQ_MTU:
+ req_value = &adapter->req_mtu;
+ name = "mtu";
+ break;
+ case PROMISC_REQUESTED:
+ req_value = &adapter->promisc;
+ name = "promisc";
+ return;
+ default:
+ dev_err(dev, "Got invalid cap request rsp %d\n",
+ crq->request_capability.capability);
+ return;
+ }
+
+ switch (crq->request_capability_rsp.rc.code) {
+ case SUCCESS:
+ break;
+ case PARTIALSUCCESS:
+ dev_info(dev, "req=%lld, rsp=%ld in %s queue, retrying.\n",
+ *req_value,
+ (long int)be32_to_cpu(crq->request_capability_rsp.
+ number), name);
+ release_sub_crqs(adapter);
+ *req_value = be32_to_cpu(crq->request_capability_rsp.number);
+ init_sub_crqs(adapter, 1);
+ return;
+ default:
+ dev_err(dev, "Error %d in request cap rsp\n",
+ crq->request_capability_rsp.rc.code);
+ return;
+ }
+
+ /* Done receiving requested capabilities, query IP offload support */
+ if (++adapter->requested_caps == 6) {
+ union ibmvnic_crq newcrq;
+ int buf_sz = sizeof(struct ibmvnic_query_ip_offload_buffer);
+ struct ibmvnic_query_ip_offload_buffer *ip_offload_buf =
+ &adapter->ip_offload_buf;
+
+ adapter->ip_offload_tok = dma_map_single(dev, ip_offload_buf,
+ buf_sz,
+ DMA_FROM_DEVICE);
+
+ if (dma_mapping_error(dev, adapter->ip_offload_tok)) {
+ if (!firmware_has_feature(FW_FEATURE_CMO))
+ dev_err(dev, "Couldn't map offload buffer\n");
+ return;
+ }
+
+ memset(&newcrq, 0, sizeof(newcrq));
+ newcrq.query_ip_offload.first = IBMVNIC_CRQ_CMD;
+ newcrq.query_ip_offload.cmd = QUERY_IP_OFFLOAD;
+ newcrq.query_ip_offload.len = cpu_to_be32(buf_sz);
+ newcrq.query_ip_offload.ioba =
+ cpu_to_be32(adapter->ip_offload_tok);
+
+ ibmvnic_send_crq(adapter, &newcrq);
+ }
+}
+
+static int handle_login_rsp(union ibmvnic_crq *login_rsp_crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ int i;
+ struct ibmvnic_login_buffer *login = adapter->login_buf;
+ struct ibmvnic_login_rsp_buffer *login_rsp = adapter->login_rsp_buf;
+ union ibmvnic_crq crq;
+
+ dma_unmap_single(dev, adapter->login_buf_token, adapter->login_buf_sz,
+ DMA_BIDIRECTIONAL);
+ dma_unmap_single(dev, adapter->login_rsp_buf_token,
+ adapter->login_rsp_buf_sz, DMA_BIDIRECTIONAL);
+
+ netdev_dbg(adapter->netdev, "Login Response Buffer:\n");
+ for (i = 0; i < (adapter->login_rsp_buf_sz - 1) / 8 + 1; i++) {
+ netdev_dbg(adapter->netdev, "%016lx\n",
+ ((long unsigned int *)(adapter->login_rsp_buf))[i]);
+ }
+
+ /* Sanity checks */
+ if (login->num_txcomp_subcrqs != login_rsp->num_txsubm_subcrqs ||
+ (be32_to_cpu(login->num_rxcomp_subcrqs) *
+ adapter->req_rx_add_queues !=
+ be32_to_cpu(login_rsp->num_rxadd_subcrqs))) {
+ dev_err(dev, "FATAL: Inconsistent login and login rsp\n");
+ ibmvnic_remove(adapter->vdev);
+ return -EIO;
+ }
+
+ complete(&adapter->init_done);
+
+ memset(&crq, 0, sizeof(crq));
+ crq.request_ras_comp_num.first = IBMVNIC_CRQ_CMD;
+ crq.request_ras_comp_num.cmd = REQUEST_RAS_COMP_NUM;
+ ibmvnic_send_crq(adapter, &crq);
+
+ return 0;
+}
+
+static void handle_query_cap_rsp(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ struct net_device *netdev = adapter->netdev;
+ long rc;
+
+ atomic_dec(&adapter->running_cap_queries);
+ netdev_dbg(netdev, "Outstanding queries: %d\n",
+ atomic_read(&adapter->running_cap_queries));
+ rc = crq->query_capability.rc.code;
+ if (rc) {
+ dev_err(dev, "Error %ld in QUERY_CAP_RSP\n", rc);
+ goto out;
+ }
+
+ switch (be16_to_cpu(crq->query_capability.capability)) {
+ case MIN_TX_QUEUES:
+ adapter->min_tx_queues =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "min_tx_queues = %lld\n",
+ adapter->min_tx_queues);
+ break;
+ case MIN_RX_QUEUES:
+ adapter->min_rx_queues =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "min_rx_queues = %lld\n",
+ adapter->min_rx_queues);
+ break;
+ case MIN_RX_ADD_QUEUES:
+ adapter->min_rx_add_queues =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "min_rx_add_queues = %lld\n",
+ adapter->min_rx_add_queues);
+ break;
+ case MAX_TX_QUEUES:
+ adapter->max_tx_queues =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "max_tx_queues = %lld\n",
+ adapter->max_tx_queues);
+ if (adapter->max_tx_queues > ibmvnic_max_tx_queues)
+ netdev_warn(netdev, "Adapter's max_tx_queues is "
+ "greater than module parameter. Set module "
+ "paramenter to %lld, for optimum "
+ "performance\n", adapter->max_tx_queues);
+ break;
+ case MAX_RX_QUEUES:
+ adapter->max_rx_queues =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "max_rx_queues = %lld\n",
+ adapter->max_rx_queues);
+ break;
+ case MAX_RX_ADD_QUEUES:
+ adapter->max_rx_add_queues =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "max_rx_add_queues = %lld\n",
+ adapter->max_rx_add_queues);
+ break;
+ case MIN_TX_ENTRIES_PER_SUBCRQ:
+ adapter->min_tx_entries_per_subcrq =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "min_tx_entries_per_subcrq = %lld\n",
+ adapter->min_tx_entries_per_subcrq);
+ break;
+ case MIN_RX_ADD_ENTRIES_PER_SUBCRQ:
+ adapter->min_rx_add_entries_per_subcrq =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "min_rx_add_entrs_per_subcrq = %lld\n",
+ adapter->min_rx_add_entries_per_subcrq);
+ break;
+ case MAX_TX_ENTRIES_PER_SUBCRQ:
+ adapter->max_tx_entries_per_subcrq =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "max_tx_entries_per_subcrq = %lld\n",
+ adapter->max_tx_entries_per_subcrq);
+ break;
+ case MAX_RX_ADD_ENTRIES_PER_SUBCRQ:
+ adapter->max_rx_add_entries_per_subcrq =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "max_rx_add_entrs_per_subcrq = %lld\n",
+ adapter->max_rx_add_entries_per_subcrq);
+ break;
+ case TCP_IP_OFFLOAD:
+ adapter->tcp_ip_offload =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "tcp_ip_offload = %lld\n",
+ adapter->tcp_ip_offload);
+ break;
+ case PROMISC_SUPPORTED:
+ adapter->promisc_supported =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "promisc_supported = %lld\n",
+ adapter->promisc_supported);
+ break;
+ case MIN_MTU:
+ adapter->min_mtu = be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "min_mtu = %lld\n", adapter->min_mtu);
+ break;
+ case MAX_MTU:
+ adapter->max_mtu = be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "max_mtu = %lld\n", adapter->max_mtu);
+ break;
+ case MAX_MULTICAST_FILTERS:
+ adapter->max_multicast_filters =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "max_multicast_filters = %lld\n",
+ adapter->max_multicast_filters);
+ break;
+ case VLAN_HEADER_INSERTION:
+ adapter->vlan_header_insertion =
+ be32_to_cpu(crq->query_capability.number);
+ if (adapter->vlan_header_insertion)
+ netdev->features |= NETIF_F_HW_VLAN_STAG_TX;
+ netdev_dbg(netdev, "vlan_header_insertion = %lld\n",
+ adapter->vlan_header_insertion);
+ break;
+ case MAX_TX_SG_ENTRIES:
+ adapter->max_tx_sg_entries =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "max_tx_sg_entries = %lld\n",
+ adapter->max_tx_sg_entries);
+ break;
+ case RX_SG_SUPPORTED:
+ adapter->rx_sg_supported =
+ be32_to_cpu(crq->query_capability.number);
+ netdev_dbg(netdev, "rx_sg_supported = %lld\n",
+ adapter->rx_sg_supported);
+ break;
+ default:
+ netdev_err(netdev, "Got invalid cap rsp %d\n",
+ crq->query_capability.capability);
+ }
+
+out:
+ if (atomic_read(&adapter->running_cap_queries) == 0)
+ init_sub_crqs(adapter, 0);
+ /* We're done querying the capabilities, initialize sub-crqs */
+}
+
+static void handle_control_ras_rsp(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ u8 correlator = crq->control_ras_rsp.correlator;
+ int found = 0;
+ int i;
+
+ if (crq->control_ras_rsp.rc.code) {
+ dev_warn(dev, "Control ras failed rc=%d\n",
+ crq->control_ras_rsp.rc.code);
+ return;
+ }
+
+ for (i = 0; i < adapter->ras_comp_num; i++) {
+ if (adapter->ras_comps[i].correlator == correlator) {
+ found = 1;
+ break;
+ }
+ }
+
+ if (!found) {
+ dev_warn(dev, "Correlator not found on control_ras_rsp\n");
+ return;
+ }
+
+ switch (crq->control_ras_rsp.op) {
+ case IBMVNIC_TRACE_LEVEL:
+ adapter->ras_comps[i].trace_level = crq->control_ras.level;
+ break;
+ case IBMVNIC_ERROR_LEVEL:
+ adapter->ras_comps[i].error_check_level =
+ crq->control_ras.level;
+ break;
+ case IBMVNIC_TRACE_PAUSE:
+ adapter->ras_comp_int[i].paused = 1;
+ break;
+ case IBMVNIC_TRACE_RESUME:
+ adapter->ras_comp_int[i].paused = 0;
+ break;
+ case IBMVNIC_TRACE_ON:
+ adapter->ras_comps[i].trace_on = 1;
+ break;
+ case IBMVNIC_TRACE_OFF:
+ adapter->ras_comps[i].trace_on = 0;
+ break;
+ case IBMVNIC_CHG_TRACE_BUFF_SZ:
+ /* trace_buff_sz is 3 bytes, stuff it into an int */
+ ((u8 *)(&adapter->ras_comps[i].trace_buff_size))[0] = 0;
+ ((u8 *)(&adapter->ras_comps[i].trace_buff_size))[1] =
+ crq->control_ras_rsp.trace_buff_sz[0];
+ ((u8 *)(&adapter->ras_comps[i].trace_buff_size))[2] =
+ crq->control_ras_rsp.trace_buff_sz[1];
+ ((u8 *)(&adapter->ras_comps[i].trace_buff_size))[3] =
+ crq->control_ras_rsp.trace_buff_sz[2];
+ break;
+ default:
+ dev_err(dev, "invalid op %d on control_ras_rsp",
+ crq->control_ras_rsp.op);
+ }
+}
+
+static int ibmvnic_fw_comp_open(struct inode *inode, struct file *file)
+{
+ file->private_data = inode->i_private;
+ return 0;
+}
+
+static ssize_t trace_read(struct file *file, char __user *user_buf, size_t len,
+ loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ struct device *dev = &adapter->vdev->dev;
+ int num = ras_comp_int->num;
+ struct ibmvnic_fw_trace_entry *trace;
+ dma_addr_t trace_tok;
+ union ibmvnic_crq crq;
+
+ if (*ppos >= be32_to_cpu(adapter->ras_comps[num].trace_buff_size))
+ return 0;
+
+ trace =
+ dma_alloc_coherent(dev,
+ be32_to_cpu(adapter->ras_comps[num].
+ trace_buff_size), &trace_tok,
+ GFP_KERNEL);
+ if (!trace) {
+ dev_err(dev, "Couldn't alloc trace buffer\n");
+ return 0;
+ }
+
+ memset(&crq, 0, sizeof(crq));
+ crq.collect_fw_trace.first = IBMVNIC_CRQ_CMD;
+ crq.collect_fw_trace.cmd = COLLECT_FW_TRACE;
+ crq.collect_fw_trace.correlator = adapter->ras_comps[num].correlator;
+ crq.collect_fw_trace.ioba = cpu_to_be32(trace_tok);
+ crq.collect_fw_trace.len = adapter->ras_comps[num].trace_buff_size;
+ ibmvnic_send_crq(adapter, &crq);
+
+ init_completion(&adapter->fw_done);
+ wait_for_completion(&adapter->fw_done);
+
+ if (*ppos + len > be32_to_cpu(adapter->ras_comps[num].trace_buff_size))
+ len =
+ be32_to_cpu(adapter->ras_comps[num].trace_buff_size) -
+ *ppos;
+
+ copy_to_user(user_buf, &((u8 *) trace)[*ppos], len);
+
+ dma_free_coherent(dev,
+ be32_to_cpu(adapter->ras_comps[num].trace_buff_size),
+ trace, trace_tok);
+ *ppos += len;
+ return len;
+}
+
+static const struct file_operations trace_ops = {
+ .owner = THIS_MODULE,
+ .open = ibmvnic_fw_comp_open,
+ .read = trace_read,
+};
+
+static ssize_t paused_read(struct file *file, char __user *user_buf, size_t len,
+ loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[5]; /* 1 or 0 plus \n and \0 */
+ int num = ras_comp_int->num;
+ int size;
+
+ size = sprintf(buff, "%d\n", adapter->ras_comp_int[num].paused);
+
+ if (*ppos >= size)
+ return 0;
+
+ copy_to_user(user_buf, buff, size);
+ *ppos += size;
+ return size;
+}
+
+static ssize_t paused_write(struct file *file, const char __user *user_buf,
+ size_t len, loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[9]; /* decimal max int plus \n and \0 */
+ unsigned long val;
+ union ibmvnic_crq crq;
+ int num = ras_comp_int->num;
+
+ copy_from_user(buff, user_buf, sizeof(buff));
+ val = simple_strtoul(buff, NULL, 10);
+
+ adapter->ras_comp_int[num].paused = val ? 1 : 0;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.control_ras.first = IBMVNIC_CRQ_CMD;
+ crq.control_ras.cmd = CONTROL_RAS;
+ crq.control_ras.correlator = adapter->ras_comps[num].correlator;
+ crq.control_ras.op = val ? IBMVNIC_TRACE_PAUSE : IBMVNIC_TRACE_RESUME;
+ ibmvnic_send_crq(adapter, &crq);
+
+ return len;
+}
+
+static const struct file_operations paused_ops = {
+ .owner = THIS_MODULE,
+ .open = ibmvnic_fw_comp_open,
+ .read = paused_read,
+ .write = paused_write,
+};
+
+static ssize_t tracing_read(struct file *file, char __user *user_buf,
+ size_t len, loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[5]; /* 1 or 0 plus \n and \0 */
+ int num = ras_comp_int->num;
+ int size;
+
+ size = sprintf(buff, "%d\n", adapter->ras_comps[num].trace_on);
+
+ if (*ppos >= size)
+ return 0;
+
+ copy_to_user(user_buf, buff, size);
+ *ppos += size;
+ return size;
+}
+
+static ssize_t tracing_write(struct file *file, const char __user *user_buf,
+ size_t len, loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[9]; /* decimal max int plus \n and \0 */
+ unsigned long val;
+ union ibmvnic_crq crq;
+ int num = ras_comp_int->num;
+
+ copy_from_user(buff, user_buf, sizeof(buff));
+ val = simple_strtoul(buff, NULL, 10);
+
+ memset(&crq, 0, sizeof(crq));
+ crq.control_ras.first = IBMVNIC_CRQ_CMD;
+ crq.control_ras.cmd = CONTROL_RAS;
+ crq.control_ras.correlator = adapter->ras_comps[num].correlator;
+ crq.control_ras.op = val ? IBMVNIC_TRACE_ON : IBMVNIC_TRACE_OFF;
+
+ return len;
+}
+
+static const struct file_operations tracing_ops = {
+ .owner = THIS_MODULE,
+ .open = ibmvnic_fw_comp_open,
+ .read = tracing_read,
+ .write = tracing_write,
+};
+
+static ssize_t error_level_read(struct file *file, char __user *user_buf,
+ size_t len, loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[5]; /* decimal max char plus \n and \0 */
+ int num = ras_comp_int->num;
+ int size;
+
+ size = sprintf(buff, "%d\n", adapter->ras_comps[num].error_check_level);
+
+ if (*ppos >= size)
+ return 0;
+
+ copy_to_user(user_buf, buff, size);
+ *ppos += size;
+ return size;
+}
+
+static ssize_t error_level_write(struct file *file, const char __user *user_buf,
+ size_t len, loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[9]; /* decimal max int plus \n and \0 */
+ unsigned long val;
+ union ibmvnic_crq crq;
+ int num = ras_comp_int->num;
+
+ copy_from_user(buff, user_buf, sizeof(buff));
+ val = simple_strtoul(buff, NULL, 10);
+
+ if (val > 9)
+ val = 9;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.control_ras.first = IBMVNIC_CRQ_CMD;
+ crq.control_ras.cmd = CONTROL_RAS;
+ crq.control_ras.correlator = adapter->ras_comps[num].correlator;
+ crq.control_ras.op = IBMVNIC_ERROR_LEVEL;
+ crq.control_ras.level = val;
+ ibmvnic_send_crq(adapter, &crq);
+
+ return len;
+}
+
+static const struct file_operations error_level_ops = {
+ .owner = THIS_MODULE,
+ .open = ibmvnic_fw_comp_open,
+ .read = error_level_read,
+ .write = error_level_write,
+};
+
+static ssize_t trace_level_read(struct file *file, char __user *user_buf,
+ size_t len, loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[5]; /* decimal max char plus \n and \0 */
+ int num = ras_comp_int->num;
+ int size;
+
+ size = sprintf(buff, "%d\n", adapter->ras_comps[num].trace_level);
+ if (*ppos >= size)
+ return 0;
+
+ copy_to_user(user_buf, buff, size);
+ *ppos += size;
+ return size;
+}
+
+static ssize_t trace_level_write(struct file *file, const char __user *user_buf,
+ size_t len, loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[9]; /* decimal max int plus \n and \0 */
+ unsigned long val;
+ union ibmvnic_crq crq;
+
+ copy_from_user(buff, user_buf, sizeof(buff));
+ val = simple_strtoul(buff, NULL, 10);
+ if (val > 9)
+ val = 9;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.control_ras.first = IBMVNIC_CRQ_CMD;
+ crq.control_ras.cmd = CONTROL_RAS;
+ crq.control_ras.correlator =
+ adapter->ras_comps[ras_comp_int->num].correlator;
+ crq.control_ras.op = IBMVNIC_TRACE_LEVEL;
+ crq.control_ras.level = val;
+ ibmvnic_send_crq(adapter, &crq);
+
+ return len;
+}
+
+static const struct file_operations trace_level_ops = {
+ .owner = THIS_MODULE,
+ .open = ibmvnic_fw_comp_open,
+ .read = trace_level_read,
+ .write = trace_level_write,
+};
+
+static ssize_t trace_buff_size_read(struct file *file, char __user *user_buf,
+ size_t len, loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[9]; /* decimal max int plus \n and \0 */
+ int num = ras_comp_int->num;
+ int size;
+
+ size = sprintf(buff, "%d\n", adapter->ras_comps[num].trace_buff_size);
+ if (*ppos >= size)
+ return 0;
+
+ copy_to_user(user_buf, buff, size);
+ *ppos += size;
+ return size;
+}
+
+static ssize_t trace_buff_size_write(struct file *file,
+ const char __user *user_buf, size_t len,
+ loff_t *ppos)
+{
+ struct ibmvnic_fw_comp_internal *ras_comp_int = file->private_data;
+ struct ibmvnic_adapter *adapter = ras_comp_int->adapter;
+ char buff[9]; /* decimal max int plus \n and \0 */
+ unsigned long val;
+ union ibmvnic_crq crq;
+
+ copy_from_user(buff, user_buf, sizeof(buff));
+ val = simple_strtoul(buff, NULL, 10);
+
+ memset(&crq, 0, sizeof(crq));
+ crq.control_ras.first = IBMVNIC_CRQ_CMD;
+ crq.control_ras.cmd = CONTROL_RAS;
+ crq.control_ras.correlator =
+ adapter->ras_comps[ras_comp_int->num].correlator;
+ crq.control_ras.op = IBMVNIC_CHG_TRACE_BUFF_SZ;
+ /* trace_buff_sz is 3 bytes, stuff an int into it */
+ crq.control_ras.trace_buff_sz[0] = ((u8 *) (&val))[5];
+ crq.control_ras.trace_buff_sz[1] = ((u8 *) (&val))[6];
+ crq.control_ras.trace_buff_sz[2] = ((u8 *) (&val))[7];
+ ibmvnic_send_crq(adapter, &crq);
+
+ return len;
+}
+
+static const struct file_operations trace_size_ops = {
+ .owner = THIS_MODULE,
+ .open = ibmvnic_fw_comp_open,
+ .read = trace_buff_size_read,
+ .write = trace_buff_size_write,
+};
+
+static void handle_request_ras_comps_rsp(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ struct dentry *dir_ent;
+ struct dentry *ent;
+ int i;
+
+ debugfs_remove_recursive(adapter->ras_comps_ent);
+
+ adapter->ras_comps_ent = debugfs_create_dir("ras_comps",
+ adapter->debugfs_dir);
+ if (!adapter->ras_comps_ent || IS_ERR(adapter->ras_comps_ent)) {
+ dev_info(dev, "debugfs create ras_comps dir failed\n");
+ return;
+ }
+
+ for (i = 0; i < adapter->ras_comp_num; i++) {
+ dir_ent = debugfs_create_dir(adapter->ras_comps[i].name,
+ adapter->ras_comps_ent);
+ if (!dir_ent || IS_ERR(dir_ent)) {
+ dev_info(dev, "debugfs create %s dir failed\n",
+ adapter->ras_comps[i].name);
+ continue;
+ }
+
+ adapter->ras_comp_int[i].adapter = adapter;
+ adapter->ras_comp_int[i].num = i;
+ adapter->ras_comp_int[i].desc_blob.data =
+ &adapter->ras_comps[i].description;
+ adapter->ras_comp_int[i].desc_blob.size =
+ sizeof(adapter->ras_comps[i].description);
+
+ /* Don't need to remember the dentry's because the debugfs dir
+ * gets removed recursively
+ */
+ ent = debugfs_create_blob("description", S_IRUGO, dir_ent,
+ &adapter->ras_comp_int[i].desc_blob);
+ ent = debugfs_create_file("trace_buf_size", S_IRUGO | S_IWUSR,
+ dir_ent, &adapter->ras_comp_int[i],
+ &trace_size_ops);
+ ent = debugfs_create_file("trace_level",
+ S_IRUGO |
+ (adapter->ras_comps[i].trace_level !=
+ 0xFF ? S_IWUSR : 0),
+ dir_ent, &adapter->ras_comp_int[i],
+ &trace_level_ops);
+ ent = debugfs_create_file("error_level",
+ S_IRUGO |
+ (adapter->
+ ras_comps[i].error_check_level !=
+ 0xFF ? S_IWUSR : 0),
+ dir_ent, &adapter->ras_comp_int[i],
+ &trace_level_ops);
+ ent = debugfs_create_file("tracing", S_IRUGO | S_IWUSR,
+ dir_ent, &adapter->ras_comp_int[i],
+ &tracing_ops);
+ ent = debugfs_create_file("paused", S_IRUGO | S_IWUSR,
+ dir_ent, &adapter->ras_comp_int[i],
+ &paused_ops);
+ ent = debugfs_create_file("trace", S_IRUGO, dir_ent,
+ &adapter->ras_comp_int[i],
+ &trace_ops);
+ }
+}
+
+static void handle_request_ras_comp_num_rsp(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ union ibmvnic_crq newcrq;
+ int len = adapter->ras_comp_num * sizeof(struct ibmvnic_fw_component);
+
+ adapter->ras_comps = dma_alloc_coherent(dev, len,
+ &adapter->ras_comps_tok,
+ GFP_KERNEL);
+ if (!adapter->ras_comps) {
+ if (!firmware_has_feature(FW_FEATURE_CMO))
+ dev_err(dev, "Couldn't alloc fw comps buffer\n");
+ return;
+ }
+
+ adapter->ras_comp_int = kmalloc(adapter->ras_comp_num *
+ sizeof(struct ibmvnic_fw_comp_internal),
+ GFP_KERNEL);
+ if (!adapter->ras_comp_int)
+ dma_free_coherent(dev, len, adapter->ras_comps,
+ adapter->ras_comps_tok);
+
+ memset(&newcrq, 0, sizeof(newcrq));
+ newcrq.request_ras_comps.first = IBMVNIC_CRQ_CMD;
+ newcrq.request_ras_comps.cmd = REQUEST_RAS_COMPS;
+ newcrq.request_ras_comps.ioba = cpu_to_be32(adapter->ras_comps_tok);
+ newcrq.request_ras_comps.len = cpu_to_be32(len);
+ ibmvnic_send_crq(adapter, &newcrq);
+}
+
+static void ibmvnic_free_inflight(struct ibmvnic_adapter *adapter)
+{
+ struct device *dev = &adapter->vdev->dev;
+ struct ibmvnic_inflight_cmd *inflight_cmd;
+ struct ibmvnic_error_buff *error_buff;
+ unsigned long flags;
+ unsigned long flags2;
+
+ spin_lock_irqsave(&adapter->inflight_lock, flags);
+ list_for_each_entry(inflight_cmd, &adapter->inflight, list) {
+ switch (inflight_cmd->crq.generic.cmd) {
+ case LOGIN:
+ dma_unmap_single(dev, adapter->login_buf_token,
+ adapter->login_buf_sz,
+ DMA_BIDIRECTIONAL);
+ dma_unmap_single(dev, adapter->login_rsp_buf_token,
+ adapter->login_rsp_buf_sz,
+ DMA_BIDIRECTIONAL);
+ kfree(adapter->login_rsp_buf);
+ kfree(adapter->login_buf);
+ break;
+ case REQUEST_DUMP:
+ complete(&adapter->fw_done);
+ break;
+ case REQUEST_ERROR_INFO:
+ spin_lock_irqsave(&adapter->error_list_lock, flags2);
+ list_for_each_entry(error_buff, &adapter->errors,
+ list) {
+ dma_unmap_single(dev, error_buff->dma,
+ error_buff->len,
+ DMA_FROM_DEVICE);
+ kfree(error_buff->buff);
+ list_del(&error_buff->list);
+ kfree(error_buff);
+ }
+ spin_unlock_irqrestore(&adapter->error_list_lock,
+ flags2);
+ break;
+ }
+ list_del(&inflight_cmd->list);
+ kfree(inflight_cmd);
+ }
+ spin_unlock_irqrestore(&adapter->inflight_lock, flags);
+}
+
+static void ibmvnic_handle_crq(union ibmvnic_crq *crq,
+ struct ibmvnic_adapter *adapter)
+{
+ struct ibmvnic_generic_crq *gen_crq = &crq->generic;
+ struct device *dev = &adapter->vdev->dev;
+ struct net_device *netdev = adapter->netdev;
+ long rc;
+
+ netdev_dbg(netdev, "Handling CRQ: %016lx %016lx\n",
+ ((long unsigned int *)crq)[0],
+ ((long unsigned int *)crq)[1]);
+ switch (gen_crq->first) {
+ case IBMVNIC_CRQ_INIT_RSP:
+ switch (gen_crq->cmd) {
+ case IBMVNIC_CRQ_INIT:
+ dev_info(dev, "Partner initialized\n");
+ /* Send back a response */
+ rc = ibmvnic_send_crq_init_complete(adapter);
+ if (rc == 0)
+ send_version_xchg(adapter);
+ else
+ dev_err(dev, "Can't send initrsp rc=%ld\n", rc);
+ break;
+ case IBMVNIC_CRQ_INIT_COMPLETE:
+ dev_info(dev, "Partner initialization complete\n");
+ send_version_xchg(adapter);
+ break;
+ default:
+ dev_err(dev, "Unknown crq cmd: %d\n", gen_crq->cmd);
+ }
+ return;
+ case IBMVNIC_CRQ_XPORT_EVENT:
+ if (gen_crq->cmd == IBMVNIC_PARTITION_MIGRATED) {
+ dev_info(dev, "Re-enabling adapter\n");
+ adapter->migrated = 1;
+ ibmvnic_free_inflight(adapter);
+ release_sub_crqs(adapter);
+ rc = ibmvnic_reenable_crq_queue(adapter);
+ if (rc)
+ dev_err(dev, "Error after enable rc=%ld\n", rc);
+ adapter->migrated = 0;
+ rc = ibmvnic_send_crq_init(adapter);
+ if (rc)
+ dev_err(dev, "Error sending init rc=%ld\n", rc);
+ } else {
+ /* The adapter lost the connection */
+ dev_err(dev, "Virtual Adapter failed (rc=%d)\n",
+ gen_crq->cmd);
+ ibmvnic_free_inflight(adapter);
+ release_sub_crqs(adapter);
+ }
+ return;
+ case IBMVNIC_CRQ_CMD_RSP:
+ break;
+ default:
+ dev_err(dev, "Got an invalid msg type 0x%02x\n",
+ gen_crq->first);
+ return;
+ }
+
+ switch (gen_crq->cmd) {
+ case VERSION_EXCHANGE_RSP:
+ rc = crq->version_exchange_rsp.rc.code;
+ if (rc) {
+ dev_err(dev, "Error %ld in VERSION_EXCHG_RSP\n", rc);
+ break;
+ }
+ dev_info(dev, "Partner protocol version is %d\n",
+ crq->version_exchange_rsp.version);
+ if (be16_to_cpu(crq->version_exchange_rsp.version) <
+ ibmvnic_version)
+ ibmvnic_version =
+ be16_to_cpu(crq->version_exchange_rsp.version);
+ send_cap_queries(adapter);
+ break;
+ case QUERY_CAPABILITY_RSP:
+ handle_query_cap_rsp(crq, adapter);
+ break;
+ case REQUEST_CAPABILITY_RSP:
+ handle_request_cap_rsp(crq, adapter);
+ break;
+ case LOGIN_RSP:
+ netdev_dbg(netdev, "Got Login Response\n");
+ handle_login_rsp(crq, adapter);
+ break;
+ case LOGICAL_LINK_STATE_RSP:
+ netdev_dbg(netdev, "Got Logical Link State Response\n");
+ adapter->logical_link_state =
+ crq->logical_link_state_rsp.link_state;
+ break;
+ case LINK_STATE_INDICATION:
+ netdev_dbg(netdev, "Got Logical Link State Indication\n");
+ adapter->phys_link_state =
+ crq->link_state_indication.phys_link_state;
+ adapter->logical_link_state =
+ crq->link_state_indication.logical_link_state;
+ break;
+ case CHANGE_MAC_ADDR_RSP:
+ netdev_dbg(netdev, "Got MAC address change Response\n");
+ handle_change_mac_rsp(crq, adapter);
+ break;
+ case ERROR_INDICATION:
+ netdev_dbg(netdev, "Got Error Indication\n");
+ handle_error_indication(crq, adapter);
+ break;
+ case REQUEST_ERROR_RSP:
+ netdev_dbg(netdev, "Got Error Detail Response\n");
+ handle_error_info_rsp(crq, adapter);
+ break;
+ case REQUEST_STATISTICS_RSP:
+ netdev_dbg(netdev, "Got Statistics Response\n");
+ complete(&adapter->stats_done);
+ break;
+ case REQUEST_DUMP_SIZE_RSP:
+ netdev_dbg(netdev, "Got Request Dump Size Response\n");
+ handle_dump_size_rsp(crq, adapter);
+ break;
+ case REQUEST_DUMP_RSP:
+ netdev_dbg(netdev, "Got Request Dump Response\n");
+ complete(&adapter->fw_done);
+ break;
+ case QUERY_IP_OFFLOAD_RSP:
+ netdev_dbg(netdev, "Got Query IP offload Response\n");
+ handle_query_ip_offload_rsp(adapter);
+ break;
+ case MULTICAST_CTRL_RSP:
+ netdev_dbg(netdev, "Got multicast control Response\n");
+ break;
+ case CONTROL_IP_OFFLOAD_RSP:
+ netdev_dbg(netdev, "Got Control IP offload Response\n");
+ dma_unmap_single(dev, adapter->ip_offload_ctrl_tok,
+ sizeof(adapter->ip_offload_ctrl),
+ DMA_TO_DEVICE);
+ /* We're done with the queries, perform the login */
+ send_login(adapter);
+ break;
+ case REQUEST_RAS_COMP_NUM_RSP:
+ netdev_dbg(netdev, "Got Request RAS Comp Num Response\n");
+ adapter->ras_comp_num =
+ be32_to_cpu(crq->request_ras_comp_num_rsp.num_components);
+ handle_request_ras_comp_num_rsp(crq, adapter);
+ break;
+ case REQUEST_RAS_COMPS_RSP:
+ netdev_dbg(netdev, "Got Request RAS Comps Response\n");
+ handle_request_ras_comps_rsp(crq, adapter);
+ break;
+ case CONTROL_RAS_RSP:
+ netdev_dbg(netdev, "Got Control RAS Response\n");
+ handle_control_ras_rsp(crq, adapter);
+ break;
+ case COLLECT_FW_TRACE_RSP:
+ netdev_dbg(netdev, "Got Collect firmware trace Response\n");
+ complete(&adapter->fw_done);
+ break;
+ default:
+ netdev_err(netdev, "Got an invalid cmd type 0x%02x\n",
+ gen_crq->cmd);
+ }
+}
+
+static irqreturn_t ibmvnic_interrupt(int irq, void *instance)
+{
+ struct ibmvnic_adapter *adapter = instance;
+ struct vio_dev *vdev = adapter->vdev;
+ union ibmvnic_crq *crq;
+ int done = 0;
+
+ vio_disable_interrupts(vdev);
+ while (!done) {
+ /* Pull all the valid messages off the CRQ */
+ while ((crq = ibmvnic_next_crq(adapter)) != NULL) {
+ ibmvnic_handle_crq(crq, adapter);
+ crq->generic.first = 0;
+ }
+ vio_enable_interrupts(vdev);
+ crq = ibmvnic_next_crq(adapter);
+ if (crq != NULL) {
+ vio_disable_interrupts(vdev);
+ ibmvnic_handle_crq(crq, adapter);
+ crq->generic.first = 0;
+ } else {
+ done = 1;
+ }
+ }
+
+ return IRQ_HANDLED;
+}
+
+static int ibmvnic_reenable_crq_queue(struct ibmvnic_adapter *adapter)
+{
+ int rc;
+ struct vio_dev *vdev = adapter->vdev;
+
+ do {
+ rc = plpar_hcall_norets(H_ENABLE_CRQ, vdev->unit_address);
+ } while (rc == H_IN_PROGRESS || rc == H_BUSY || H_IS_LONG_BUSY(rc));
+
+ if (rc)
+ dev_err(&vdev->dev, "Error enabling adapter (rc=%d)\n", rc);
+
+ return rc;
+}
+
+static int ibmvnic_reset_crq(struct ibmvnic_adapter *adapter)
+{
+ int rc;
+ struct vio_dev *vdev = adapter->vdev;
+ struct ibmvnic_crq_queue *crq = &adapter->crq;
+ struct device *dev = &adapter->vdev->dev;
+
+ /* Close the CRQ */
+ do {
+ rc = plpar_hcall_norets(H_FREE_CRQ, vdev->unit_address);
+ } while (rc == H_BUSY || H_IS_LONG_BUSY(rc));
+
+ /* Clean out the queue */
+ memset(crq->msgs, 0, PAGE_SIZE);
+ crq->cur = 0;
+
+ /* And re-open it again */
+ rc = plpar_hcall_norets(H_REG_CRQ, vdev->unit_address,
+ crq->msg_token, PAGE_SIZE);
+
+ if (rc == H_CLOSED)
+ /* Adapter is good, but other end is not ready */
+ dev_warn(dev, "Partner adapter not ready\n");
+ else if (rc != 0)
+ dev_warn(dev, "Couldn't register crq (rc=%d)\n", rc);
+
+ return rc;
+}
+
+static void ibmvnic_release_crq_queue(struct ibmvnic_adapter *adapter)
+{
+ long rc;
+ struct vio_dev *vdev = adapter->vdev;
+ struct ibmvnic_crq_queue *crq = &adapter->crq;
+
+ netdev_dbg(adapter->netdev, "Releasing CRQ\n");
+ free_irq(vdev->irq, adapter);
+ do {
+ rc = plpar_hcall_norets(H_FREE_CRQ, vdev->unit_address);
+ } while (rc == H_BUSY || H_IS_LONG_BUSY(rc));
+
+ dma_unmap_single(&vdev->dev, crq->msg_token, PAGE_SIZE,
+ DMA_BIDIRECTIONAL);
+ free_page((unsigned long)crq->msgs);
+}
+
+static int ibmvnic_init_crq_queue(struct ibmvnic_adapter *adapter)
+{
+ int rc, retrc = -ENOMEM;
+ struct ibmvnic_crq_queue *crq = &adapter->crq;
+ struct vio_dev *vdev = adapter->vdev;
+ struct device *dev = &adapter->vdev->dev;
+
+ crq->msgs = (union ibmvnic_crq *)get_zeroed_page(GFP_KERNEL);
+ /* Should we allocate more than one page? */
+
+ if (!crq->msgs)
+ return -ENOMEM;
+
+ crq->size = PAGE_SIZE / sizeof(*crq->msgs);
+ crq->msg_token = dma_map_single(dev, crq->msgs, PAGE_SIZE,
+ DMA_BIDIRECTIONAL);
+ if (dma_mapping_error(dev, crq->msg_token))
+ goto map_failed;
+
+ retrc = rc = plpar_hcall_norets(H_REG_CRQ, vdev->unit_address,
+ crq->msg_token, PAGE_SIZE);
+
+ if (rc == H_RESOURCE)
+ /* maybe kexecing and resource is busy. try a reset */
+ retrc = rc = ibmvnic_reset_crq(adapter);
+
+ if (rc == H_CLOSED) {
+ dev_warn(dev, "Partner adapter not ready\n");
+ } else if (rc) {
+ dev_warn(dev, "Error %d opening adapter\n", rc);
+ goto reg_crq_failed;
+ }
+
+ retrc = 0;
+
+ netdev_dbg(adapter->netdev, "registering irq 0x%x\n", vdev->irq);
+ rc = request_irq(vdev->irq, ibmvnic_interrupt, 0, IBMVNIC_NAME,
+ adapter);
+ if (rc) {
+ dev_err(dev, "Couldn't register irq 0x%x. rc=%d\n",
+ vdev->irq, rc);
+ goto req_irq_failed;
+ }
+
+ rc = vio_enable_interrupts(vdev);
+ if (rc) {
+ dev_err(dev, "Error %d enabling interrupts\n", rc);
+ goto req_irq_failed;
+ }
+
+ crq->cur = 0;
+ spin_lock_init(&crq->lock);
+
+ return retrc;
+
+req_irq_failed:
+ do {
+ rc = plpar_hcall_norets(H_FREE_CRQ, vdev->unit_address);
+ } while (rc == H_BUSY || H_IS_LONG_BUSY(rc));
+reg_crq_failed:
+ dma_unmap_single(dev, crq->msg_token, PAGE_SIZE, DMA_BIDIRECTIONAL);
+map_failed:
+ free_page((unsigned long)crq->msgs);
+ return retrc;
+}
+
+/* debugfs for dump */
+static int ibmvnic_dump_show(struct seq_file *seq, void *v)
+{
+ struct net_device *netdev = seq->private;
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ struct device *dev = &adapter->vdev->dev;
+ union ibmvnic_crq crq;
+
+ memset(&crq, 0, sizeof(crq));
+ crq.request_dump_size.first = IBMVNIC_CRQ_CMD;
+ crq.request_dump_size.cmd = REQUEST_DUMP_SIZE;
+ ibmvnic_send_crq(adapter, &crq);
+
+ init_completion(&adapter->fw_done);
+ wait_for_completion(&adapter->fw_done);
+
+ seq_write(seq, adapter->dump_data, adapter->dump_data_size);
+
+ dma_unmap_single(dev, adapter->dump_data_token, adapter->dump_data_size,
+ DMA_BIDIRECTIONAL);
+
+ kfree(adapter->dump_data);
+
+ return 0;
+}
+
+static int ibmvnic_dump_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, ibmvnic_dump_show, inode->i_private);
+}
+
+static const struct file_operations ibmvnic_dump_ops = {
+ .owner = THIS_MODULE,
+ .open = ibmvnic_dump_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+};
+
+static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id)
+{
+ struct net_device *netdev;
+ struct ibmvnic_adapter *adapter;
+ unsigned char *mac_addr_p;
+ struct dentry *ent;
+ char buf[16];
+ int rc;
+
+ dev_dbg(&dev->dev, "entering ibmvnic_probe for UA 0x%x\n",
+ dev->unit_address);
+
+ mac_addr_p = (unsigned char *)vio_get_attribute(dev,
+ VETH_MAC_ADDR, NULL);
+ if (!mac_addr_p) {
+ dev_err(&dev->dev,
+ "(%s:%3.3d) ERROR: Can't find MAC_ADDR attribute\n",
+ __FILE__, __LINE__);
+ return 0;
+ }
+
+ netdev = alloc_etherdev_mq(sizeof(struct ibmvnic_adapter),
+ ibmvnic_max_tx_queues);
+ if (!netdev)
+ return -ENOMEM;
+
+ adapter = netdev_priv(netdev);
+ dev_set_drvdata(&dev->dev, netdev);
+ adapter->vdev = dev;
+ adapter->netdev = netdev;
+
+ memcpy(&adapter->mac_addr, mac_addr_p, netdev->addr_len);
+ memcpy(netdev->dev_addr, &adapter->mac_addr, netdev->addr_len);
+ netdev->irq = dev->irq;
+ netdev->netdev_ops = &ibmvnic_netdev_ops;
+ netdev->ethtool_ops = &ibmvnic_ethtool_ops;
+ SET_NETDEV_DEV(netdev, &dev->dev);
+ spin_lock_init(&adapter->stats_lock);
+
+ rc = ibmvnic_init_crq_queue(adapter);
+ if (rc) {
+ dev_err(&dev->dev, "Couldn't initialize crq. rc=%d\n", rc);
+ goto free_netdev;
+ }
+
+ INIT_LIST_HEAD(&adapter->errors);
+ INIT_LIST_HEAD(&adapter->inflight);
+ spin_lock_init(&adapter->error_list_lock);
+ spin_lock_init(&adapter->inflight_lock);
+
+ adapter->stats_token = dma_map_single(&dev->dev, &adapter->stats,
+ sizeof(struct ibmvnic_statistics),
+ DMA_FROM_DEVICE);
+ if (dma_mapping_error(&dev->dev, adapter->stats_token)) {
+ if (!firmware_has_feature(FW_FEATURE_CMO))
+ dev_err(&dev->dev, "Couldn't map stats buffer\n");
+ goto free_crq;
+ }
+
+ snprintf(buf, sizeof(buf), "ibmvnic_%x", dev->unit_address);
+ ent = debugfs_create_dir(buf, NULL);
+ if (!ent || IS_ERR(ent)) {
+ dev_info(&dev->dev, "debugfs create directory failed\n");
+ adapter->debugfs_dir = NULL;
+ } else {
+ adapter->debugfs_dir = ent;
+ ent = debugfs_create_file("dump", S_IRUGO, adapter->debugfs_dir,
+ netdev, &ibmvnic_dump_ops);
+ if (!ent || IS_ERR(ent)) {
+ dev_info(&dev->dev,
+ "debugfs create dump file failed\n");
+ adapter->debugfs_dump = NULL;
+ } else {
+ adapter->debugfs_dump = ent;
+ }
+ }
+ ibmvnic_send_crq_init(adapter);
+
+ init_completion(&adapter->init_done);
+ wait_for_completion(&adapter->init_done);
+
+ netdev->real_num_tx_queues = adapter->req_tx_queues;
+
+ rc = register_netdev(netdev);
+ if (rc) {
+ dev_err(&dev->dev, "failed to register netdev rc=%d\n", rc);
+ goto free_debugfs;
+ }
+ dev_info(&dev->dev, "ibmvnic registered\n");
+
+ return 0;
+
+free_debugfs:
+ if (adapter->debugfs_dir && !IS_ERR(adapter->debugfs_dir))
+ debugfs_remove_recursive(adapter->debugfs_dir);
+free_crq:
+ ibmvnic_release_crq_queue(adapter);
+free_netdev:
+ free_netdev(netdev);
+ return rc;
+}
+
+static int ibmvnic_remove(struct vio_dev *dev)
+{
+ struct net_device *netdev = dev_get_drvdata(&dev->dev);
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+
+ unregister_netdev(netdev);
+
+ release_sub_crqs(adapter);
+
+ ibmvnic_release_crq_queue(adapter);
+
+ if (adapter->debugfs_dir && !IS_ERR(adapter->debugfs_dir))
+ debugfs_remove_recursive(adapter->debugfs_dir);
+
+ if (adapter->ras_comps)
+ dma_free_coherent(&dev->dev,
+ adapter->ras_comp_num *
+ sizeof(struct ibmvnic_fw_component),
+ adapter->ras_comps, adapter->ras_comps_tok);
+
+ kfree(adapter->ras_comp_int);
+
+ free_netdev(netdev);
+ dev_set_drvdata(&dev->dev, NULL);
+
+ return 0;
+}
+
+static unsigned long ibmvnic_get_desired_dma(struct vio_dev *vdev)
+{
+ struct net_device *netdev = dev_get_drvdata(&vdev->dev);
+ struct ibmvnic_adapter *adapter;
+ unsigned long ret = 0;
+ struct iommu_table *tbl;
+ int i;
+
+ tbl = get_iommu_table_base(&vdev->dev);
+
+ /* netdev inits at probe time along with the structures we need below*/
+ if (netdev == NULL)
+ return IOMMU_PAGE_ALIGN(IBMVNIC_IO_ENTITLEMENT_DEFAULT, tbl);
+
+ adapter = netdev_priv(netdev);
+
+ ret += PAGE_SIZE; /* the crq message queue */
+ ret += adapter->bounce_buffer_size;
+ ret += IOMMU_PAGE_ALIGN(sizeof(struct ibmvnic_statistics), tbl);
+
+ for (i = 0; i < adapter->req_tx_queues + adapter->req_rx_queues; i++)
+ ret += PAGE_SIZE; /* the scrq message queue */
+
+ for (i = 0; i < be32_to_cpu(adapter->login_rsp_buf->num_rxadd_subcrqs);
+ i++)
+ ret += adapter->rx_pool[i].size *
+ IOMMU_PAGE_ALIGN(adapter->rx_pool[i].buff_size, tbl);
+
+ return ret;
+}
+
+static int ibmvnic_resume(struct device *dev)
+{
+ struct net_device *netdev = dev_get_drvdata(dev);
+ struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+ int i;
+
+ /* kick the interrupt handlers just in case we lost an interrupt */
+ for (i = 0; i < adapter->req_rx_queues; i++)
+ ibmvnic_interrupt_rx(adapter->rx_scrq[i]->irq,
+ adapter->rx_scrq[i]);
+
+ return 0;
+}
+
+static struct vio_device_id ibmvnic_device_table[] = {
+ {"network", "IBM,ibmvnic"},
+ {"", "" }
+};
+MODULE_DEVICE_TABLE(vio, ibmvnic_device_table);
+
+static const struct dev_pm_ops ibmvnic_pm_ops = {
+ .resume = ibmvnic_resume
+};
+
+static struct vio_driver ibmvnic_driver = {
+ .id_table = ibmvnic_device_table,
+ .probe = ibmvnic_probe,
+ .remove = ibmvnic_remove,
+ .get_desired_dma = ibmvnic_get_desired_dma,
+ .driver = {
+ .name = ibmvnic_driver_name,
+ .owner = THIS_MODULE,
+ .pm = &ibmvnic_pm_ops,
+ }
+};
+
+/* module functions */
+static int __init ibmvnic_module_init(void)
+{
+ printk(KERN_INFO "%s: %s %s\n",
+ ibmvnic_driver_name, ibmvnic_driver_string,
+ IBMVNIC_DRIVER_VERSION);
+
+ return vio_register_driver(&ibmvnic_driver);
+}
+
+static void __exit ibmvnic_module_exit(void)
+{
+ vio_unregister_driver(&ibmvnic_driver);
+}
+
+module_init(ibmvnic_module_init);
+module_exit(ibmvnic_module_exit);
--- a/arch/powerpc/include/asm/hvcall.h 2014-08-16 11:40:26.000000000 -0500
+++ b/arch/powerpc/include/asm/hvcall.h 2014-08-25 07:39:00.000000000 -0500
@@ -258,11 +258,16 @@
#define H_DEL_CONN 0x288
#define H_JOIN 0x298
#define H_VASI_STATE 0x2A4
+#define H_VIOCTL 0x2A8
#define H_ENABLE_CRQ 0x2B0
#define H_GET_EM_PARMS 0x2B8
#define H_SET_MPP 0x2D0
#define H_GET_MPP 0x2D4
+#define H_REG_SUB_CRQ 0x2DC
#define H_HOME_NODE_ASSOCIATIVITY 0x2EC
+#define H_FREE_SUB_CRQ 0x2E0
+#define H_SEND_SUB_CRQ 0x2E4
+#define H_SEND_SUB_CRQ_INDIRECT 0x2E8
#define H_BEST_ENERGY 0x2F4
#define H_XIRR_X 0x2FC
#define H_RANDOM 0x300
@@ -271,6 +276,21 @@
#define H_SET_MODE 0x31C
#define MAX_HCALL_OPCODE H_SET_MODE
+/* H_VIOCTL functions */
+#define GET_VIOA_DUMP_SIZE 0x01
+#define GET_VIOA_DUMP 0x02
+#define GET_ILLAN_NUM_VLAN_IDS 0x03
+#define GET_ILLAN_VLAN_ID_LIST 0x04
+#define GET_ILLAN_SWITCH_ID 0x05
+#define DISABLE_MIGRATION 0x06
+#define ENABLE_MIGRATION 0x07
+#define GET_PARTNER_INFO 0x08
+#define GET_PARTNER_WWPN_LIST 0x09
+#define DISABLE_ALL_VIO_INTS 0x0A
+#define DISABLE_VIO_INTERRUPT 0x0B
+#define ENABLE_VIO_INTERRUPT 0x0C
+
+
/* Platform specific hcalls, used by KVM */
#define H_RTAS 0xf000
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox