[PATCH 1/1] net/{octeon_ep,enetfec}: remove unused value
Coverity issue: 385414, 374016
Fixes: c836a7ba33e ("net/octeon_ep: support mailbox between VF and PF")
Fixes: bb5b5bf1e5c ("net/enetfec: support queue configuration")
Signed-off-by: Ariel Otilibili
--
Cc: sta...@dpdk.org
Cc: Vamsi Attunuru
Cc: Apeksha Gupta
Cc: Sachin Saxena
---
drivers/net/enetfec/enet_ethdev.c | 1 -
drivers/net/octeon_ep/otx_ep_mbox.c | 3 ---
2 files changed, 4 deletions(-)
diff --git a/drivers/net/enetfec/enet_ethdev.c
b/drivers/net/enetfec/enet_ethdev.c
index 91c0f60490..0d27b63953 100644
--- a/drivers/net/enetfec/enet_ethdev.c
+++ b/drivers/net/enetfec/enet_ethdev.c
@@ -415,7 +415,6 @@ enetfec_tx_queue_setup(struct rte_eth_dev *dev,
offset_des_active_txq[queue_idx];
bd_base = (struct bufdesc *)(((uintptr_t)bd_base) + size);
txq->bd.last = (struct bufdesc *)(((uintptr_t)bd_base) - dsize);
- bdp = txq->bd.base;
bdp = txq->bd.cur;
for (i = 0; i < txq->bd.ring_size; i++) {
diff --git a/drivers/net/octeon_ep/otx_ep_mbox.c
b/drivers/net/octeon_ep/otx_ep_mbox.c
index 64a51c1fd0..1d7e08d2cc 100644
--- a/drivers/net/octeon_ep/otx_ep_mbox.c
+++ b/drivers/net/octeon_ep/otx_ep_mbox.c
@@ -256,9 +256,6 @@ int otx_ep_mbox_get_link_info(struct rte_eth_dev *eth_dev,
}
link->link_status = RTE_ETH_LINK_UP;
link->link_duplex = RTE_ETH_LINK_FULL_DUPLEX;
- link->link_autoneg = (link_info.autoneg ==
- OTX_EP_LINK_AUTONEG) ? RTE_ETH_LINK_AUTONEG :
RTE_ETH_LINK_FIXED;
-
link->link_autoneg = link_info.autoneg;
link->link_speed = link_info.speed;
return 0;
--
2.47.1
[PATCH 0/1] net/{octeon_ep,enetfec}: remove unused value
Hello,
This patch clears out the Coverity issues 385414 & 374016.
Thank you,
Ariel Otilibili (1):
net/{octeon_ep,enetfec}: remove unused value
drivers/net/enetfec/enet_ethdev.c | 1 -
drivers/net/octeon_ep/otx_ep_mbox.c | 3 ---
2 files changed, 4 deletions(-)
--
2.47.1
RE: [PATCH] net/nfp: implement the burst mode get operation
> On Wed, 18 Dec 2024 14:30:36 +0800
> Chaoyong He wrote:
>
> > +int
> > +nfp_net_rx_burst_mode_get(struct rte_eth_dev *eth_dev,
> > + uint16_t queue_id __rte_unused,
> > + struct rte_eth_burst_mode *mode)
> > +{
> > + eth_rx_burst_t pkt_burst;
> > +
> > + pkt_burst = eth_dev->rx_pkt_burst;
> > + if (pkt_burst == nfp_net_recv_pkts) {
> > + snprintf(mode->info, RTE_ETH_BURST_MODE_INFO_SIZE,
> "%s",
> > + "Scalar");
> > + } else if (pkt_burst == nfp_net_vec_avx2_recv_pkts) {
> > + snprintf(mode->info, RTE_ETH_BURST_MODE_INFO_SIZE,
> "%s",
> > + "Vector AVX2");
> > + } else {
> > + return -EINVAL;
> > + }
> > +
> > + return 0;
> > +}
> > +
>
> The coccinelle script want to replace that snprintf with strlcpy
>
> Also don't need {} for one line statement but its ok as is.
Okay, I will send a new version patch.
Re: [PATCH] net/virtio: fix Rx checksum calculation
On 2024-12-18 09:59, Maxime Coquelin wrote:
Hi,
On 12/18/24 08:34, Wangyunjian(wangyunjian,TongTu) wrote:
-Original Message-
From: Maxime Coquelin [mailto:maxime.coque...@redhat.com]
Sent: Tuesday, December 17, 2024 11:33 PM
To: dev@dpdk.org
Cc: Olivier Matz ; Maxime Gouin
; Maxime Coquelin
Subject: [PATCH] net/virtio: fix Rx checksum calculation
From: Olivier Matz
If hdr->csum_start is larger than packet length, the len argument passed
to rte_raw_cksum_mbuf() overflows and causes a segmentation fault.
Ignore checksum computation in this case.
CVE-2024-11614
Fixes: ca7036b4af3a ("vhost: fix offload flags in Rx path")
Signed-off-by: Maxime Gouin
Signed-off-by: Olivier Matz
Reviewed-by: Maxime Coquelin
---
lib/vhost/virtio_net.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
index d764d4bc6a..69901ab3b5 100644
--- a/lib/vhost/virtio_net.c
+++ b/lib/vhost/virtio_net.c
@@ -2823,6 +2823,9 @@ vhost_dequeue_offload(struct virtio_net *dev,
struct virtio_net_hdr *hdr,
*/
uint16_t csum = 0, off;
+ if (hdr->csum_start >= rte_pktmbuf_pkt_len(m))
+ return;
+
The hdr->csum_start does two successive reads from user space to read
a variable length data structure. The result overflow if the data
structure
changes between the two reads.
You don't know if the resulting object code will perform one, two or
more loads from memory.
If you want to be sure it's exactly one load, you need to go through a
volatile pointer. This seems like a use case for RTE_READ_ONCE() (which
doesn't exist).
An alternative is to do a relaxed atomic load, which is really what you
ask for.
We can prevent double fetch issue by using the temporary variable
csum_start.
I don't think that makes a difference. The compiler is still free to
generate object code which load from the same location multiple times.
Right, that's a good catch! The exploitation od this issue seem
difficult though.
We may systematically copy the full header, as we only do it for ones
not contiguous in host VA space.
I think you would need to go through a volatile pointer here as well, to
make sure the copy actually occur. At least if the target is
stack-allocated object.
What do you think? Are you willing to contribute a fix?
Thanks,
Maxime
Thanks,
Yunjian
if (rte_raw_cksum_mbuf(m, hdr->csum_start,
rte_pktmbuf_pkt_len(m) - hdr->csum_start, &csum) <
0)
return;
--
2.47.0
[PATCH] bus/fslmc: fix use after rte_free
The cleanup loop would derefence the dpio_dev after freeing.
Use TAILQ_FOREACH_SAFE to fix that.
Found by building with sanitizer undefined flag.
Fixes: e55d0494ab98 ("bus/fslmc: support secondary process")
Cc: shreyansh.j...@nxp.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index 2dfcf7a498..fb7d1968d1 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -403,6 +403,7 @@ dpaa2_create_dpio_device(int vdev_fd,
struct rte_dpaa2_device *obj)
{
struct dpaa2_dpio_dev *dpio_dev = NULL;
+ struct dpaa2_dpio_dev *dpio_tmp;
struct vfio_region_info reg_info = { .argsz = sizeof(reg_info)};
struct qbman_swp_desc p_des;
struct dpio_attr attr;
@@ -588,7 +589,7 @@ dpaa2_create_dpio_device(int vdev_fd,
rte_free(dpio_dev);
/* For each element in the list, cleanup */
- TAILQ_FOREACH(dpio_dev, &dpio_dev_list, next) {
+ RTE_TAILQ_FOREACH_SAFE(dpio_dev, &dpio_dev_list, next, dpio_tmp) {
if (dpio_dev->dpio) {
dpio_disable(dpio_dev->dpio, CMD_PRI_LOW,
dpio_dev->token);
--
2.45.2
Re: [PATCH v2] MAINTAINERS: change maintainer for next-net
10/12/2024 21:41, Stephen Hemminger: > Change of maintainers for next-net tree. > > Signed-off-by: Stephen Hemminger > Acked-by: Ferruh Yigit > --- > Next-net Tree > -M: Ferruh Yigit > -M: Andrew Rybchenko > +M: Stephen Hemminger > T: git://dpdk.org/next/dpdk-next-net Applied, thanks all for the help.
Re: [PATCH 0/3] Defer lcore variables allocation
On 2024-12-16 10:49, David Marchand wrote: On Mon, Dec 16, 2024 at 10:42 AM Burakov, Anatoly wrote: On 12/5/2024 6:57 PM, David Marchand wrote: As I had reported in rc2, the lcore variables allocation have a noticeable impact on applications consuming DPDK, even when such applications does not use DPDK, or use features associated to some lcore variables. While the amount has been reduced in a rush before rc2, there are still cases when the increased memory footprint is noticed like in scaling tests. See https://bugs.launchpad.net/ubuntu/+source/dpdk/+bug/2090931 lcore variable allocations in constructor is a bad idea, as the application consuming DPDK has no control over such allocation: linking some code does not mean that all of it will be used at runtime. The general question on whether lcore variables in constructor should be forbidden, is left to a later discussion. For now, this series only focus on fixing subsystems using lcore variables so that those allocations are deferred either in rte_eal_init() or in the path that does require such lcore variables. An idle question: would this have any consequences in use case of eal init -> eal cleanup -> eal init with different arguments? Hum, interesting question. I would say that initialising lcore variables in constructors means that this usecase is broken, since lcore variables are freed in eal_lcore_var_cleanup(). After rte_eal_cleanup() is called, no DPDK calls may be made. So, with the current API, there is no such use case to break.
[PATCH] drivers/common: avoid truncation of constant value
This issue was flagged by MSVC warning below:
drivers\common\idpf\base/virtchnl2.h(269): warning C4309:
'initializing': truncation of constant value
The problem is that 64-bit numbers are initialized in an enum.
The C11 standard states: The expression that defines the value of an
enumeration constant shall be an integer constant expression that
has a value representable as an int. [Section 6.7.2.2]
As a result compilers end up truncating these numbers. MSVC and Clang
do that. Gcc has an extension that makes it use a 64-bit size for the
enum to avoid the truncation.
At the moment this truncation is only a real problem for
VIRTCHNL2_CAP_OEM, but clearly all these capability flags are
intended to be 64 bits long and therefore should better not be defined
in an enum.
The fix is to use "defines" for them, as the code was using about
7 months ago before enums were introduced here. In this patch the
flag names and values are being preserved, only the enum is being
removed.
Signed-off-by: Andre Muezerie
---
drivers/common/idpf/base/virtchnl2.h | 59 ++--
1 file changed, 30 insertions(+), 29 deletions(-)
diff --git a/drivers/common/idpf/base/virtchnl2.h
b/drivers/common/idpf/base/virtchnl2.h
index 3285a2b674..b6e28ade97 100644
--- a/drivers/common/idpf/base/virtchnl2.h
+++ b/drivers/common/idpf/base/virtchnl2.h
@@ -239,35 +239,36 @@ enum virtchnl2_cap_rsc {
VIRTCHNL2_CAP_RSC_IPV6_SCTP = BIT(3),
};
-/* Other capability flags */
-enum virtchnl2_cap_other {
- VIRTCHNL2_CAP_RDMA = BIT_ULL(0),
- VIRTCHNL2_CAP_SRIOV = BIT_ULL(1),
- VIRTCHNL2_CAP_MACFILTER = BIT_ULL(2),
- VIRTCHNL2_CAP_FLOW_DIRECTOR = BIT_ULL(3),
- VIRTCHNL2_CAP_SPLITQ_QSCHED = BIT_ULL(4),
- VIRTCHNL2_CAP_CRC = BIT_ULL(5),
- VIRTCHNL2_CAP_FLOW_STEER= BIT_ULL(6),
- VIRTCHNL2_CAP_WB_ON_ITR = BIT_ULL(7),
- VIRTCHNL2_CAP_PROMISC = BIT_ULL(8),
- VIRTCHNL2_CAP_LINK_SPEED= BIT_ULL(9),
- VIRTCHNL2_CAP_INLINE_IPSEC = BIT_ULL(10),
- VIRTCHNL2_CAP_LARGE_NUM_QUEUES = BIT_ULL(11),
- /* Require additional info */
- VIRTCHNL2_CAP_VLAN = BIT_ULL(12),
- VIRTCHNL2_CAP_PTP = BIT_ULL(13),
- VIRTCHNL2_CAP_ADV_RSS = BIT_ULL(15),
- VIRTCHNL2_CAP_FDIR = BIT_ULL(16),
- VIRTCHNL2_CAP_RX_FLEX_DESC = BIT_ULL(17),
- VIRTCHNL2_CAP_PTYPE = BIT_ULL(18),
- VIRTCHNL2_CAP_LOOPBACK = BIT_ULL(19),
- /* Enable miss completion types plus ability to detect a miss completion
-* if a reserved bit is set in a standard completion's tag.
-*/
- VIRTCHNL2_CAP_MISS_COMPL_TAG= BIT_ULL(20),
- /* This must be the last capability */
- VIRTCHNL2_CAP_OEM = BIT_ULL(63),
-};
+/* Other capability flags.
+ * Note: Not using an enum for these flags because these are 64-bit numbers,
+ * which might not fit in an int the enum maps to.
+ */
+#define VIRTCHNL2_CAP_RDMA RTE_BIT64(0)
+#define VIRTCHNL2_CAP_SRIOVRTE_BIT64(1)
+#define VIRTCHNL2_CAP_MACFILTERRTE_BIT64(2)
+#define VIRTCHNL2_CAP_FLOW_DIRECTORRTE_BIT64(3)
+#define VIRTCHNL2_CAP_SPLITQ_QSCHEDRTE_BIT64(4)
+#define VIRTCHNL2_CAP_CRC RTE_BIT64(5)
+#define VIRTCHNL2_CAP_FLOW_STEER RTE_BIT64(6)
+#define VIRTCHNL2_CAP_WB_ON_ITRRTE_BIT64(7)
+#define VIRTCHNL2_CAP_PROMISC RTE_BIT64(8)
+#define VIRTCHNL2_CAP_LINK_SPEED RTE_BIT64(9)
+#define VIRTCHNL2_CAP_INLINE_IPSEC RTE_BIT64(10)
+#define VIRTCHNL2_CAP_LARGE_NUM_QUEUES RTE_BIT64(11)
+/* Require additional info */
+#define VIRTCHNL2_CAP_VLAN RTE_BIT64(12)
+#define VIRTCHNL2_CAP_PTP RTE_BIT64(13)
+#define VIRTCHNL2_CAP_ADV_RSS RTE_BIT64(15)
+#define VIRTCHNL2_CAP_FDIR RTE_BIT64(16)
+#define VIRTCHNL2_CAP_RX_FLEX_DESC RTE_BIT64(17)
+#define VIRTCHNL2_CAP_PTYPERTE_BIT64(18)
+#define VIRTCHNL2_CAP_LOOPBACK RTE_BIT64(19)
+/* Enable miss completion types plus ability to detect a miss completion
+ * if a reserved bit is set in a standard completion's tag.
+ */
+#define VIRTCHNL2_CAP_MISS_COMPL_TAG RTE_BIT64(20)
+/* This must be the last capability */
+#define VIRTCHNL2_CAP_OEM RTE_BIT64(63)
/**
* enum virtchnl2_action_types - Available actions for sideband flow steering
--
2.47.0.vfs.0.3
Re: [PATCH v2 2/5] random: defer seeding to EAL init
On Tue, 17 Dec 2024 09:59:49 +0100
David Marchand wrote:
> The RNG is documented as being seeded as part of EAL init.
>
> Move the initialisation (seeding) helper out of a constructor and
> call it explicitly from rte_eal_init() as it was done before commit
> 3f002f069612 ("eal: replace libc-based random generation with LFSR").
>
> This also moves the unconditional lcore variable allocation out of a
> constructor.
>
> While at it, mark local symbol rand_state as static.
>
> Fixes: 29c39cd3d54d ("random: keep PRNG state in lcore variable")
> Cc: sta...@dpdk.org
>
> Signed-off-by: David Marchand
> Reviewed-by: Mattias Rönnblom
> Acked-by: Anatoly Burakov
Probably need to add a check to rte_random() so it crashes
if called before initialization, rather than returning an un-random
number which could be a hidden long term bug.
Re: [PATCH] net/virtio: fix Rx checksum calculation
On Wed, 18 Dec 2024 10:20:47 +0100
Olivier Matz wrote:
> Hi,
>
> On Wed, Dec 18, 2024 at 09:59:05AM +0100, Maxime Coquelin wrote:
> > Hi,
> >
> > On 12/18/24 08:34, Wangyunjian(wangyunjian,TongTu) wrote:
> > > > -Original Message-
> > > > From: Maxime Coquelin [mailto:maxime.coque...@redhat.com]
> > > > Sent: Tuesday, December 17, 2024 11:33 PM
> > > > To: dev@dpdk.org
> > > > Cc: Olivier Matz ; Maxime Gouin
> > > > ; Maxime Coquelin
> > > >
> > > > Subject: [PATCH] net/virtio: fix Rx checksum calculation
> > > >
> > > > From: Olivier Matz
> > > >
> > > > If hdr->csum_start is larger than packet length, the len argument passed
> > > > to rte_raw_cksum_mbuf() overflows and causes a segmentation fault.
> > > >
> > > > Ignore checksum computation in this case.
> > > >
> > > > CVE-2024-11614
> > > >
> > > > Fixes: ca7036b4af3a ("vhost: fix offload flags in Rx path")
> > > >
> > > > Signed-off-by: Maxime Gouin
> > > > Signed-off-by: Olivier Matz
> > > > Reviewed-by: Maxime Coquelin
> > > > ---
> > > > lib/vhost/virtio_net.c | 3 +++
> > > > 1 file changed, 3 insertions(+)
> > > >
> > > > diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
> > > > index d764d4bc6a..69901ab3b5 100644
> > > > --- a/lib/vhost/virtio_net.c
> > > > +++ b/lib/vhost/virtio_net.c
> > > > @@ -2823,6 +2823,9 @@ vhost_dequeue_offload(struct virtio_net *dev,
> > > > struct virtio_net_hdr *hdr,
> > > > */
> > > > uint16_t csum = 0, off;
> > > >
> > > > + if (hdr->csum_start >= rte_pktmbuf_pkt_len(m))
> > > > + return;
> > > > +
> > >
> > > The hdr->csum_start does two successive reads from user space to read
> > > a variable length data structure. The result overflow if the data
> > > structure
> > > changes between the two reads.
> > >
> > > We can prevent double fetch issue by using the temporary variable
> > > csum_start.
>
> This is an interesting remark, thanks!
>
> However, in practical, I'd say that the hdr->csum_start is fetched in a
> register
> only once if using optimized compilation, because the compiler has no reason
> to
> think that hdr->csum_start can be modified.
>
> Olivier
True, but security never depend on optimization.
Needs a fetch and compiler barrier to be truly safe against compilers.
Re: [PATCH] net/virtio: fix Rx checksum calculation
Hi,
On Wed, Dec 18, 2024 at 09:59:05AM +0100, Maxime Coquelin wrote:
> Hi,
>
> On 12/18/24 08:34, Wangyunjian(wangyunjian,TongTu) wrote:
> > > -Original Message-
> > > From: Maxime Coquelin [mailto:maxime.coque...@redhat.com]
> > > Sent: Tuesday, December 17, 2024 11:33 PM
> > > To: dev@dpdk.org
> > > Cc: Olivier Matz ; Maxime Gouin
> > > ; Maxime Coquelin
> > >
> > > Subject: [PATCH] net/virtio: fix Rx checksum calculation
> > >
> > > From: Olivier Matz
> > >
> > > If hdr->csum_start is larger than packet length, the len argument passed
> > > to rte_raw_cksum_mbuf() overflows and causes a segmentation fault.
> > >
> > > Ignore checksum computation in this case.
> > >
> > > CVE-2024-11614
> > >
> > > Fixes: ca7036b4af3a ("vhost: fix offload flags in Rx path")
> > >
> > > Signed-off-by: Maxime Gouin
> > > Signed-off-by: Olivier Matz
> > > Reviewed-by: Maxime Coquelin
> > > ---
> > > lib/vhost/virtio_net.c | 3 +++
> > > 1 file changed, 3 insertions(+)
> > >
> > > diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
> > > index d764d4bc6a..69901ab3b5 100644
> > > --- a/lib/vhost/virtio_net.c
> > > +++ b/lib/vhost/virtio_net.c
> > > @@ -2823,6 +2823,9 @@ vhost_dequeue_offload(struct virtio_net *dev,
> > > struct virtio_net_hdr *hdr,
> > >*/
> > > uint16_t csum = 0, off;
> > >
> > > + if (hdr->csum_start >= rte_pktmbuf_pkt_len(m))
> > > + return;
> > > +
> >
> > The hdr->csum_start does two successive reads from user space to read
> > a variable length data structure. The result overflow if the data structure
> > changes between the two reads.
> >
> > We can prevent double fetch issue by using the temporary variable
> > csum_start.
This is an interesting remark, thanks!
However, in practical, I'd say that the hdr->csum_start is fetched in a register
only once if using optimized compilation, because the compiler has no reason to
think that hdr->csum_start can be modified.
Olivier
>
> Right, that's a good catch! The exploitation od this issue seem
> difficult though.
>
> We may systematically copy the full header, as we only do it for ones
> not contiguous in host VA space.
>
> What do you think? Are you willing to contribute a fix?
>
> Thanks,
> Maxime
> >
> > Thanks,
> > Yunjian
> >
> > > if (rte_raw_cksum_mbuf(m, hdr->csum_start,
> > > rte_pktmbuf_pkt_len(m) -
> > > hdr->csum_start, &csum) <
> > > 0)
> > > return;
> > > --
> > > 2.47.0
> >
>
Re: [PATCH v2 4/5] power: reduce memory footprint of per-lcore state
On 12/17/2024 9:59 AM, David Marchand wrote:
Now that the per-lcore state was moved into a lcore variable,
there is no reason to align a per-lcore state on a cache line to avoid
false sharing.
Remove this alignment and save a few bytes.
Fixes: 130643319579 ("power: keep per-lcore state in lcore variable")
Signed-off-by: David Marchand
---
lib/power/rte_power_pmd_mgmt.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/power/rte_power_pmd_mgmt.c b/lib/power/rte_power_pmd_mgmt.c
index 369ce3c354..8ec925bd65 100644
--- a/lib/power/rte_power_pmd_mgmt.c
+++ b/lib/power/rte_power_pmd_mgmt.c
@@ -56,7 +56,7 @@ struct queue_list_entry {
const struct rte_eth_rxtx_callback *cb;
};
-struct __rte_cache_aligned pmd_core_cfg {
+struct pmd_core_cfg {
TAILQ_HEAD(queue_list_head, queue_list_entry) head;
/**< List of queues associated with this lcore */
size_t n_queues;
Acked-by: Anatoly Burakov
--
Thanks,
Anatoly
Re: [PATCH v2 3/5] power: defer lcore variable allocation
On 12/17/2024 9:59 AM, David Marchand wrote:
The lcore variable in this code unit is only used through
rte_power_ethdev_pmgmt_queue_*() public symbols.
Defer the unconditional lcore variable allocation in those symbols.
Fixes: 130643319579 ("power: keep per-lcore state in lcore variable")
Cc: sta...@dpdk.org
Signed-off-by: David Marchand
---
Acked-by: Anatoly Burakov
--
Thanks,
Anatoly
Re: [PATCH v2 5/5] eal/x86: defer power intrinsics variable allocation
On 12/17/2024 9:59 AM, David Marchand wrote:
The lcore variable in this code unit is only used through
rte_power_monitor*() public symbols.
Defer the unconditional lcore variable allocation in those symbols.
Fixes: 18b5049ab4fe ("eal/x86: keep power intrinsics state in lcore variable")
Cc: sta...@dpdk.org
Signed-off-by: David Marchand
---
Acked-by: Anatoly Burakov
--
Thanks,
Anatoly
Re: [PATCH v2] app/test: fix stack overflow in lpm6_perf_autotest
On Fri, Dec 13, 2024 at 09:15:40AM -0800, Stephen Hemminger wrote:
> On Fri, 13 Dec 2024 09:08:22 -0800
> Andre Muezerie wrote:
>
> > + struct rte_ipv6_addr *ip_batch =
> > + (struct rte_ipv6_addr *)rte_malloc("ip_batch",
> > + sizeof(struct rte_ipv6_addr) * NUM_IPS_ENTRIES, 0
>
> Cast is not needed here.
> If you are going to allocate an array, use calloc() or rte_calloc()
Thanks for the comments. I'll update the patch.
[PATCH v3] app/test: fix stack overflow in lpm6_perf_autotest
Test lpm6_perf_autotest was hitting a stack overflow on Windows
with both MSVC and Clang.
The fix is to move some of the data from the stack to the heap.
Signed-off-by: Andre Muezerie
---
app/test/test_lpm6_perf.c | 13 +++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/app/test/test_lpm6_perf.c b/app/test/test_lpm6_perf.c
index 1860a99ed6..59df0a958a 100644
--- a/app/test/test_lpm6_perf.c
+++ b/app/test/test_lpm6_perf.c
@@ -10,6 +10,7 @@
#include
#include
+#include
#include
#include
@@ -117,8 +118,13 @@ test_lpm6_perf(void)
total_time = 0;
count = 0;
- struct rte_ipv6_addr ip_batch[NUM_IPS_ENTRIES];
- int32_t next_hops[NUM_IPS_ENTRIES];
+ struct rte_ipv6_addr *ip_batch = rte_calloc("ip_batch",
+ NUM_IPS_ENTRIES, sizeof(struct rte_ipv6_addr), 0);
+ TEST_LPM_ASSERT(ip_batch != NULL);
+
+ int32_t *next_hops = rte_calloc("next_hops",
+ NUM_IPS_ENTRIES, sizeof(int32_t), 0);
+ TEST_LPM_ASSERT(next_hops != NULL);
for (i = 0; i < NUM_IPS_ENTRIES; i++)
ip_batch[i] = large_ips_table[i].ip;
@@ -153,6 +159,9 @@ test_lpm6_perf(void)
printf("Average LPM Delete: %g cycles\n",
(double)total_time / NUM_ROUTE_ENTRIES);
+ rte_free(next_hops);
+ rte_free(ip_batch);
+
rte_lpm6_delete_all(lpm);
rte_lpm6_free(lpm);
--
2.47.0.vfs.0.3
Re: [DPDK/core Bug 1605] lib/eal build failure with address and undefined behavior sanitizers
On Tue, 17 Dec 2024 20:39:12 + bugzi...@dpdk.org wrote: > https://bugs.dpdk.org/show_bug.cgi?id=1605 > > Bug ID: 1605 >Summary: lib/eal build failure with address and undefined > behavior sanitizers >Product: DPDK >Version: 25.03 > Hardware: All > OS: All > Status: UNCONFIRMED > Severity: normal > Priority: Normal > Component: core > Assignee: dev@dpdk.org > Reporter: alia...@nvidia.com > Target Milestone: --- > > $ meson --werror -Db_sanitize=address,undefined build && ninja -C build > [..] > /usr/include/x86_64-linux-gnu/bits/string_fortified.h:29:10: error: writing 4 > bytes into a region of size 0 [-Werror=stringop-overflow=] > [..] > > OS: Ubuntu 22.04 > Compiler: gcc 11.2.0 > Meson: 0.61.2 > Ninja-build: 1.10.1 > > Build passes with "-Db_sanitize=address". > Doing this found some other bugs: [182/3218] Compiling C object drivers/libtmp_rte_bus_fslmc.a.p/bus_fslmc_portal_dpaa2_hw_dpio.c.o ../drivers/bus/fslmc/portal/dpaa2_hw_dpio.c: In function ‘dpaa2_create_dpio_device’: ../drivers/bus/fslmc/portal/dpaa2_hw_dpio.c:607:1: warning: pointer ‘dpio_dev’ used after ‘rte_free’ [-Wuse-after-free] 607 | } | ^ ../drivers/bus/fslmc/portal/dpaa2_hw_dpio.c:600:17: note: call to ‘rte_free’ here 600 | rte_free(dpio_dev); | ^~
Re: [PATCH v2 2/5] random: defer seeding to EAL init
On 2024-12-18 17:35, Stephen Hemminger wrote:
On Tue, 17 Dec 2024 09:59:49 +0100
David Marchand wrote:
The RNG is documented as being seeded as part of EAL init.
Move the initialisation (seeding) helper out of a constructor and
call it explicitly from rte_eal_init() as it was done before commit
3f002f069612 ("eal: replace libc-based random generation with LFSR").
This also moves the unconditional lcore variable allocation out of a
constructor.
While at it, mark local symbol rand_state as static.
Fixes: 29c39cd3d54d ("random: keep PRNG state in lcore variable")
Cc: sta...@dpdk.org
Signed-off-by: David Marchand
Reviewed-by: Mattias Rönnblom
Acked-by: Anatoly Burakov
Probably need to add a check to rte_random() so it crashes
if called before initialization, rather than returning an un-random
number which could be a hidden long term bug.
If you do, do it either as a RTE_ASSERT() or an RTE_VERIFY() in the
(lcore_id == LCORE_ID_ANY) path, since that is what will be taken.
Preferably, you should have as little as possible in rte_rand() fast
path, because this function is used in packet processing.
That said, the "unrandom" number will always be 0, so it shouldn't go
unnoticed for too long.
Re: [PATCH] net/virtio: fix Rx checksum calculation
Hi,
On 12/18/24 08:34, Wangyunjian(wangyunjian,TongTu) wrote:
-Original Message-
From: Maxime Coquelin [mailto:maxime.coque...@redhat.com]
Sent: Tuesday, December 17, 2024 11:33 PM
To: dev@dpdk.org
Cc: Olivier Matz ; Maxime Gouin
; Maxime Coquelin
Subject: [PATCH] net/virtio: fix Rx checksum calculation
From: Olivier Matz
If hdr->csum_start is larger than packet length, the len argument passed
to rte_raw_cksum_mbuf() overflows and causes a segmentation fault.
Ignore checksum computation in this case.
CVE-2024-11614
Fixes: ca7036b4af3a ("vhost: fix offload flags in Rx path")
Signed-off-by: Maxime Gouin
Signed-off-by: Olivier Matz
Reviewed-by: Maxime Coquelin
---
lib/vhost/virtio_net.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
index d764d4bc6a..69901ab3b5 100644
--- a/lib/vhost/virtio_net.c
+++ b/lib/vhost/virtio_net.c
@@ -2823,6 +2823,9 @@ vhost_dequeue_offload(struct virtio_net *dev,
struct virtio_net_hdr *hdr,
*/
uint16_t csum = 0, off;
+ if (hdr->csum_start >= rte_pktmbuf_pkt_len(m))
+ return;
+
The hdr->csum_start does two successive reads from user space to read
a variable length data structure. The result overflow if the data structure
changes between the two reads.
We can prevent double fetch issue by using the temporary variable csum_start.
Right, that's a good catch! The exploitation od this issue seem
difficult though.
We may systematically copy the full header, as we only do it for ones
not contiguous in host VA space.
What do you think? Are you willing to contribute a fix?
Thanks,
Maxime
Thanks,
Yunjian
if (rte_raw_cksum_mbuf(m, hdr->csum_start,
rte_pktmbuf_pkt_len(m) - hdr->csum_start,
&csum) <
0)
return;
--
2.47.0
Re: [PATCH v2 1/5] eal: check lcore variable handle
On 12/17/2024 9:59 AM, David Marchand wrote: Add an assert to double check the passed handle is not NULL, as it points at an initialisation/allocation issue prior to accessing this lcore variable. Signed-off-by: David Marchand --- lib/eal/include/rte_lcore_var.h | 2 ++ 1 file changed, 2 insertions(+) Acked-by: Anatoly Burakov -- Thanks, Anatoly
[DPDK/examples Bug 1606] flow_filtering/flow_filtering_mismatch_rule: Some mismatch rule packets match the rule
https://bugs.dpdk.org/show_bug.cgi?id=1606 Bug ID: 1606 Summary: flow_filtering/flow_filtering_mismatch_rule: Some mismatch rule packets match the rule Product: DPDK Version: 24.11 Hardware: x86 OS: All Status: UNCONFIRMED Severity: normal Priority: Normal Component: examples Assignee: dev@dpdk.org Reporter: linglix.c...@intel.com Target Milestone: --- Environment === DPDK version: a4f455560f version: 24.11-rc4 OS: Ubuntu24.04.1 LTS (Noble Numbat)/6.8.0-41-generic Compiler: gcc version 13.2.0 Hardware platform: Intel(R) Xeon(R) Gold 6139 CPU @ 2.30GHz NIC hardware: Ethernet Controller XXV710 for 25GbE SFP28 158b NIC firmware: FW: 9.50 0x8000f4e1 1.3682.0 Driver: i40e-2.26.8 Test Setup Steps to reproduce == 1. bind ports to dpdk ./usertools/dpdk-devbind.py -b vfio-pci :af:00.0 :af:00.1 2. build flow_filtering meson configure -Dexamples=flow_filtering x86_64-native-linuxapp-gcc ninja -C x86_64-native-linuxapp-gcc 3. launch flow_filtering x86_64-native-linuxapp-gcc/examples/dpdk-flow_filtering -l 1 -n 4 -a :af:00.0 -a :af:00.1 -- --non-template 4. Send packets which mismatches the defined rule from tester sendp([Ether(dst='3c:fd:fe:cf:33:74')/IP(src='0.0.0.0', dst='192.169.1.1')/Raw(load=b''),Ether(dst='3c:fd:fe:cf:33:74')/IP(src='0.0.0.0', dst='193.168.1.1')/Raw(load=b'')],iface="ens2f0",count=1,loop=0,inter=0.01) 5. Check the packets are not received by queue 1 Results: EAL: Detected CPU lcores: 72 EAL: Detected NUMA nodes: 2 EAL: Detected static linkage of DPDK EAL: Multi-process socket /var/run/dpdk/rte/mp_socket EAL: Selected IOVA mode 'VA' EAL: 1024 hugepages of size 2097152 reserved, but no mounted hugetlbfs found for that size EAL: VFIO support initialized EAL: Using IOMMU type 1 (Type 1) I40E_DRIVER: i40e_GLQF_reg_init(): i40e device :af:00.0 changed global register [0x002689a0]. original: 0x0021, new: 0x0029 :: warn: 2 ports detected, but we use only one: port 0 :: initializing port: 0 :: initializing port: 0 done Flow created!!: src=00:00:00:00:00:00 - dst=3C:FD:FE:CF:33:74 - queue=0x1 src=00:00:00:00:00:00 - dst=3C:FD:FE:CF:33:74 - queue=0x1 Expected Result: src=00:00:00:00:00:00 - dst=3C:FD:FE:CF:33:74 - queue=0x0 src=00:00:00:00:00:00 - dst=3C:FD:FE:CF:33:74 - queue=0x0 bad commit: commit 16158f34900075f2f30b879bf3708e54e07455f4 Author: Shani Peretz Date: Mon Sep 30 22:53:21 2024 +0300 examples/flow_filtering: introduce use cases snippets These code snippets demonstrate rule creation using template and non-template APIs. They include functions that enable developers to create rules. The purpose of providing these snippets is to allow developers to reuse them, thereby saving time and effort during the implementation of flow rules. The code snippets are categorized based on their usage and can be copied, paste and modified to suit any requirements. The snippets provided here are kept up to date and are being compiled along with the rest of the examples. There is a skeleton that demonstrates rule creation using both template and non template APIs. Developers can change the functions in the skeleton to the corresponding snippet functions with the appropriate suffix and create rules using the snippets themselves for easy testing. Each snippet has the same functions to implement the actions and patterns for the corresponding feature. Signed-off-by: Shani Peretz Acked-by: Stephen Hemminger -- You are receiving this mail because: You are the assignee for the bug.
[PATCH v4 04/15] net/zxdh: port tables unint implementations
delete port tables in host.
Signed-off-by: Junlong Wang
---
drivers/net/zxdh/zxdh_ethdev.c | 18 ++
drivers/net/zxdh/zxdh_msg.h| 1 +
drivers/net/zxdh/zxdh_np.c | 103 +
drivers/net/zxdh/zxdh_np.h | 9 +++
drivers/net/zxdh/zxdh_tables.c | 33 ++-
drivers/net/zxdh/zxdh_tables.h | 1 +
6 files changed, 164 insertions(+), 1 deletion(-)
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index ff44816384..717a1d2b0b 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -887,12 +887,30 @@ zxdh_np_uninit(struct rte_eth_dev *dev)
zxdh_np_dtb_data_res_free(hw);
}
+static int
+zxdh_tables_uninit(struct rte_eth_dev *dev)
+{
+ int ret;
+
+ ret = zxdh_port_attr_uninit(dev);
+ if (ret)
+ PMD_DRV_LOG(ERR, "zxdh_port_attr_uninit failed");
+
+ return ret;
+}
+
static int
zxdh_dev_close(struct rte_eth_dev *dev)
{
struct zxdh_hw *hw = dev->data->dev_private;
int ret = 0;
+ ret = zxdh_tables_uninit(dev);
+ if (ret != 0) {
+ PMD_DRV_LOG(ERR, "%s :tables uninit %s failed ", __func__,
dev->device->name);
+ return -1;
+ }
+
zxdh_intr_release(dev);
zxdh_np_uninit(dev);
zxdh_pci_reset(hw);
diff --git a/drivers/net/zxdh/zxdh_msg.h b/drivers/net/zxdh/zxdh_msg.h
index b7b17b8696..613ca71170 100644
--- a/drivers/net/zxdh/zxdh_msg.h
+++ b/drivers/net/zxdh/zxdh_msg.h
@@ -167,6 +167,7 @@ enum pciebar_layout_type {
enum zxdh_msg_type {
ZXDH_NULL = 0,
ZXDH_VF_PORT_INIT = 1,
+ ZXDH_VF_PORT_UNINIT = 2,
ZXDH_MSG_TYPE_END,
};
diff --git a/drivers/net/zxdh/zxdh_np.c b/drivers/net/zxdh/zxdh_np.c
index db536d96e3..99a7dc11b4 100644
--- a/drivers/net/zxdh/zxdh_np.c
+++ b/drivers/net/zxdh/zxdh_np.c
@@ -25,6 +25,7 @@ ZXDH_RISCV_DTB_MGR
*p_riscv_dtb_queue_mgr[ZXDH_DEV_CHANNEL_MAX];
ZXDH_TLB_MGR_T *g_p_dpp_tlb_mgr[ZXDH_DEV_CHANNEL_MAX];
ZXDH_REG_T g_dpp_reg_info[4];
ZXDH_DTB_TABLE_T g_dpp_dtb_table_info[4];
+ZXDH_SDT_TBL_DATA_T g_sdt_info[ZXDH_DEV_CHANNEL_MAX][ZXDH_DEV_SDT_ID_MAX];
#define ZXDH_SDT_MGR_PTR_GET()(&g_sdt_mgr)
#define ZXDH_SDT_SOFT_TBL_GET(id) (g_sdt_mgr.sdt_tbl_array[id])
@@ -1454,3 +1455,105 @@ zxdh_np_dtb_table_entry_write(uint32_t dev_id,
return rc;
}
+
+static uint32_t
+zxdh_np_sdt_tbl_data_get(uint32_t dev_id, uint32_t sdt_no, ZXDH_SDT_TBL_DATA_T
*p_sdt_data)
+{
+ uint32_t rc = 0;
+
+ p_sdt_data->data_high32 = g_sdt_info[dev_id][sdt_no].data_high32;
+ p_sdt_data->data_low32 = g_sdt_info[dev_id][sdt_no].data_low32;
+
+ return rc;
+}
+
+int
+zxdh_np_dtb_table_entry_delete(uint32_t dev_id,
+uint32_t queue_id,
+uint32_t entrynum,
+ZXDH_DTB_USER_ENTRY_T *delete_entries)
+{
+ ZXDH_SDT_TBL_DATA_T sdt_tbl = {0};
+ ZXDH_DTB_USER_ENTRY_T *pentry = NULL;
+ ZXDH_DTB_ENTRY_T dtb_one_entry = {0};
+ uint8_t entry_cmd[ZXDH_DTB_TABLE_CMD_SIZE_BIT / 8] = {0};
+ uint8_t entry_data[ZXDH_ETCAM_WIDTH_MAX / 8] = {0};
+ uint8_t *p_data_buff = NULL;
+ uint8_t *p_data_buff_ex = NULL;
+ uint32_t tbl_type = 0;
+ uint32_t element_id = 0xff;
+ uint32_t one_dtb_len = 0;
+ uint32_t dtb_len = 0;
+ uint32_t entry_index;
+ uint32_t sdt_no;
+ uint32_t addr_offset;
+ uint32_t max_size;
+ uint32_t rc;
+
+ ZXDH_COMM_CHECK_POINT(delete_entries);
+
+ p_data_buff = rte_calloc(NULL, 1, ZXDH_DTB_TABLE_DATA_BUFF_SIZE, 0);
+ ZXDH_COMM_CHECK_POINT(p_data_buff);
+
+ p_data_buff_ex = rte_calloc(NULL, 1, ZXDH_DTB_TABLE_DATA_BUFF_SIZE, 0);
+ ZXDH_COMM_CHECK_POINT_MEMORY_FREE(p_data_buff_ex, p_data_buff);
+
+ dtb_one_entry.cmd = entry_cmd;
+ dtb_one_entry.data = entry_data;
+
+ max_size = (ZXDH_DTB_TABLE_DATA_BUFF_SIZE / 16) - 1;
+
+ for (entry_index = 0; entry_index < entrynum; entry_index++) {
+ pentry = delete_entries + entry_index;
+
+ sdt_no = pentry->sdt_no;
+ rc = zxdh_np_sdt_tbl_data_get(dev_id, sdt_no, &sdt_tbl);
+ switch (tbl_type) {
+ case ZXDH_SDT_TBLT_ERAM:
+ {
+ rc = zxdh_np_dtb_eram_one_entry(dev_id, sdt_no,
ZXDH_DTB_ITEM_DELETE,
+ pentry->p_entry_data, &one_dtb_len,
&dtb_one_entry);
+ break;
+ }
+ default:
+ {
+ PMD_DRV_LOG(ERR, "SDT table_type[ %d ] is invalid!",
tbl_type);
+ rte_free(p_data_buff);
+ rte_free(p_data_buff_ex);
+ return 1;
+ }
+ }
+
+ addr_offset = dtb_len * ZXDH_DTB_LEN_POS_SETP;
+ dtb_len += one_dtb_len;
+ if (dtb_len > max_s
[PATCH v4 08/15] net/zxdh: provided dev simple rx implementations
provided dev simple rx implementations.
Signed-off-by: Junlong Wang
---
doc/guides/nics/features/zxdh.ini | 1 +
doc/guides/nics/zxdh.rst | 1 +
drivers/net/zxdh/zxdh_ethdev.c| 1 +
drivers/net/zxdh/zxdh_rxtx.c | 313 ++
drivers/net/zxdh/zxdh_rxtx.h | 2 +
5 files changed, 318 insertions(+)
diff --git a/doc/guides/nics/features/zxdh.ini
b/doc/guides/nics/features/zxdh.ini
index 7b72be5f25..bb44e93fad 100644
--- a/doc/guides/nics/features/zxdh.ini
+++ b/doc/guides/nics/features/zxdh.ini
@@ -9,3 +9,4 @@ x86-64 = Y
ARMv8= Y
SR-IOV = Y
Multiprocess aware = Y
+Scattered Rx = Y
diff --git a/doc/guides/nics/zxdh.rst b/doc/guides/nics/zxdh.rst
index eb970a888f..f42db9c1f1 100644
--- a/doc/guides/nics/zxdh.rst
+++ b/doc/guides/nics/zxdh.rst
@@ -20,6 +20,7 @@ Features of the ZXDH PMD are:
- Multi arch support: x86_64, ARMv8.
- Multiple queues for TX and RX
- SR-IOV VF
+- Scattered and gather for TX and RX
Driver compilation and testing
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index b6057edeaf..0d63129d8d 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -967,6 +967,7 @@ zxdh_set_rxtx_funcs(struct rte_eth_dev *eth_dev)
}
eth_dev->tx_pkt_prepare = zxdh_xmit_pkts_prepare;
eth_dev->tx_pkt_burst = &zxdh_xmit_pkts_packed;
+ eth_dev->rx_pkt_burst = &zxdh_recv_pkts_packed;
return 0;
}
diff --git a/drivers/net/zxdh/zxdh_rxtx.c b/drivers/net/zxdh/zxdh_rxtx.c
index 10034a0e98..06290d48bb 100644
--- a/drivers/net/zxdh/zxdh_rxtx.c
+++ b/drivers/net/zxdh/zxdh_rxtx.c
@@ -31,6 +31,93 @@
#define ZXDH_TX_MAX_SEGS 31
#define ZXDH_RX_MAX_SEGS 31
+uint32_t zxdh_outer_l2_type[16] = {
+ 0,
+ RTE_PTYPE_L2_ETHER,
+ RTE_PTYPE_L2_ETHER_TIMESYNC,
+ RTE_PTYPE_L2_ETHER_ARP,
+ RTE_PTYPE_L2_ETHER_LLDP,
+ RTE_PTYPE_L2_ETHER_NSH,
+ RTE_PTYPE_L2_ETHER_VLAN,
+ RTE_PTYPE_L2_ETHER_QINQ,
+ RTE_PTYPE_L2_ETHER_PPPOE,
+ RTE_PTYPE_L2_ETHER_FCOE,
+ RTE_PTYPE_L2_ETHER_MPLS,
+};
+
+uint32_t zxdh_outer_l3_type[16] = {
+ 0,
+ RTE_PTYPE_L3_IPV4,
+ RTE_PTYPE_L3_IPV4_EXT,
+ RTE_PTYPE_L3_IPV6,
+ RTE_PTYPE_L3_IPV4_EXT_UNKNOWN,
+ RTE_PTYPE_L3_IPV6_EXT,
+ RTE_PTYPE_L3_IPV6_EXT_UNKNOWN,
+};
+
+uint32_t zxdh_outer_l4_type[16] = {
+ 0,
+ RTE_PTYPE_L4_TCP,
+ RTE_PTYPE_L4_UDP,
+ RTE_PTYPE_L4_FRAG,
+ RTE_PTYPE_L4_SCTP,
+ RTE_PTYPE_L4_ICMP,
+ RTE_PTYPE_L4_NONFRAG,
+ RTE_PTYPE_L4_IGMP,
+};
+
+uint32_t zxdh_tunnel_type[16] = {
+ 0,
+ RTE_PTYPE_TUNNEL_IP,
+ RTE_PTYPE_TUNNEL_GRE,
+ RTE_PTYPE_TUNNEL_VXLAN,
+ RTE_PTYPE_TUNNEL_NVGRE,
+ RTE_PTYPE_TUNNEL_GENEVE,
+ RTE_PTYPE_TUNNEL_GRENAT,
+ RTE_PTYPE_TUNNEL_GTPC,
+ RTE_PTYPE_TUNNEL_GTPU,
+ RTE_PTYPE_TUNNEL_ESP,
+ RTE_PTYPE_TUNNEL_L2TP,
+ RTE_PTYPE_TUNNEL_VXLAN_GPE,
+ RTE_PTYPE_TUNNEL_MPLS_IN_GRE,
+ RTE_PTYPE_TUNNEL_MPLS_IN_UDP,
+};
+
+uint32_t zxdh_inner_l2_type[16] = {
+ 0,
+ RTE_PTYPE_INNER_L2_ETHER,
+ 0,
+ 0,
+ 0,
+ 0,
+ RTE_PTYPE_INNER_L2_ETHER_VLAN,
+ RTE_PTYPE_INNER_L2_ETHER_QINQ,
+ 0,
+ 0,
+ 0,
+};
+
+uint32_t zxdh_inner_l3_type[16] = {
+ 0,
+ RTE_PTYPE_INNER_L3_IPV4,
+ RTE_PTYPE_INNER_L3_IPV4_EXT,
+ RTE_PTYPE_INNER_L3_IPV6,
+ RTE_PTYPE_INNER_L3_IPV4_EXT_UNKNOWN,
+ RTE_PTYPE_INNER_L3_IPV6_EXT,
+ RTE_PTYPE_INNER_L3_IPV6_EXT_UNKNOWN,
+};
+
+uint32_t zxdh_inner_l4_type[16] = {
+ 0,
+ RTE_PTYPE_INNER_L4_TCP,
+ RTE_PTYPE_INNER_L4_UDP,
+ RTE_PTYPE_INNER_L4_FRAG,
+ RTE_PTYPE_INNER_L4_SCTP,
+ RTE_PTYPE_INNER_L4_ICMP,
+ 0,
+ 0,
+};
+
static void
zxdh_xmit_cleanup_inorder_packed(struct zxdh_virtqueue *vq, int32_t num)
{
@@ -394,3 +481,229 @@ uint16_t zxdh_xmit_pkts_prepare(void *tx_queue
__rte_unused, struct rte_mbuf **t
}
return nb_tx;
}
+
+static uint16_t zxdh_dequeue_burst_rx_packed(struct zxdh_virtqueue *vq,
+ struct rte_mbuf **rx_pkts,
+ uint32_t *len,
+ uint16_t num)
+{
+ struct zxdh_vring_packed_desc *desc = vq->vq_packed.ring.desc;
+ struct rte_mbuf *cookie = NULL;
+ uint16_t i, used_idx;
+ uint16_t id;
+
+ for (i = 0; i < num; i++) {
+ used_idx = vq->vq_used_cons_idx;
+ /**
+* desc_is_used has a load-acquire or rte_io_rmb inside
+* and wait for used desc in virtqueue.
+*/
+ if (!zxdh_desc_used(&desc[used_idx], vq))
+ return i;
+ len[i] = desc[used_idx].len;
+ id
[PATCH v4 09/15] net/zxdh: link info update, set link up/down
provided link info update, set link up /down,
and link intr.
Signed-off-by: Junlong Wang
---
doc/guides/nics/features/zxdh.ini | 2 +
doc/guides/nics/zxdh.rst | 3 +
drivers/net/zxdh/meson.build | 1 +
drivers/net/zxdh/zxdh_ethdev.c | 14 ++-
drivers/net/zxdh/zxdh_ethdev.h | 2 +
drivers/net/zxdh/zxdh_ethdev_ops.c | 166
drivers/net/zxdh/zxdh_ethdev_ops.h | 14 +++
drivers/net/zxdh/zxdh_msg.c| 57 ++
drivers/net/zxdh/zxdh_msg.h| 40 +++
drivers/net/zxdh/zxdh_np.c | 172 -
drivers/net/zxdh/zxdh_np.h | 20
drivers/net/zxdh/zxdh_tables.c | 15 +++
drivers/net/zxdh/zxdh_tables.h | 6 +-
13 files changed, 503 insertions(+), 9 deletions(-)
create mode 100644 drivers/net/zxdh/zxdh_ethdev_ops.c
create mode 100644 drivers/net/zxdh/zxdh_ethdev_ops.h
diff --git a/doc/guides/nics/features/zxdh.ini
b/doc/guides/nics/features/zxdh.ini
index bb44e93fad..7da3aaced1 100644
--- a/doc/guides/nics/features/zxdh.ini
+++ b/doc/guides/nics/features/zxdh.ini
@@ -10,3 +10,5 @@ ARMv8= Y
SR-IOV = Y
Multiprocess aware = Y
Scattered Rx = Y
+Link status = Y
+Link status event= Y
diff --git a/doc/guides/nics/zxdh.rst b/doc/guides/nics/zxdh.rst
index f42db9c1f1..fdbc3b3923 100644
--- a/doc/guides/nics/zxdh.rst
+++ b/doc/guides/nics/zxdh.rst
@@ -21,6 +21,9 @@ Features of the ZXDH PMD are:
- Multiple queues for TX and RX
- SR-IOV VF
- Scattered and gather for TX and RX
+- Link Auto-negotiation
+- Link state information
+- Set Link down or up
Driver compilation and testing
diff --git a/drivers/net/zxdh/meson.build b/drivers/net/zxdh/meson.build
index 20b2cf484a..48f8f5e1ee 100644
--- a/drivers/net/zxdh/meson.build
+++ b/drivers/net/zxdh/meson.build
@@ -22,4 +22,5 @@ sources = files(
'zxdh_np.c',
'zxdh_tables.c',
'zxdh_rxtx.c',
+'zxdh_ethdev_ops.c',
)
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index 0d63129d8d..d3876ec9b3 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -16,6 +16,7 @@
#include "zxdh_np.h"
#include "zxdh_tables.h"
#include "zxdh_rxtx.h"
+#include "zxdh_ethdev_ops.h"
struct zxdh_hw_internal zxdh_hw_internal[RTE_MAX_ETHPORTS];
struct zxdh_shared_data *zxdh_shared_data;
@@ -105,12 +106,18 @@ static void
zxdh_devconf_intr_handler(void *param)
{
struct rte_eth_dev *dev = param;
+ struct zxdh_hw *hw = dev->data->dev_private;
+
+ uint8_t isr = zxdh_pci_isr(hw);
if (zxdh_intr_unmask(dev) < 0)
PMD_DRV_LOG(ERR, "interrupt enable failed");
+ if (isr & ZXDH_PCI_ISR_CONFIG) {
+ if (zxdh_dev_link_update(dev, 0) == 0)
+ rte_eth_dev_callback_process(dev,
RTE_ETH_EVENT_INTR_LSC, NULL);
+ }
}
-
/* Interrupt handler triggered by NIC for handling specific interrupt. */
static void
zxdh_fromriscv_intr_handler(void *param)
@@ -1006,6 +1013,8 @@ zxdh_dev_start(struct rte_eth_dev *dev)
vq = hw->vqs[logic_qidx];
zxdh_queue_notify(vq);
}
+ zxdh_dev_set_link_up(dev);
+
return 0;
}
@@ -1020,6 +1029,9 @@ static const struct eth_dev_ops zxdh_eth_dev_ops = {
.tx_queue_setup = zxdh_dev_tx_queue_setup,
.rx_queue_intr_enable= zxdh_dev_rx_queue_intr_enable,
.rx_queue_intr_disable = zxdh_dev_rx_queue_intr_disable,
+ .link_update = zxdh_dev_link_update,
+ .dev_set_link_up = zxdh_dev_set_link_up,
+ .dev_set_link_down = zxdh_dev_set_link_down,
};
static int32_t
diff --git a/drivers/net/zxdh/zxdh_ethdev.h b/drivers/net/zxdh/zxdh_ethdev.h
index b1f398b28e..c0b719062c 100644
--- a/drivers/net/zxdh/zxdh_ethdev.h
+++ b/drivers/net/zxdh/zxdh_ethdev.h
@@ -72,6 +72,7 @@ struct zxdh_hw {
uint64_t guest_features;
uint32_t max_queue_pairs;
uint32_t speed;
+ uint32_t speed_mode;
uint32_t notify_off_multiplier;
uint16_t *notify_base;
uint16_t pcie_id;
@@ -93,6 +94,7 @@ struct zxdh_hw {
uint8_t panel_id;
uint8_t has_tx_offload;
uint8_t has_rx_offload;
+ uint8_t admin_status;
};
struct zxdh_dtb_shared_data {
diff --git a/drivers/net/zxdh/zxdh_ethdev_ops.c
b/drivers/net/zxdh/zxdh_ethdev_ops.c
new file mode 100644
index 00..5a0af98cc0
--- /dev/null
+++ b/drivers/net/zxdh/zxdh_ethdev_ops.c
@@ -0,0 +1,166 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2024 ZTE Corporation
+ */
+
+#include "zxdh_ethdev.h"
+#include "zxdh_pci.h"
+#include "zxdh_msg.h"
+#include "zxdh_ethdev_ops.h"
+#include "zxdh_tables.h"
+#include "zxdh_logs.h"
+
+static int32_t zxdh_config_port_status(struct rte_eth_dev *dev, uint16_t
link_status)
+{
+ struct zxdh_hw *hw = dev->data->dev_priva
[PATCH v4 02/15] net/zxdh: zxdh np uninit implementation
(np)network processor release resources in host.
Signed-off-by: Junlong Wang
---
drivers/net/zxdh/zxdh_ethdev.c | 48
drivers/net/zxdh/zxdh_np.c | 470 +
drivers/net/zxdh/zxdh_np.h | 107
3 files changed, 625 insertions(+)
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index b8f4415e00..4e114d95da 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -841,6 +841,51 @@ zxdh_dev_configure(struct rte_eth_dev *dev)
return ret;
}
+static void
+zxdh_np_dtb_data_res_free(struct zxdh_hw *hw)
+{
+ struct rte_eth_dev *dev = hw->eth_dev;
+ int ret;
+ int i;
+
+ if (g_dtb_data.init_done && g_dtb_data.bind_device == dev) {
+ ret = zxdh_np_online_uninit(0, dev->data->name,
g_dtb_data.queueid);
+ if (ret)
+ PMD_DRV_LOG(ERR, "%s dpp_np_online_uninstall failed",
dev->data->name);
+
+ if (g_dtb_data.dtb_table_conf_mz)
+ rte_memzone_free(g_dtb_data.dtb_table_conf_mz);
+
+ if (g_dtb_data.dtb_table_dump_mz) {
+ rte_memzone_free(g_dtb_data.dtb_table_dump_mz);
+ g_dtb_data.dtb_table_dump_mz = NULL;
+ }
+
+ for (i = 0; i < ZXDH_MAX_BASE_DTB_TABLE_COUNT; i++) {
+ if (g_dtb_data.dtb_table_bulk_dump_mz[i]) {
+
rte_memzone_free(g_dtb_data.dtb_table_bulk_dump_mz[i]);
+ g_dtb_data.dtb_table_bulk_dump_mz[i] = NULL;
+ }
+ }
+ g_dtb_data.init_done = 0;
+ g_dtb_data.bind_device = NULL;
+ }
+ if (zxdh_shared_data != NULL)
+ zxdh_shared_data->np_init_done = 0;
+}
+
+static void
+zxdh_np_uninit(struct rte_eth_dev *dev)
+{
+ struct zxdh_hw *hw = dev->data->dev_private;
+
+ if (!g_dtb_data.init_done && !g_dtb_data.dev_refcnt)
+ return;
+
+ if (--g_dtb_data.dev_refcnt == 0)
+ zxdh_np_dtb_data_res_free(hw);
+}
+
static int
zxdh_dev_close(struct rte_eth_dev *dev)
{
@@ -848,6 +893,7 @@ zxdh_dev_close(struct rte_eth_dev *dev)
int ret = 0;
zxdh_intr_release(dev);
+ zxdh_np_uninit(dev);
zxdh_pci_reset(hw);
zxdh_dev_free_mbufs(dev);
@@ -1010,6 +1056,7 @@ zxdh_np_dtb_res_init(struct rte_eth_dev *dev)
return 0;
free_res:
+ zxdh_np_dtb_data_res_free(hw);
rte_free(dpp_ctrl);
return ret;
}
@@ -1177,6 +1224,7 @@ zxdh_eth_dev_init(struct rte_eth_dev *eth_dev)
err_zxdh_init:
zxdh_intr_release(eth_dev);
+ zxdh_np_uninit(eth_dev);
zxdh_bar_msg_chan_exit();
rte_free(eth_dev->data->mac_addrs);
eth_dev->data->mac_addrs = NULL;
diff --git a/drivers/net/zxdh/zxdh_np.c b/drivers/net/zxdh/zxdh_np.c
index e44d7ff501..28728b0c68 100644
--- a/drivers/net/zxdh/zxdh_np.c
+++ b/drivers/net/zxdh/zxdh_np.c
@@ -18,10 +18,21 @@ static ZXDH_DEV_MGR_T g_dev_mgr;
static ZXDH_SDT_MGR_T g_sdt_mgr;
ZXDH_PPU_CLS_BITMAP_T g_ppu_cls_bit_map[ZXDH_DEV_CHANNEL_MAX];
ZXDH_DTB_MGR_T *p_dpp_dtb_mgr[ZXDH_DEV_CHANNEL_MAX];
+ZXDH_RISCV_DTB_MGR *p_riscv_dtb_queue_mgr[ZXDH_DEV_CHANNEL_MAX];
+ZXDH_TLB_MGR_T *g_p_dpp_tlb_mgr[ZXDH_DEV_CHANNEL_MAX];
+ZXDH_REG_T g_dpp_reg_info[4];
#define ZXDH_SDT_MGR_PTR_GET()(&g_sdt_mgr)
#define ZXDH_SDT_SOFT_TBL_GET(id) (g_sdt_mgr.sdt_tbl_array[id])
+#define ZXDH_COMM_MASK_BIT(_bitnum_)\
+ (0x1U << (_bitnum_))
+
+#define ZXDH_COMM_GET_BIT_MASK(_inttype_, _bitqnt_)\
+ ((_inttype_)(((_bitqnt_) < 32)))
+
+#define ZXDH_REG_DATA_MAX (128)
+
#define ZXDH_COMM_CHECK_DEV_POINT(dev_id, point)\
do {\
if (NULL == (point)) {\
@@ -338,3 +349,462 @@ zxdh_np_host_init(uint32_t dev_id,
return 0;
}
+
+static ZXDH_RISCV_DTB_MGR *
+zxdh_np_riscv_dtb_queue_mgr_get(uint32_t dev_id)
+{
+ if (dev_id >= ZXDH_DEV_CHANNEL_MAX)
+ return NULL;
+ else
+ return p_riscv_dtb_queue_mgr[dev_id];
+}
+
+static uint32_t
+zxdh_np_riscv_dtb_mgr_queue_info_delete(uint32_t dev_id, uint32_t queue_id)
+{
+ ZXDH_RISCV_DTB_MGR *p_riscv_dtb_mgr = NULL;
+
+ p_riscv_dtb_mgr = zxdh_np_riscv_dtb_queue_mgr_get(dev_id);
+ if (p_riscv_dtb_mgr == NULL)
+ return 1;
+
+ p_riscv_dtb_mgr->queue_alloc_count--;
+ p_riscv_dtb_mgr->queue_user_info[queue_id].alloc_flag = 0;
+ p_riscv_dtb_mgr->queue_user_info[queue_id].queue_id = 0xFF;
+ p_riscv_dtb_mgr->queue_user_info[queue_id].vport = 0;
+ memset(p_riscv_dtb_mgr->queue_user_info[queue_id].user_name, 0,
ZXDH_PORT_NAME_MAX);
+
+ return 0;
+}
+
+static uint32_t
+zxdh_np_dev_get_dev_type(uint32_t dev_id)
+{
+ ZXDH_DEV_MGR_T *p_dev_mgr = NULL;
+ ZXDH_DEV_CFG_T *p_dev_info = NULL;
+
+ p_dev_mgr = &g_dev_mgr;
+ p_dev_info = p_dev_mgr->p_dev_array[de
[PATCH v4 06/15] net/zxdh: dev start/stop ops implementations
dev start/stop implementations, start/stop the rx/tx queues.
Signed-off-by: Junlong Wang
---
doc/guides/nics/features/zxdh.ini | 2 +
doc/guides/nics/zxdh.rst | 2 +
drivers/net/zxdh/zxdh_ethdev.c| 61 +
drivers/net/zxdh/zxdh_pci.c | 24
drivers/net/zxdh/zxdh_pci.h | 1 +
drivers/net/zxdh/zxdh_queue.c | 91 +++
drivers/net/zxdh/zxdh_queue.h | 69 +++
drivers/net/zxdh/zxdh_rxtx.h | 14 ++---
8 files changed, 256 insertions(+), 8 deletions(-)
diff --git a/doc/guides/nics/features/zxdh.ini
b/doc/guides/nics/features/zxdh.ini
index 05c8091ed7..7b72be5f25 100644
--- a/doc/guides/nics/features/zxdh.ini
+++ b/doc/guides/nics/features/zxdh.ini
@@ -7,3 +7,5 @@
Linux= Y
x86-64 = Y
ARMv8= Y
+SR-IOV = Y
+Multiprocess aware = Y
diff --git a/doc/guides/nics/zxdh.rst b/doc/guides/nics/zxdh.rst
index 2144753d75..eb970a888f 100644
--- a/doc/guides/nics/zxdh.rst
+++ b/doc/guides/nics/zxdh.rst
@@ -18,6 +18,8 @@ Features
Features of the ZXDH PMD are:
- Multi arch support: x86_64, ARMv8.
+- Multiple queues for TX and RX
+- SR-IOV VF
Driver compilation and testing
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index 521d7ed433..59ee942bdd 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -899,12 +899,35 @@ zxdh_tables_uninit(struct rte_eth_dev *dev)
return ret;
}
+static int
+zxdh_dev_stop(struct rte_eth_dev *dev)
+{
+ int ret = 0;
+
+ if (dev->data->dev_started == 0)
+ return 0;
+
+ ret = zxdh_intr_disable(dev);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "intr disable failed");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
static int
zxdh_dev_close(struct rte_eth_dev *dev)
{
struct zxdh_hw *hw = dev->data->dev_private;
int ret = 0;
+ ret = zxdh_dev_stop(dev);
+ if (ret != 0) {
+ PMD_DRV_LOG(ERR, " stop port %s failed.", dev->device->name);
+ return -1;
+ }
+
ret = zxdh_tables_uninit(dev);
if (ret != 0) {
PMD_DRV_LOG(ERR, "%s :tables uninit %s failed ", __func__,
dev->device->name);
@@ -928,9 +951,47 @@ zxdh_dev_close(struct rte_eth_dev *dev)
return ret;
}
+static int
+zxdh_dev_start(struct rte_eth_dev *dev)
+{
+ struct zxdh_hw *hw = dev->data->dev_private;
+ struct zxdh_virtqueue *vq;
+ int32_t ret;
+ uint16_t logic_qidx;
+ uint16_t i;
+
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ logic_qidx = 2 * i + ZXDH_RQ_QUEUE_IDX;
+ ret = zxdh_dev_rx_queue_setup_finish(dev, logic_qidx);
+ if (ret < 0)
+ return ret;
+ }
+ ret = zxdh_intr_enable(dev);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "interrupt enable failed");
+ return -EINVAL;
+ }
+
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ logic_qidx = 2 * i + ZXDH_RQ_QUEUE_IDX;
+ vq = hw->vqs[logic_qidx];
+ /* Flush the old packets */
+ zxdh_queue_rxvq_flush(vq);
+ zxdh_queue_notify(vq);
+ }
+ for (i = 0; i < dev->data->nb_tx_queues; i++) {
+ logic_qidx = 2 * i + ZXDH_TQ_QUEUE_IDX;
+ vq = hw->vqs[logic_qidx];
+ zxdh_queue_notify(vq);
+ }
+ return 0;
+}
+
/* dev_ops for zxdh, bare necessities for basic operation */
static const struct eth_dev_ops zxdh_eth_dev_ops = {
.dev_configure = zxdh_dev_configure,
+ .dev_start = zxdh_dev_start,
+ .dev_stop= zxdh_dev_stop,
.dev_close = zxdh_dev_close,
.dev_infos_get = zxdh_dev_infos_get,
.rx_queue_setup = zxdh_dev_rx_queue_setup,
diff --git a/drivers/net/zxdh/zxdh_pci.c b/drivers/net/zxdh/zxdh_pci.c
index 250e67d560..83164a5c79 100644
--- a/drivers/net/zxdh/zxdh_pci.c
+++ b/drivers/net/zxdh/zxdh_pci.c
@@ -202,6 +202,29 @@ zxdh_del_queue(struct zxdh_hw *hw, struct zxdh_virtqueue
*vq)
rte_write16(0, &hw->common_cfg->queue_enable);
}
+static void
+zxdh_notify_queue(struct zxdh_hw *hw, struct zxdh_virtqueue *vq)
+{
+ uint32_t notify_data = 0;
+
+ if (!zxdh_pci_with_feature(hw, ZXDH_F_NOTIFICATION_DATA)) {
+ rte_write16(vq->vq_queue_index, vq->notify_addr);
+ return;
+ }
+
+ if (zxdh_pci_with_feature(hw, ZXDH_F_RING_PACKED)) {
+ notify_data = ((uint32_t)(!!(vq->vq_packed.cached_flags &
+
ZXDH_VRING_PACKED_DESC_F_AVAIL)) << 31) |
+ ((uint32_t)vq->vq_avail_idx <<
16) |
+
[PATCH v4 05/15] net/zxdh: rx/tx queue setup and intr enable
rx/tx queue setup and intr enable implementations.
Signed-off-by: Junlong Wang
---
drivers/net/zxdh/zxdh_ethdev.c | 4 +
drivers/net/zxdh/zxdh_queue.c | 149 +
drivers/net/zxdh/zxdh_queue.h | 33
3 files changed, 186 insertions(+)
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index 717a1d2b0b..521d7ed433 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -933,6 +933,10 @@ static const struct eth_dev_ops zxdh_eth_dev_ops = {
.dev_configure = zxdh_dev_configure,
.dev_close = zxdh_dev_close,
.dev_infos_get = zxdh_dev_infos_get,
+ .rx_queue_setup = zxdh_dev_rx_queue_setup,
+ .tx_queue_setup = zxdh_dev_tx_queue_setup,
+ .rx_queue_intr_enable= zxdh_dev_rx_queue_intr_enable,
+ .rx_queue_intr_disable = zxdh_dev_rx_queue_intr_disable,
};
static int32_t
diff --git a/drivers/net/zxdh/zxdh_queue.c b/drivers/net/zxdh/zxdh_queue.c
index b4ef90ea36..af21f046ad 100644
--- a/drivers/net/zxdh/zxdh_queue.c
+++ b/drivers/net/zxdh/zxdh_queue.c
@@ -12,6 +12,11 @@
#include "zxdh_common.h"
#include "zxdh_msg.h"
+#define ZXDH_MBUF_MIN_SIZE sizeof(struct zxdh_net_hdr_dl)
+#define ZXDH_MBUF_SIZE_4K 4096
+#define ZXDH_RX_FREE_THRESH 32
+#define ZXDH_TX_FREE_THRESH 32
+
struct rte_mbuf *
zxdh_queue_detach_unused(struct zxdh_virtqueue *vq)
{
@@ -125,3 +130,147 @@ zxdh_free_queues(struct rte_eth_dev *dev)
return 0;
}
+
+static int
+zxdh_check_mempool(struct rte_mempool *mp, uint16_t offset, uint16_t
min_length)
+{
+ uint16_t data_room_size;
+
+ if (mp == NULL)
+ return -EINVAL;
+ data_room_size = rte_pktmbuf_data_room_size(mp);
+ if (data_room_size < offset + min_length) {
+ PMD_RX_LOG(ERR,
+ "%s mbuf_data_room_size %u < %u (%u + %u)",
+ mp->name, data_room_size,
+ offset + min_length, offset, min_length);
+ return -EINVAL;
+ }
+ return 0;
+}
+
+int32_t
+zxdh_dev_rx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t queue_idx,
+ uint16_t nb_desc,
+ uint32_t socket_id __rte_unused,
+ const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mp)
+{
+ struct zxdh_hw *hw = dev->data->dev_private;
+ uint16_t vtpci_logic_qidx = 2 * queue_idx + ZXDH_RQ_QUEUE_IDX;
+ struct zxdh_virtqueue *vq = hw->vqs[vtpci_logic_qidx];
+ int32_t ret = 0;
+
+ if (rx_conf->rx_deferred_start) {
+ PMD_RX_LOG(ERR, "Rx deferred start is not supported");
+ return -EINVAL;
+ }
+ uint16_t rx_free_thresh = rx_conf->rx_free_thresh;
+
+ if (rx_free_thresh == 0)
+ rx_free_thresh = RTE_MIN(vq->vq_nentries / 4,
ZXDH_RX_FREE_THRESH);
+
+ /* rx_free_thresh must be multiples of four. */
+ if (rx_free_thresh & 0x3) {
+ PMD_RX_LOG(ERR, "(rx_free_thresh=%u port=%u queue=%u)",
+ rx_free_thresh, dev->data->port_id, queue_idx);
+ return -EINVAL;
+ }
+ /* rx_free_thresh must be less than the number of RX entries */
+ if (rx_free_thresh >= vq->vq_nentries) {
+ PMD_RX_LOG(ERR, "RX entries (%u). (rx_free_thresh=%u port=%u
queue=%u)",
+ vq->vq_nentries, rx_free_thresh, dev->data->port_id,
queue_idx);
+ return -EINVAL;
+ }
+ vq->vq_free_thresh = rx_free_thresh;
+ nb_desc = ZXDH_QUEUE_DEPTH;
+
+ vq->vq_free_cnt = RTE_MIN(vq->vq_free_cnt, nb_desc);
+ struct zxdh_virtnet_rx *rxvq = &vq->rxq;
+
+ rxvq->queue_id = vtpci_logic_qidx;
+
+ int mbuf_min_size = ZXDH_MBUF_MIN_SIZE;
+
+ if (rx_conf->offloads & RTE_ETH_RX_OFFLOAD_TCP_LRO)
+ mbuf_min_size = ZXDH_MBUF_SIZE_4K;
+
+ ret = zxdh_check_mempool(mp, RTE_PKTMBUF_HEADROOM, mbuf_min_size);
+ if (ret != 0) {
+ PMD_RX_LOG(ERR,
+ "rxq setup but mpool size too small(<%d) failed",
mbuf_min_size);
+ return -EINVAL;
+ }
+ rxvq->mpool = mp;
+ if (queue_idx < dev->data->nb_rx_queues)
+ dev->data->rx_queues[queue_idx] = rxvq;
+
+ return 0;
+}
+
+int32_t
+zxdh_dev_tx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t queue_idx,
+ uint16_t nb_desc,
+ uint32_t socket_id __rte_unused,
+ const struct rte_eth_txconf *tx_conf)
+{
+ uint16_t vtpci_logic_qidx = 2 * queue_idx + ZXDH_TQ_QUEUE_IDX;
+ struct zxdh_hw *hw = dev->data->dev_private;
+ struct zxdh_virtqueue *vq = hw->vqs[vtpc
[PATCH v4 00/15] net/zxdh: updated net zxdh driver
V4: - resolved ci compile issues. V3: - use rte_zmalloc and rte_calloc to avoid memset. - remove unnecessary initialization, which first usage will set. - adjust some function which is always return 0, changed to void and skip the ASSERTION later. - resolved some WARNING:MACRO_ARG_UNUSED issues. - resolved some other issues. V2: - resolve code style and github-robot build issue. V1: - updated net zxdh driver provided insert/delete/get table code funcs. provided link/mac/vlan/promiscuous/rss/mtu ops. Junlong Wang (15): net/zxdh: zxdh np init implementation net/zxdh: zxdh np uninit implementation net/zxdh: port tables init implementations net/zxdh: port tables unint implementations net/zxdh: rx/tx queue setup and intr enable net/zxdh: dev start/stop ops implementations net/zxdh: provided dev simple tx implementations net/zxdh: provided dev simple rx implementations net/zxdh: link info update, set link up/down net/zxdh: mac set/add/remove ops implementations net/zxdh: promisc/allmulti ops implementations net/zxdh: vlan filter/ offload ops implementations net/zxdh: rss hash config/update, reta update/get net/zxdh: basic stats ops implementations net/zxdh: mtu update ops implementations doc/guides/nics/features/zxdh.ini | 18 + doc/guides/nics/zxdh.rst | 17 + drivers/net/zxdh/meson.build |4 + drivers/net/zxdh/zxdh_common.c | 24 + drivers/net/zxdh/zxdh_common.h |1 + drivers/net/zxdh/zxdh_ethdev.c | 575 +++- drivers/net/zxdh/zxdh_ethdev.h | 40 + drivers/net/zxdh/zxdh_ethdev_ops.c | 1595 + drivers/net/zxdh/zxdh_ethdev_ops.h | 80 ++ drivers/net/zxdh/zxdh_msg.c| 166 +++ drivers/net/zxdh/zxdh_msg.h| 232 drivers/net/zxdh/zxdh_np.c | 2060 drivers/net/zxdh/zxdh_np.h | 579 drivers/net/zxdh/zxdh_pci.c| 26 +- drivers/net/zxdh/zxdh_pci.h|9 +- drivers/net/zxdh/zxdh_queue.c | 242 +++- drivers/net/zxdh/zxdh_queue.h | 144 +- drivers/net/zxdh/zxdh_rxtx.c | 804 +++ drivers/net/zxdh/zxdh_rxtx.h | 20 +- drivers/net/zxdh/zxdh_tables.c | 794 +++ drivers/net/zxdh/zxdh_tables.h | 231 21 files changed, 7615 insertions(+), 46 deletions(-) create mode 100644 drivers/net/zxdh/zxdh_ethdev_ops.c create mode 100644 drivers/net/zxdh/zxdh_ethdev_ops.h create mode 100644 drivers/net/zxdh/zxdh_np.c create mode 100644 drivers/net/zxdh/zxdh_np.h create mode 100644 drivers/net/zxdh/zxdh_rxtx.c create mode 100644 drivers/net/zxdh/zxdh_tables.c create mode 100644 drivers/net/zxdh/zxdh_tables.h -- 2.27.0
[PATCH v4 11/15] net/zxdh: promisc/allmulti ops implementations
provided promiscuous/allmulticast ops.
Signed-off-by: Junlong Wang
---
doc/guides/nics/features/zxdh.ini | 2 +
doc/guides/nics/zxdh.rst | 2 +
drivers/net/zxdh/zxdh_ethdev.c | 21 ++-
drivers/net/zxdh/zxdh_ethdev.h | 2 +
drivers/net/zxdh/zxdh_ethdev_ops.c | 132 +
drivers/net/zxdh/zxdh_ethdev_ops.h | 4 +
drivers/net/zxdh/zxdh_msg.h| 10 ++
drivers/net/zxdh/zxdh_tables.c | 223 +
drivers/net/zxdh/zxdh_tables.h | 22 +++
9 files changed, 417 insertions(+), 1 deletion(-)
diff --git a/doc/guides/nics/features/zxdh.ini
b/doc/guides/nics/features/zxdh.ini
index dc09fe3453..e9b237e102 100644
--- a/doc/guides/nics/features/zxdh.ini
+++ b/doc/guides/nics/features/zxdh.ini
@@ -14,3 +14,5 @@ Link status = Y
Link status event= Y
Unicast MAC filter = Y
Multicast MAC filter = Y
+Promiscuous mode = Y
+Allmulticast mode= Y
diff --git a/doc/guides/nics/zxdh.rst b/doc/guides/nics/zxdh.rst
index e0b0776aca..0399df1302 100644
--- a/doc/guides/nics/zxdh.rst
+++ b/doc/guides/nics/zxdh.rst
@@ -26,6 +26,8 @@ Features of the ZXDH PMD are:
- Set Link down or up
- Unicast MAC filter
- Multicast MAC filter
+- Promiscuous mode
+- Multicast mode
Driver compilation and testing
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index 85ada87cdc..1d64b877c1 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -901,8 +901,16 @@ zxdh_tables_uninit(struct rte_eth_dev *dev)
int ret;
ret = zxdh_port_attr_uninit(dev);
- if (ret)
+ if (ret) {
PMD_DRV_LOG(ERR, "zxdh_port_attr_uninit failed");
+ return ret;
+ }
+
+ ret = zxdh_promisc_table_uninit(dev);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "uninit promisc_table failed");
+ return ret;
+ }
return ret;
}
@@ -1055,6 +1063,10 @@ static const struct eth_dev_ops zxdh_eth_dev_ops = {
.mac_addr_add= zxdh_dev_mac_addr_add,
.mac_addr_remove = zxdh_dev_mac_addr_remove,
.mac_addr_set= zxdh_dev_mac_addr_set,
+ .promiscuous_enable = zxdh_dev_promiscuous_enable,
+ .promiscuous_disable = zxdh_dev_promiscuous_disable,
+ .allmulticast_enable = zxdh_dev_allmulticast_enable,
+ .allmulticast_disable= zxdh_dev_allmulticast_disable,
};
static int32_t
@@ -1306,6 +1318,13 @@ zxdh_tables_init(struct rte_eth_dev *dev)
PMD_DRV_LOG(ERR, " panel table init failed");
return ret;
}
+
+ ret = zxdh_promisc_table_init(dev);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "promisc_table_init failed");
+ return ret;
+ }
+
return ret;
}
diff --git a/drivers/net/zxdh/zxdh_ethdev.h b/drivers/net/zxdh/zxdh_ethdev.h
index 5b95cb1c2a..3cdac5de73 100644
--- a/drivers/net/zxdh/zxdh_ethdev.h
+++ b/drivers/net/zxdh/zxdh_ethdev.h
@@ -98,6 +98,8 @@ struct zxdh_hw {
uint8_t has_tx_offload;
uint8_t has_rx_offload;
uint8_t admin_status;
+ uint8_t promisc_status;
+ uint8_t allmulti_status;
};
struct zxdh_dtb_shared_data {
diff --git a/drivers/net/zxdh/zxdh_ethdev_ops.c
b/drivers/net/zxdh/zxdh_ethdev_ops.c
index 751f80e9b4..aed4e6410c 100644
--- a/drivers/net/zxdh/zxdh_ethdev_ops.c
+++ b/drivers/net/zxdh/zxdh_ethdev_ops.c
@@ -397,3 +397,135 @@ void zxdh_dev_mac_addr_remove(struct rte_eth_dev *dev
__rte_unused, uint32_t ind
}
memset(&dev->data->mac_addrs[index], 0, sizeof(struct rte_ether_addr));
}
+
+int zxdh_dev_promiscuous_enable(struct rte_eth_dev *dev)
+{
+ struct zxdh_hw *hw = dev->data->dev_private;
+ struct zxdh_msg_info msg_info = {0};
+ int16_t ret = 0;
+
+ if (hw->promisc_status == 0) {
+ if (hw->is_pf) {
+ ret = zxdh_dev_unicast_table_set(hw, hw->vport.vport,
true);
+ if (hw->allmulti_status == 0)
+ ret = zxdh_dev_multicast_table_set(hw,
hw->vport.vport, true);
+
+ } else {
+ struct zxdh_port_promisc_msg *promisc_msg =
&msg_info.data.port_promisc_msg;
+
+ zxdh_msg_head_build(hw, ZXDH_PORT_PROMISC_SET,
&msg_info);
+ promisc_msg->mode = ZXDH_PROMISC_MODE;
+ promisc_msg->value = true;
+ if (hw->allmulti_status == 0)
+ promisc_msg->mc_follow = true;
+
+ ret = zxdh_vf_send_msg_to_pf(dev, &msg_info,
sizeof(msg_info), NULL, 0);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "Failed to send msg: port 0x%x
msg type %d",
+ hw->vport.vport,
ZXDH_PROMISC_MODE);
+ return ret;
+
[PATCH v4 13/15] net/zxdh: rss hash config/update, reta update/get
provided rss hash config/update, reta update/get ops.
Signed-off-by: Junlong Wang
---
doc/guides/nics/features/zxdh.ini | 3 +
doc/guides/nics/zxdh.rst | 1 +
drivers/net/zxdh/zxdh_ethdev.c | 52
drivers/net/zxdh/zxdh_ethdev.h | 3 +
drivers/net/zxdh/zxdh_ethdev_ops.c | 410 +
drivers/net/zxdh/zxdh_ethdev_ops.h | 26 ++
drivers/net/zxdh/zxdh_msg.h| 22 ++
drivers/net/zxdh/zxdh_tables.c | 82 ++
drivers/net/zxdh/zxdh_tables.h | 7 +
9 files changed, 606 insertions(+)
diff --git a/doc/guides/nics/features/zxdh.ini
b/doc/guides/nics/features/zxdh.ini
index 6fb006c2da..415ca547d0 100644
--- a/doc/guides/nics/features/zxdh.ini
+++ b/doc/guides/nics/features/zxdh.ini
@@ -19,3 +19,6 @@ Allmulticast mode= Y
VLAN filter = Y
VLAN offload = Y
QinQ offload = Y
+RSS hash = Y
+RSS reta update = Y
+Inner RSS= Y
diff --git a/doc/guides/nics/zxdh.rst b/doc/guides/nics/zxdh.rst
index 3a7585d123..3cc6a1d348 100644
--- a/doc/guides/nics/zxdh.rst
+++ b/doc/guides/nics/zxdh.rst
@@ -31,6 +31,7 @@ Features of the ZXDH PMD are:
- VLAN filter and VLAN offload
- VLAN stripping and inserting
- QINQ stripping and inserting
+- Receive Side Scaling (RSS)
Driver compilation and testing
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index cc32b467a9..1349559c9b 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -61,6 +61,9 @@ zxdh_dev_infos_get(struct rte_eth_dev *dev,
dev_info->rx_offload_capa |= RTE_ETH_RX_OFFLOAD_TCP_LRO;
dev_info->rx_offload_capa |= RTE_ETH_RX_OFFLOAD_RSS_HASH;
+ dev_info->reta_size = RTE_ETH_RSS_RETA_SIZE_256;
+ dev_info->flow_type_rss_offloads = ZXDH_RSS_HF;
+
dev_info->tx_offload_capa = (RTE_ETH_TX_OFFLOAD_MULTI_SEGS);
dev_info->tx_offload_capa |= (RTE_ETH_TX_OFFLOAD_TCP_TSO |
RTE_ETH_TX_OFFLOAD_UDP_TSO);
@@ -784,9 +787,48 @@ zxdh_dev_conf_offload(struct rte_eth_dev *dev)
return ret;
}
+ ret = zxdh_rss_configure(dev);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "rss configure failed");
+ return ret;
+ }
+
return 0;
}
+static int
+zxdh_rss_qid_config(struct rte_eth_dev *dev)
+{
+ struct zxdh_hw *hw = dev->data->dev_private;
+ struct zxdh_port_attr_table port_attr = {0};
+ struct zxdh_msg_info msg_info = {0};
+ int ret = 0;
+
+ if (hw->is_pf) {
+ ret = zxdh_get_port_attr(hw->vport.vfid, &port_attr);
+ port_attr.port_base_qid = hw->channel_context[0].ph_chno &
0xfff;
+
+ ret = zxdh_set_port_attr(hw->vport.vfid, &port_attr);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "PF:%d port_base_qid insert failed",
hw->vfid);
+ return ret;
+ }
+ } else {
+ struct zxdh_port_attr_set_msg *attr_msg =
&msg_info.data.port_attr_msg;
+
+ zxdh_msg_head_build(hw, ZXDH_PORT_ATTRS_SET, &msg_info);
+ attr_msg->mode = ZXDH_PORT_BASE_QID_FLAG;
+ attr_msg->value = hw->channel_context[0].ph_chno & 0xfff;
+ ret = zxdh_vf_send_msg_to_pf(dev, &msg_info, sizeof(msg_info),
NULL, 0);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "Failed to send msg: port 0x%x msg
type %d ",
+ hw->vport.vport,
ZXDH_PORT_BASE_QID_FLAG);
+ return ret;
+ }
+ }
+ return ret;
+}
+
static int32_t
zxdh_dev_configure(struct rte_eth_dev *dev)
{
@@ -873,6 +915,12 @@ zxdh_dev_configure(struct rte_eth_dev *dev)
return -1;
}
+ ret = zxdh_rss_qid_config(dev);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "Failed to configure base qid!");
+ return -1;
+ }
+
zxdh_pci_reinit_complete(hw);
end:
@@ -1099,6 +1147,10 @@ static const struct eth_dev_ops zxdh_eth_dev_ops = {
.allmulticast_disable= zxdh_dev_allmulticast_disable,
.vlan_filter_set = zxdh_dev_vlan_filter_set,
.vlan_offload_set= zxdh_dev_vlan_offload_set,
+ .reta_update = zxdh_dev_rss_reta_update,
+ .reta_query = zxdh_dev_rss_reta_query,
+ .rss_hash_update = zxdh_rss_hash_update,
+ .rss_hash_conf_get = zxdh_rss_hash_conf_get,
};
static int32_t
diff --git a/drivers/net/zxdh/zxdh_ethdev.h b/drivers/net/zxdh/zxdh_ethdev.h
index 3cdac5de73..2934fa264a 100644
--- a/drivers/net/zxdh/zxdh_ethdev.h
+++ b/drivers/net/zxdh/zxdh_ethdev.h
@@ -82,6 +82,7 @@ struct zxdh_hw {
uint16_t queue_num;
uint16_t mc_num;
uint16_t uc_num;
+ uint16_t *rss_reta;
uint8_t *isr;
uint8_t wea
[PATCH v4 10/15] net/zxdh: mac set/add/remove ops implementations
provided mac set/add/remove ops.
Signed-off-by: Junlong Wang
---
doc/guides/nics/features/zxdh.ini | 2 +
doc/guides/nics/zxdh.rst | 2 +
drivers/net/zxdh/zxdh_common.c | 24 +++
drivers/net/zxdh/zxdh_common.h | 1 +
drivers/net/zxdh/zxdh_ethdev.c | 32 +++-
drivers/net/zxdh/zxdh_ethdev.h | 3 +
drivers/net/zxdh/zxdh_ethdev_ops.c | 233 +
drivers/net/zxdh/zxdh_ethdev_ops.h | 4 +
drivers/net/zxdh/zxdh_msg.h| 12 ++
drivers/net/zxdh/zxdh_np.h | 5 +
drivers/net/zxdh/zxdh_tables.c | 197
drivers/net/zxdh/zxdh_tables.h | 36 +
12 files changed, 549 insertions(+), 2 deletions(-)
diff --git a/doc/guides/nics/features/zxdh.ini
b/doc/guides/nics/features/zxdh.ini
index 7da3aaced1..dc09fe3453 100644
--- a/doc/guides/nics/features/zxdh.ini
+++ b/doc/guides/nics/features/zxdh.ini
@@ -12,3 +12,5 @@ Multiprocess aware = Y
Scattered Rx = Y
Link status = Y
Link status event= Y
+Unicast MAC filter = Y
+Multicast MAC filter = Y
diff --git a/doc/guides/nics/zxdh.rst b/doc/guides/nics/zxdh.rst
index fdbc3b3923..e0b0776aca 100644
--- a/doc/guides/nics/zxdh.rst
+++ b/doc/guides/nics/zxdh.rst
@@ -24,6 +24,8 @@ Features of the ZXDH PMD are:
- Link Auto-negotiation
- Link state information
- Set Link down or up
+- Unicast MAC filter
+- Multicast MAC filter
Driver compilation and testing
diff --git a/drivers/net/zxdh/zxdh_common.c b/drivers/net/zxdh/zxdh_common.c
index 4f18c97ed7..75883a8897 100644
--- a/drivers/net/zxdh/zxdh_common.c
+++ b/drivers/net/zxdh/zxdh_common.c
@@ -256,6 +256,30 @@ zxdh_panelid_get(struct rte_eth_dev *dev, uint8_t *panelid)
return ret;
}
+static int
+zxdh_get_res_hash_id(struct zxdh_res_para *in, uint8_t *hash_id)
+{
+ uint8_t reps = 0;
+ uint16_t reps_len = 0;
+
+ if (zxdh_get_res_info(in, ZXDH_TBL_FIELD_HASHID, &reps, &reps_len) !=
ZXDH_BAR_MSG_OK)
+ return -1;
+
+ *hash_id = reps;
+ return ZXDH_BAR_MSG_OK;
+}
+
+int32_t
+zxdh_hashidx_get(struct rte_eth_dev *dev, uint8_t *hash_idx)
+{
+ struct zxdh_res_para param;
+
+ zxdh_fill_res_para(dev, ¶m);
+ int32_t ret = zxdh_get_res_hash_id(¶m, hash_idx);
+
+ return ret;
+}
+
uint32_t
zxdh_read_bar_reg(struct rte_eth_dev *dev, uint32_t bar, uint32_t reg)
{
diff --git a/drivers/net/zxdh/zxdh_common.h b/drivers/net/zxdh/zxdh_common.h
index 72c29e1522..826f1fb95d 100644
--- a/drivers/net/zxdh/zxdh_common.h
+++ b/drivers/net/zxdh/zxdh_common.h
@@ -22,6 +22,7 @@ struct zxdh_res_para {
int32_t zxdh_phyport_get(struct rte_eth_dev *dev, uint8_t *phyport);
int32_t zxdh_panelid_get(struct rte_eth_dev *dev, uint8_t *pannelid);
+int32_t zxdh_hashidx_get(struct rte_eth_dev *dev, uint8_t *hash_idx);
uint32_t zxdh_read_bar_reg(struct rte_eth_dev *dev, uint32_t bar, uint32_t
reg);
void zxdh_write_bar_reg(struct rte_eth_dev *dev, uint32_t bar, uint32_t reg,
uint32_t val);
void zxdh_release_lock(struct zxdh_hw *hw);
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index d3876ec9b3..85ada87cdc 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -979,6 +979,23 @@ zxdh_set_rxtx_funcs(struct rte_eth_dev *eth_dev)
return 0;
}
+static int
+zxdh_mac_config(struct rte_eth_dev *eth_dev)
+{
+ struct zxdh_hw *hw = eth_dev->data->dev_private;
+ int ret = 0;
+
+ if (hw->is_pf) {
+ ret = zxdh_set_mac_table(hw->vport.vport,
+ ð_dev->data->mac_addrs[0],
hw->hash_search_index);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "Failed to add mac: port 0x%x",
hw->vport.vport);
+ return ret;
+ }
+ }
+ return ret;
+}
+
static int
zxdh_dev_start(struct rte_eth_dev *dev)
{
@@ -1014,6 +1031,9 @@ zxdh_dev_start(struct rte_eth_dev *dev)
zxdh_queue_notify(vq);
}
zxdh_dev_set_link_up(dev);
+ ret = zxdh_mac_config(hw->eth_dev);
+ if (ret)
+ PMD_DRV_LOG(ERR, " mac config failed");
return 0;
}
@@ -1032,6 +1052,9 @@ static const struct eth_dev_ops zxdh_eth_dev_ops = {
.link_update = zxdh_dev_link_update,
.dev_set_link_up = zxdh_dev_set_link_up,
.dev_set_link_down = zxdh_dev_set_link_down,
+ .mac_addr_add= zxdh_dev_mac_addr_add,
+ .mac_addr_remove = zxdh_dev_mac_addr_remove,
+ .mac_addr_set= zxdh_dev_mac_addr_set,
};
static int32_t
@@ -1073,15 +1096,20 @@ zxdh_agent_comm(struct rte_eth_dev *eth_dev, struct
zxdh_hw *hw)
PMD_DRV_LOG(ERR, "Failed to get phyport");
return -1;
}
- PMD_DRV_LOG(INFO, "Get phyport success: 0x%x", hw->phyport);
+ PMD_DRV_LOG(DEBUG, "Get phyport success: 0x%x", hw
[PATCH v4 07/15] net/zxdh: provided dev simple tx implementations
provided dev simple tx implementations.
Signed-off-by: Junlong Wang
---
drivers/net/zxdh/meson.build | 1 +
drivers/net/zxdh/zxdh_ethdev.c | 21 ++
drivers/net/zxdh/zxdh_queue.h | 26 ++-
drivers/net/zxdh/zxdh_rxtx.c | 396 +
drivers/net/zxdh/zxdh_rxtx.h | 4 +
5 files changed, 447 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/zxdh/zxdh_rxtx.c
diff --git a/drivers/net/zxdh/meson.build b/drivers/net/zxdh/meson.build
index 5b3af87c5b..20b2cf484a 100644
--- a/drivers/net/zxdh/meson.build
+++ b/drivers/net/zxdh/meson.build
@@ -21,4 +21,5 @@ sources = files(
'zxdh_queue.c',
'zxdh_np.c',
'zxdh_tables.c',
+'zxdh_rxtx.c',
)
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index 59ee942bdd..b6057edeaf 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -15,6 +15,7 @@
#include "zxdh_queue.h"
#include "zxdh_np.h"
#include "zxdh_tables.h"
+#include "zxdh_rxtx.h"
struct zxdh_hw_internal zxdh_hw_internal[RTE_MAX_ETHPORTS];
struct zxdh_shared_data *zxdh_shared_data;
@@ -951,6 +952,25 @@ zxdh_dev_close(struct rte_eth_dev *dev)
return ret;
}
+static int32_t
+zxdh_set_rxtx_funcs(struct rte_eth_dev *eth_dev)
+{
+ struct zxdh_hw *hw = eth_dev->data->dev_private;
+
+ if (!zxdh_pci_packed_queue(hw)) {
+ PMD_DRV_LOG(ERR, " port %u not support packed queue",
eth_dev->data->port_id);
+ return -1;
+ }
+ if (!zxdh_pci_with_feature(hw, ZXDH_NET_F_MRG_RXBUF)) {
+ PMD_DRV_LOG(ERR, " port %u not support rx mergeable",
eth_dev->data->port_id);
+ return -1;
+ }
+ eth_dev->tx_pkt_prepare = zxdh_xmit_pkts_prepare;
+ eth_dev->tx_pkt_burst = &zxdh_xmit_pkts_packed;
+
+ return 0;
+}
+
static int
zxdh_dev_start(struct rte_eth_dev *dev)
{
@@ -966,6 +986,7 @@ zxdh_dev_start(struct rte_eth_dev *dev)
if (ret < 0)
return ret;
}
+ zxdh_set_rxtx_funcs(dev);
ret = zxdh_intr_enable(dev);
if (ret) {
PMD_DRV_LOG(ERR, "interrupt enable failed");
diff --git a/drivers/net/zxdh/zxdh_queue.h b/drivers/net/zxdh/zxdh_queue.h
index 6513aec3f0..9343df81ac 100644
--- a/drivers/net/zxdh/zxdh_queue.h
+++ b/drivers/net/zxdh/zxdh_queue.h
@@ -21,8 +21,15 @@ enum { ZXDH_VTNET_RQ = 0, ZXDH_VTNET_TQ = 1 };
#define ZXDH_TQ_QUEUE_IDX 1
#define ZXDH_MAX_TX_INDIRECT 8
+/* This marks a buffer as continuing via the next field. */
+#define ZXDH_VRING_DESC_F_NEXT 1
+
/* This marks a buffer as write-only (otherwise read-only). */
-#define ZXDH_VRING_DESC_F_WRITE 2
+#define ZXDH_VRING_DESC_F_WRITE2
+
+/* This means the buffer contains a list of buffer descriptors. */
+#define ZXDH_VRING_DESC_F_INDIRECT 4
+
/* This flag means the descriptor was made available by the driver */
#define ZXDH_VRING_PACKED_DESC_F_AVAIL (1 << (7))
#define ZXDH_VRING_PACKED_DESC_F_USED(1 << (15))
@@ -35,11 +42,17 @@ enum { ZXDH_VTNET_RQ = 0, ZXDH_VTNET_TQ = 1 };
#define ZXDH_RING_EVENT_FLAGS_DISABLE 0x1
#define ZXDH_RING_EVENT_FLAGS_DESC0x2
+#define ZXDH_RING_F_INDIRECT_DESC 28
+
#define ZXDH_VQ_RING_DESC_CHAIN_END 32768
#define ZXDH_QUEUE_DEPTH 1024
#define ZXDH_RQ_QUEUE_IDX 0
#define ZXDH_TQ_QUEUE_IDX 1
+#define ZXDH_TYPE_HDR_SIZEsizeof(struct zxdh_type_hdr)
+#define ZXDH_PI_HDR_SIZE sizeof(struct zxdh_pi_hdr)
+#define ZXDH_DL_NET_HDR_SIZE sizeof(struct zxdh_net_hdr_dl)
+#define ZXDH_UL_NET_HDR_SIZE sizeof(struct zxdh_net_hdr_ul)
/*
* ring descriptors: 16 bytes.
@@ -355,6 +368,17 @@ static inline void zxdh_queue_notify(struct zxdh_virtqueue
*vq)
ZXDH_VTPCI_OPS(vq->hw)->notify_queue(vq->hw, vq);
}
+static inline int32_t
+zxdh_queue_kick_prepare_packed(struct zxdh_virtqueue *vq)
+{
+ uint16_t flags = 0;
+
+ zxdh_mb(vq->hw->weak_barriers);
+ flags = vq->vq_packed.ring.device->desc_event_flags;
+
+ return (flags != ZXDH_RING_EVENT_FLAGS_DISABLE);
+}
+
struct rte_mbuf *zxdh_queue_detach_unused(struct zxdh_virtqueue *vq);
int32_t zxdh_free_queues(struct rte_eth_dev *dev);
int32_t zxdh_get_queue_type(uint16_t vtpci_queue_idx);
diff --git a/drivers/net/zxdh/zxdh_rxtx.c b/drivers/net/zxdh/zxdh_rxtx.c
new file mode 100644
index 00..10034a0e98
--- /dev/null
+++ b/drivers/net/zxdh/zxdh_rxtx.c
@@ -0,0 +1,396 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2024 ZTE Corporation
+ */
+
+#include
+#include
+
+#include
+
+#include "zxdh_logs.h"
+#include "zxdh_pci.h"
+#include "zxdh_queue.h"
+
+#define ZXDH_PKT_FORM_CPU 0x20/* 1-cpu 0-np */
+#define ZXDH_NO_IP_FRAGMENT 0x2000 /* ip fragment flag */
+#define ZXDH_NO_IPID_UPDATE
[PATCH v4 12/15] net/zxdh: vlan filter/ offload ops implementations
provided vlan filter, vlan offload ops.
Signed-off-by: Junlong Wang
---
doc/guides/nics/features/zxdh.ini | 3 +
doc/guides/nics/zxdh.rst | 3 +
drivers/net/zxdh/zxdh_ethdev.c | 40 +-
drivers/net/zxdh/zxdh_ethdev_ops.c | 223 +
drivers/net/zxdh/zxdh_ethdev_ops.h | 2 +
drivers/net/zxdh/zxdh_msg.h| 22 +++
drivers/net/zxdh/zxdh_rxtx.c | 18 +++
drivers/net/zxdh/zxdh_tables.c | 99 +
drivers/net/zxdh/zxdh_tables.h | 10 +-
9 files changed, 417 insertions(+), 3 deletions(-)
diff --git a/doc/guides/nics/features/zxdh.ini
b/doc/guides/nics/features/zxdh.ini
index e9b237e102..6fb006c2da 100644
--- a/doc/guides/nics/features/zxdh.ini
+++ b/doc/guides/nics/features/zxdh.ini
@@ -16,3 +16,6 @@ Unicast MAC filter = Y
Multicast MAC filter = Y
Promiscuous mode = Y
Allmulticast mode= Y
+VLAN filter = Y
+VLAN offload = Y
+QinQ offload = Y
diff --git a/doc/guides/nics/zxdh.rst b/doc/guides/nics/zxdh.rst
index 0399df1302..3a7585d123 100644
--- a/doc/guides/nics/zxdh.rst
+++ b/doc/guides/nics/zxdh.rst
@@ -28,6 +28,9 @@ Features of the ZXDH PMD are:
- Multicast MAC filter
- Promiscuous mode
- Multicast mode
+- VLAN filter and VLAN offload
+- VLAN stripping and inserting
+- QINQ stripping and inserting
Driver compilation and testing
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index 1d64b877c1..cc32b467a9 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -758,6 +758,34 @@ zxdh_alloc_queues(struct rte_eth_dev *dev, uint16_t nr_vq)
return 0;
}
+static int
+zxdh_vlan_offload_configure(struct rte_eth_dev *dev)
+{
+ int ret;
+ int mask = RTE_ETH_VLAN_STRIP_MASK | RTE_ETH_VLAN_FILTER_MASK |
RTE_ETH_QINQ_STRIP_MASK;
+
+ ret = zxdh_dev_vlan_offload_set(dev, mask);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "vlan offload set error");
+ return -1;
+ }
+
+ return 0;
+}
+
+static int
+zxdh_dev_conf_offload(struct rte_eth_dev *dev)
+{
+ int ret = 0;
+
+ ret = zxdh_vlan_offload_configure(dev);
+ if (ret) {
+ PMD_DRV_LOG(ERR, "zxdh_vlan_offload_configure failed");
+ return ret;
+ }
+
+ return 0;
+}
static int32_t
zxdh_dev_configure(struct rte_eth_dev *dev)
@@ -815,7 +843,7 @@ zxdh_dev_configure(struct rte_eth_dev *dev)
nr_vq = dev->data->nb_rx_queues + dev->data->nb_tx_queues;
if (nr_vq == hw->queue_num)
- return 0;
+ goto end;
PMD_DRV_LOG(DEBUG, "queue changed need reset ");
/* Reset the device although not necessary at startup */
@@ -847,6 +875,8 @@ zxdh_dev_configure(struct rte_eth_dev *dev)
zxdh_pci_reinit_complete(hw);
+end:
+ zxdh_dev_conf_offload(dev);
return ret;
}
@@ -1067,6 +1097,8 @@ static const struct eth_dev_ops zxdh_eth_dev_ops = {
.promiscuous_disable = zxdh_dev_promiscuous_disable,
.allmulticast_enable = zxdh_dev_allmulticast_enable,
.allmulticast_disable= zxdh_dev_allmulticast_disable,
+ .vlan_filter_set = zxdh_dev_vlan_filter_set,
+ .vlan_offload_set= zxdh_dev_vlan_offload_set,
};
static int32_t
@@ -1325,6 +1357,12 @@ zxdh_tables_init(struct rte_eth_dev *dev)
return ret;
}
+ ret = zxdh_vlan_filter_table_init(dev);
+ if (ret) {
+ PMD_DRV_LOG(ERR, " vlan filter table init failed");
+ return ret;
+ }
+
return ret;
}
diff --git a/drivers/net/zxdh/zxdh_ethdev_ops.c
b/drivers/net/zxdh/zxdh_ethdev_ops.c
index aed4e6410c..94c5e6dbc8 100644
--- a/drivers/net/zxdh/zxdh_ethdev_ops.c
+++ b/drivers/net/zxdh/zxdh_ethdev_ops.c
@@ -2,6 +2,8 @@
* Copyright(c) 2024 ZTE Corporation
*/
+#include
+
#include "zxdh_ethdev.h"
#include "zxdh_pci.h"
#include "zxdh_msg.h"
@@ -9,6 +11,8 @@
#include "zxdh_tables.h"
#include "zxdh_logs.h"
+#define ZXDH_VLAN_FILTER_GROUPS 64
+
static int32_t zxdh_config_port_status(struct rte_eth_dev *dev, uint16_t
link_status)
{
struct zxdh_hw *hw = dev->data->dev_private;
@@ -529,3 +533,222 @@ int zxdh_dev_allmulticast_disable(struct rte_eth_dev *dev)
hw->allmulti_status = 0;
return ret;
}
+
+int
+zxdh_dev_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
+{
+ struct zxdh_hw *hw = (struct zxdh_hw *)dev->data->dev_private;
+ uint16_t idx = 0;
+ uint16_t bit_idx = 0;
+ uint8_t msg_type = 0;
+ int ret = 0;
+
+ vlan_id &= RTE_VLAN_ID_MASK;
+ if (vlan_id == 0 || vlan_id == RTE_ETHER_MAX_VLAN_ID) {
+ PMD_DRV_LOG(ERR, "vlan id (%d) is reserved", vlan_id);
+ return -EINVAL;
+ }
+
+ if (dev->data->dev_started == 0) {
+ PMD_DRV_LOG(ERR, "vlan_filter dev not start");
+
[PATCH v4 15/15] net/zxdh: mtu update ops implementations
mtu update ops implementations.
Signed-off-by: Junlong Wang
---
doc/guides/nics/features/zxdh.ini | 1 +
doc/guides/nics/zxdh.rst | 2 +
drivers/net/zxdh/zxdh_ethdev.c | 5 ++
drivers/net/zxdh/zxdh_ethdev_ops.c | 78 ++
drivers/net/zxdh/zxdh_ethdev_ops.h | 3 ++
drivers/net/zxdh/zxdh_tables.c | 42
drivers/net/zxdh/zxdh_tables.h | 4 ++
7 files changed, 135 insertions(+)
diff --git a/doc/guides/nics/features/zxdh.ini
b/doc/guides/nics/features/zxdh.ini
index 98c141cf95..3561e31666 100644
--- a/doc/guides/nics/features/zxdh.ini
+++ b/doc/guides/nics/features/zxdh.ini
@@ -24,3 +24,4 @@ RSS reta update = Y
Inner RSS= Y
Basic stats = Y
Stats per queue = Y
+MTU update = Y
diff --git a/doc/guides/nics/zxdh.rst b/doc/guides/nics/zxdh.rst
index c8a52b587c..58e0c49a2e 100644
--- a/doc/guides/nics/zxdh.rst
+++ b/doc/guides/nics/zxdh.rst
@@ -33,6 +33,8 @@ Features of the ZXDH PMD are:
- QINQ stripping and inserting
- Receive Side Scaling (RSS)
- Port hardware statistics
+- MTU update
+- Jumbo frames
Driver compilation and testing
diff --git a/drivers/net/zxdh/zxdh_ethdev.c b/drivers/net/zxdh/zxdh_ethdev.c
index a1822e1556..2906e3be6e 100644
--- a/drivers/net/zxdh/zxdh_ethdev.c
+++ b/drivers/net/zxdh/zxdh_ethdev.c
@@ -64,6 +64,10 @@ zxdh_dev_infos_get(struct rte_eth_dev *dev,
dev_info->reta_size = RTE_ETH_RSS_RETA_SIZE_256;
dev_info->flow_type_rss_offloads = ZXDH_RSS_HF;
+ dev_info->max_mtu = ZXDH_MAX_RX_PKTLEN - RTE_ETHER_HDR_LEN -
+ RTE_VLAN_HLEN - ZXDH_DL_NET_HDR_SIZE;
+ dev_info->min_mtu = ZXDH_ETHER_MIN_MTU;
+
dev_info->tx_offload_capa = (RTE_ETH_TX_OFFLOAD_MULTI_SEGS);
dev_info->tx_offload_capa |= (RTE_ETH_TX_OFFLOAD_TCP_TSO |
RTE_ETH_TX_OFFLOAD_UDP_TSO);
@@ -1153,6 +1157,7 @@ static const struct eth_dev_ops zxdh_eth_dev_ops = {
.rss_hash_conf_get = zxdh_rss_hash_conf_get,
.stats_get = zxdh_dev_stats_get,
.stats_reset = zxdh_dev_stats_reset,
+ .mtu_set = zxdh_dev_mtu_set,
};
static int32_t
diff --git a/drivers/net/zxdh/zxdh_ethdev_ops.c
b/drivers/net/zxdh/zxdh_ethdev_ops.c
index 2c10f171aa..8e72a9a6b2 100644
--- a/drivers/net/zxdh/zxdh_ethdev_ops.c
+++ b/drivers/net/zxdh/zxdh_ethdev_ops.c
@@ -13,6 +13,7 @@
#include "zxdh_logs.h"
#include "zxdh_rxtx.h"
#include "zxdh_np.h"
+#include "zxdh_queue.h"
#define ZXDH_VLAN_FILTER_GROUPS 64
#define ZXDH_INVALID_LOGIC_QID0xU
@@ -1515,3 +1516,80 @@ int zxdh_dev_stats_reset(struct rte_eth_dev *dev)
return 0;
}
+
+int zxdh_dev_mtu_set(struct rte_eth_dev *dev, uint16_t new_mtu)
+{
+ struct zxdh_hw *hw = dev->data->dev_private;
+ struct zxdh_panel_table panel = {0};
+ struct zxdh_port_attr_table vport_att = {0};
+ uint16_t vfid = zxdh_vport_to_vfid(hw->vport);
+ uint16_t max_mtu = 0;
+ int ret = 0;
+
+ max_mtu = ZXDH_MAX_RX_PKTLEN - RTE_ETHER_HDR_LEN - RTE_VLAN_HLEN -
ZXDH_DL_NET_HDR_SIZE;
+ if (new_mtu < ZXDH_ETHER_MIN_MTU || new_mtu > max_mtu) {
+ PMD_DRV_LOG(ERR, "invalid mtu:%d, range[%d, %d]",
+ new_mtu, ZXDH_ETHER_MIN_MTU, max_mtu);
+ return -EINVAL;
+ }
+
+ if (dev->data->mtu == new_mtu)
+ return 0;
+
+ if (hw->is_pf) {
+ memset(&panel, 0, sizeof(panel));
+ memset(&vport_att, 0, sizeof(vport_att));
+ ret = zxdh_get_panel_attr(dev, &panel);
+ if (ret != 0) {
+ PMD_DRV_LOG(ERR, "get_panel_attr ret:%d", ret);
+ return -1;
+ }
+
+ ret = zxdh_get_port_attr(vfid, &vport_att);
+ if (ret != 0) {
+ PMD_DRV_LOG(ERR,
+ "[vfid:%d] zxdh_dev_mtu, get vport dpp_ret:%d",
vfid, ret);
+ return -1;
+ }
+
+ panel.mtu = new_mtu;
+ panel.mtu_enable = 1;
+ ret = zxdh_set_panel_attr(dev, &panel);
+ if (ret != 0) {
+ PMD_DRV_LOG(ERR, "set zxdh_dev_mtu failed, ret:%u",
ret);
+ return ret;
+ }
+
+ vport_att.mtu_enable = 1;
+ vport_att.mtu = new_mtu;
+ ret = zxdh_set_port_attr(vfid, &vport_att);
+ if (ret != 0) {
+ PMD_DRV_LOG(ERR,
+ "[vfid:%d] zxdh_dev_mtu, set vport dpp_ret:%d",
vfid, ret);
+ return ret;
+ }
+ } else {
+ struct zxdh_msg_info msg_info = {0};
+ struct zxdh_port_attr_set_msg *attr_msg =
&msg_info.data.port_attr_msg;
+
+ zxdh_ms
[PATCH v2] net/nfp: implement the burst mode get operation
Implement the burst mode get operation functions for both Rx and Tx.
Signed-off-by: Chaoyong He
Reviewed-by: Peng Zhang
---
v2:
* Replace 'snprintf()' with 'strlcpy()'.
---
drivers/net/nfp/nfp_ethdev.c| 2 ++
drivers/net/nfp/nfp_ethdev_vf.c | 2 ++
drivers/net/nfp/nfp_rxtx.c | 46 +
drivers/net/nfp/nfp_rxtx.h | 4 +++
4 files changed, 54 insertions(+)
diff --git a/drivers/net/nfp/nfp_ethdev.c b/drivers/net/nfp/nfp_ethdev.c
index f54483822f..df5482f74a 100644
--- a/drivers/net/nfp/nfp_ethdev.c
+++ b/drivers/net/nfp/nfp_ethdev.c
@@ -985,6 +985,8 @@ static const struct eth_dev_ops nfp_net_eth_dev_ops = {
.get_module_eeprom = nfp_net_get_module_eeprom,
.dev_led_on = nfp_net_led_on,
.dev_led_off= nfp_net_led_off,
+ .rx_burst_mode_get = nfp_net_rx_burst_mode_get,
+ .tx_burst_mode_get = nfp_net_tx_burst_mode_get,
};
static inline void
diff --git a/drivers/net/nfp/nfp_ethdev_vf.c b/drivers/net/nfp/nfp_ethdev_vf.c
index 36b98dc0c2..23fa5b82ad 100644
--- a/drivers/net/nfp/nfp_ethdev_vf.c
+++ b/drivers/net/nfp/nfp_ethdev_vf.c
@@ -232,6 +232,8 @@ static const struct eth_dev_ops nfp_netvf_eth_dev_ops = {
.tx_queue_release = nfp_net_tx_queue_release,
.rx_queue_intr_enable = nfp_rx_queue_intr_enable,
.rx_queue_intr_disable = nfp_rx_queue_intr_disable,
+ .rx_burst_mode_get = nfp_net_rx_burst_mode_get,
+ .tx_burst_mode_get = nfp_net_tx_burst_mode_get,
};
static inline void
diff --git a/drivers/net/nfp/nfp_rxtx.c b/drivers/net/nfp/nfp_rxtx.c
index 35fb637b21..57c0b7351b 100644
--- a/drivers/net/nfp/nfp_rxtx.c
+++ b/drivers/net/nfp/nfp_rxtx.c
@@ -12,6 +12,7 @@
#include "nfd3/nfp_nfd3.h"
#include "nfdk/nfp_nfdk.h"
+#include "nfdk/nfp_nfdk_vec.h"
#include "flower/nfp_flower.h"
#include "nfp_ipsec.h"
@@ -893,3 +894,48 @@ nfp_net_recv_pkts_set(struct rte_eth_dev *eth_dev)
else
eth_dev->rx_pkt_burst = nfp_net_recv_pkts;
}
+
+int
+nfp_net_rx_burst_mode_get(struct rte_eth_dev *eth_dev,
+ uint16_t queue_id __rte_unused,
+ struct rte_eth_burst_mode *mode)
+{
+ eth_rx_burst_t pkt_burst;
+
+ pkt_burst = eth_dev->rx_pkt_burst;
+ if (pkt_burst == nfp_net_recv_pkts) {
+ strlcpy(mode->info, "Scalar",
+ RTE_ETH_BURST_MODE_INFO_SIZE);
+ } else if (pkt_burst == nfp_net_vec_avx2_recv_pkts) {
+ strlcpy(mode->info, "Vector AVX2",
+ RTE_ETH_BURST_MODE_INFO_SIZE);
+ } else {
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+int
+nfp_net_tx_burst_mode_get(struct rte_eth_dev *eth_dev,
+ uint16_t queue_id __rte_unused,
+ struct rte_eth_burst_mode *mode)
+{
+ eth_tx_burst_t pkt_burst;
+
+ pkt_burst = eth_dev->tx_pkt_burst;
+ if (pkt_burst == nfp_net_nfd3_xmit_pkts) {
+ strlcpy(mode->info, "NFD3 Scalar",
+ RTE_ETH_BURST_MODE_INFO_SIZE);
+ } else if (pkt_burst == nfp_net_nfdk_xmit_pkts) {
+ strlcpy(mode->info, "NFDk Scalar",
+ RTE_ETH_BURST_MODE_INFO_SIZE);
+ } else if (pkt_burst == nfp_net_nfdk_vec_avx2_xmit_pkts) {
+ strlcpy(mode->info, "NFDk Vector AVX2",
+ RTE_ETH_BURST_MODE_INFO_SIZE);
+ } else {
+ return -EINVAL;
+ }
+
+ return 0;
+}
diff --git a/drivers/net/nfp/nfp_rxtx.h b/drivers/net/nfp/nfp_rxtx.h
index c717d97003..48cbd83787 100644
--- a/drivers/net/nfp/nfp_rxtx.h
+++ b/drivers/net/nfp/nfp_rxtx.h
@@ -245,6 +245,10 @@ void nfp_net_tx_queue_info_get(struct rte_eth_dev *dev,
uint16_t queue_id,
struct rte_eth_txq_info *qinfo);
void nfp_net_recv_pkts_set(struct rte_eth_dev *eth_dev);
+int nfp_net_rx_burst_mode_get(struct rte_eth_dev *eth_dev, uint16_t queue_id,
+ struct rte_eth_burst_mode *mode);
+int nfp_net_tx_burst_mode_get(struct rte_eth_dev *eth_dev, uint16_t queue_id,
+ struct rte_eth_burst_mode *mode);
void nfp_net_parse_ptype(struct nfp_net_rxq *rxq,
struct nfp_net_rx_desc *rxds,
struct rte_mbuf *mb);
--
2.43.5
Re: [PATCH] net/gve: Allocate qpl pages using malloc if memzone allocation fails
On Wed, 18 Dec 2024 15:46:35 -0800 Joshua Washington wrote: > From: Praveen Kaligineedi > > Allocating QPL for an RX queue might fail if enough contiguous IOVA > memory cannot be allocated. However, the only requirement for QPL > for RX is that each 4K buffer be IOVA contiguous, not the entire > QPL. Therefore, use malloc to allocate 4K buffers if the allocation > using memzone fails. > > Signed-off-by: Praveen Kaligineedi > Reviewed-by: Joshua Washington > --- Why keep the memzone code? rte_malloc and memzone are both coming from huge pages. Is there any advantage to memzone for what you are doing? Better to not have two potential allocation paths to test.
RE: [PATCH v2 2/2] net/bonding: add command to set dedicated queue size
> > On Wed, 4 Dec 2024 06:21:00 + > > Chaoyong He wrote: > > > > > > The definition of what a "dedicated queue" is a bit confusing. > > > > If it is only for LACP packets, it should never need to be very big. > > > > Only under a mis-configuration and DoS kind of flood should there > > > > ever be many packets. > > > > > > Yes, the dedicated queue is only for LACP packets now and it doesn't > > > need be > > set very big. > > > > > > But if we use a hardware queue as the "dedicated queue", we must > > > consider the hardware capability. The minimum queue size of some > > > NICs may be larger than the hardcode dedicated queue size. In this > > > case, I think it > > is better to add an interface to set the dedicated queue size. > > > > How about using the existing descriptor queue limits api for that? > > It is reported by info get > > Using existing descriptor queue limits api is good enough for current > problem(hardware capability), but I think it is not very flexible. > Now we use a macro as a default value for dedicated queue size, but we can > replace the macro with queue limit while still retaining the interface for > modifying queue size. > What do you think of this? A gentle ping ~
[PATCH 1/1] vhost: fix a double fetch when dequeue offloading
The hdr->csum_start does two successive reads from user space to read a
variable length data structure. The result overflow if the data structure
changes between the two reads.
To fix this, we can prevent double fetch issue by copying virtio_hdr to
the temporary variable.
Fixes: 4dc4e33ffa10 ("net/virtio: fix Rx checksum calculation")
Cc: sta...@dpdk.org
Signed-off-by: Yunjian Wang
---
lib/vhost/virtio_net.c | 13 -
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
index 69901ab3b5..5c40ae7069 100644
--- a/lib/vhost/virtio_net.c
+++ b/lib/vhost/virtio_net.c
@@ -2914,10 +2914,12 @@ desc_to_mbuf(struct virtio_net *dev, struct
vhost_virtqueue *vq,
* in a contiguous virtual area.
*/
copy_vnet_hdr_from_desc(&tmp_hdr, buf_vec);
- hdr = &tmp_hdr;
} else {
- hdr = (struct virtio_net_hdr
*)((uintptr_t)buf_vec[0].buf_addr);
+ rte_memcpy((void *)(uintptr_t)&tmp_hdr,
+ (void *)(uintptr_t)buf_vec[0].buf_addr,
+ sizeof(struct virtio_net_hdr));
}
+ hdr = &tmp_hdr;
}
for (vec_idx = 0; vec_idx < nr_vec; vec_idx++) {
@@ -3363,7 +3365,7 @@ virtio_dev_tx_batch_packed(struct virtio_net *dev,
{
uint16_t avail_idx = vq->last_avail_idx;
uint32_t buf_offset = sizeof(struct virtio_net_hdr_mrg_rxbuf);
- struct virtio_net_hdr *hdr;
+ struct virtio_net_hdr hdr;
uintptr_t desc_addrs[PACKED_BATCH_SIZE];
uint16_t ids[PACKED_BATCH_SIZE];
uint16_t i;
@@ -3382,8 +3384,9 @@ virtio_dev_tx_batch_packed(struct virtio_net *dev,
if (virtio_net_with_host_offload(dev)) {
vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) {
- hdr = (struct virtio_net_hdr *)(desc_addrs[i]);
- vhost_dequeue_offload(dev, hdr, pkts[i],
legacy_ol_flags);
+ rte_memcpy((void *)(uintptr_t)&hdr,
+ (void *)(uintptr_t)desc_addrs[i], sizeof(struct
virtio_net_hdr));
+ vhost_dequeue_offload(dev, &hdr, pkts[i],
legacy_ol_flags);
}
}
--
2.33.0
[RFC PATCH] eventdev: adapter API to configure multiple Rx queues
This RFC introduces a new API, rte_event_eth_rx_adapter_queues_add(),
designed to enhance the flexibility of configuring multiple Rx queues
in eventdev Rx adapter.
The existing rte_event_eth_rx_adapter_queue_add() API supports adding
multiple queues by specifying rx_queue_id = -1, but it lacks the ability
to apply specific configurations to each of the added queues.
The proposed API, rte_event_eth_rx_adapter_queues_add, addresses this
limitation by:
- Enabling users to specify an array of rx_queue_id values alongside
individual configurations for each queue.
- Supporting a nb_rx_queues argument to define the number of queues to
configure. When set to 0, the API applies a common configuration to
all queues, similar to the existing rx_queue_id = -1 behavior.
This enhancement allows for more granular control when configuring
multiple Rx queues. Additionally, the API can act as a replacement for
the older API, offering both flexibility and improved functionality.
Signed-off-by: Shijith Thotton
---
lib/eventdev/eventdev_pmd.h | 34 +
lib/eventdev/rte_event_eth_rx_adapter.h | 34 +
2 files changed, 68 insertions(+)
diff --git a/lib/eventdev/eventdev_pmd.h b/lib/eventdev/eventdev_pmd.h
index 36148f8d86..2e458a9779 100644
--- a/lib/eventdev/eventdev_pmd.h
+++ b/lib/eventdev/eventdev_pmd.h
@@ -25,6 +25,7 @@
#include
#include "event_timer_adapter_pmd.h"
+#include "rte_event_eth_rx_adapter.h"
#include "rte_eventdev.h"
#ifdef __cplusplus
@@ -708,6 +709,37 @@ typedef int (*eventdev_eth_rx_adapter_queue_add_t)(
int32_t rx_queue_id,
const struct rte_event_eth_rx_adapter_queue_conf *queue_conf);
+/**
+ * Add ethernet Rx queues to event device. This callback is invoked if
+ * the caps returned from rte_eventdev_eth_rx_adapter_caps_get(, eth_port_id)
+ * has RTE_EVENT_ETH_RX_ADAPTER_CAP_INTERNAL_PORT set.
+ *
+ * @param dev
+ * Event device pointer
+ *
+ * @param eth_dev
+ * Ethernet device pointer
+ *
+ * @param rx_queue_id
+ * Ethernet device receive queue index array
+ *
+ * @param queue_conf
+ * Additional configuration structure array
+ *
+ * @param nb_rx_queues
+ * Number of ethernet device receive queues
+ *
+ * @return
+ * - 0: Success, ethernet receive queues added successfully.
+ * - <0: Error code returned by the driver function.
+ */
+typedef int (*eventdev_eth_rx_adapter_queues_add_t)(
+ const struct rte_eventdev *dev,
+ const struct rte_eth_dev *eth_dev,
+ int32_t rx_queue_id[],
+ const struct rte_event_eth_rx_adapter_queue_conf queue_conf[],
+ uint16_t nb_rx_queues);
+
/**
* Delete ethernet Rx queues from event device. This callback is invoked if
* the caps returned from eventdev_eth_rx_adapter_caps_get(, eth_port_id)
@@ -1578,6 +1610,8 @@ struct eventdev_ops {
/**< Get ethernet Rx adapter capabilities */
eventdev_eth_rx_adapter_queue_add_t eth_rx_adapter_queue_add;
/**< Add Rx queues to ethernet Rx adapter */
+ eventdev_eth_rx_adapter_queues_add_t eth_rx_adapter_queues_add;
+ /**< Add Rx queues to ethernet Rx adapter */
eventdev_eth_rx_adapter_queue_del_t eth_rx_adapter_queue_del;
/**< Delete Rx queues from ethernet Rx adapter */
eventdev_eth_rx_adapter_queue_conf_get_t eth_rx_adapter_queue_conf_get;
diff --git a/lib/eventdev/rte_event_eth_rx_adapter.h
b/lib/eventdev/rte_event_eth_rx_adapter.h
index 9237e198a7..9a5c560b67 100644
--- a/lib/eventdev/rte_event_eth_rx_adapter.h
+++ b/lib/eventdev/rte_event_eth_rx_adapter.h
@@ -553,6 +553,40 @@ int rte_event_eth_rx_adapter_queue_add(uint8_t id,
int32_t rx_queue_id,
const struct rte_event_eth_rx_adapter_queue_conf *conf);
+/**
+ * Add multiple receive queues to an event adapter.
+ *
+ * @param id
+ * Adapter identifier.
+ *
+ * @param eth_dev_id
+ * Port identifier of Ethernet device.
+ *
+ * @param rx_queue_id
+ * Array of Ethernet device receive queue indices.
+ * If nb_rx_queues is 0, then rx_queue_id is ignored.
+ *
+ * @param conf
+ * Array of additional configuration structures of type
+ * *rte_event_eth_rx_adapter_queue_conf*. conf[i] is used for rx_queue_id[i].
+ * If nb_rx_queues is 0, then conf[0] is used for all Rx queues.
+ *
+ * @param nb_rx_queues
+ * Number of receive queues to add.
+ * If nb_rx_queues is 0, then all Rx queues configured for
+ * the device are added with the same configuration in conf[0].
+ * @see RTE_EVENT_ETH_RX_ADAPTER_CAP_MULTI_EVENTQ
+ *
+ * @return
+ * - 0: Success, Receive queues added correctly.
+ * - <0: Error code on failure.
+ */
+__rte_experimental
+int rte_event_eth_rx_adapter_queues_add(
+ uint8_t id, uint16_t eth_dev_id, int32_t rx_queue_id[],
+ const struct rte_event_eth_rx_adapter_queue_conf conf[],
+ uint16_t nb_rx_queues);
+
[RFC PATCH] eventdev: adapter API to configure multiple Rx queues
This RFC introduces a new API, rte_event_eth_rx_adapter_queues_add(),
designed to enhance the flexibility of configuring multiple Rx queues
in eventdev Rx adapter.
The existing rte_event_eth_rx_adapter_queue_add() API supports adding
multiple queues by specifying rx_queue_id = -1, but it lacks the ability
to apply specific configurations to each of the added queues.
The proposed API, rte_event_eth_rx_adapter_queues_add, addresses this
limitation by:
- Enabling users to specify an array of rx_queue_id values alongside
individual configurations for each queue.
- Supporting a nb_rx_queues argument to define the number of queues to
configure. When set to 0, the API applies a common configuration to
all queues, similar to the existing rx_queue_id = -1 behavior.
This enhancement allows for more granular control when configuring
multiple Rx queues. Additionally, the API can act as a replacement for
the older API, offering both flexibility and improved functionality.
Signed-off-by: Shijith Thotton
---
lib/eventdev/eventdev_pmd.h | 34 +
lib/eventdev/rte_event_eth_rx_adapter.h | 34 +
2 files changed, 68 insertions(+)
diff --git a/lib/eventdev/eventdev_pmd.h b/lib/eventdev/eventdev_pmd.h
index 36148f8d86..2e458a9779 100644
--- a/lib/eventdev/eventdev_pmd.h
+++ b/lib/eventdev/eventdev_pmd.h
@@ -25,6 +25,7 @@
#include
#include "event_timer_adapter_pmd.h"
+#include "rte_event_eth_rx_adapter.h"
#include "rte_eventdev.h"
#ifdef __cplusplus
@@ -708,6 +709,37 @@ typedef int (*eventdev_eth_rx_adapter_queue_add_t)(
int32_t rx_queue_id,
const struct rte_event_eth_rx_adapter_queue_conf *queue_conf);
+/**
+ * Add ethernet Rx queues to event device. This callback is invoked if
+ * the caps returned from rte_eventdev_eth_rx_adapter_caps_get(, eth_port_id)
+ * has RTE_EVENT_ETH_RX_ADAPTER_CAP_INTERNAL_PORT set.
+ *
+ * @param dev
+ * Event device pointer
+ *
+ * @param eth_dev
+ * Ethernet device pointer
+ *
+ * @param rx_queue_id
+ * Ethernet device receive queue index array
+ *
+ * @param queue_conf
+ * Additional configuration structure array
+ *
+ * @param nb_rx_queues
+ * Number of ethernet device receive queues
+ *
+ * @return
+ * - 0: Success, ethernet receive queues added successfully.
+ * - <0: Error code returned by the driver function.
+ */
+typedef int (*eventdev_eth_rx_adapter_queues_add_t)(
+ const struct rte_eventdev *dev,
+ const struct rte_eth_dev *eth_dev,
+ int32_t rx_queue_id[],
+ const struct rte_event_eth_rx_adapter_queue_conf queue_conf[],
+ uint16_t nb_rx_queues);
+
/**
* Delete ethernet Rx queues from event device. This callback is invoked if
* the caps returned from eventdev_eth_rx_adapter_caps_get(, eth_port_id)
@@ -1578,6 +1610,8 @@ struct eventdev_ops {
/**< Get ethernet Rx adapter capabilities */
eventdev_eth_rx_adapter_queue_add_t eth_rx_adapter_queue_add;
/**< Add Rx queues to ethernet Rx adapter */
+ eventdev_eth_rx_adapter_queues_add_t eth_rx_adapter_queues_add;
+ /**< Add Rx queues to ethernet Rx adapter */
eventdev_eth_rx_adapter_queue_del_t eth_rx_adapter_queue_del;
/**< Delete Rx queues from ethernet Rx adapter */
eventdev_eth_rx_adapter_queue_conf_get_t eth_rx_adapter_queue_conf_get;
diff --git a/lib/eventdev/rte_event_eth_rx_adapter.h
b/lib/eventdev/rte_event_eth_rx_adapter.h
index 9237e198a7..9a5c560b67 100644
--- a/lib/eventdev/rte_event_eth_rx_adapter.h
+++ b/lib/eventdev/rte_event_eth_rx_adapter.h
@@ -553,6 +553,40 @@ int rte_event_eth_rx_adapter_queue_add(uint8_t id,
int32_t rx_queue_id,
const struct rte_event_eth_rx_adapter_queue_conf *conf);
+/**
+ * Add multiple receive queues to an event adapter.
+ *
+ * @param id
+ * Adapter identifier.
+ *
+ * @param eth_dev_id
+ * Port identifier of Ethernet device.
+ *
+ * @param rx_queue_id
+ * Array of Ethernet device receive queue indices.
+ * If nb_rx_queues is 0, then rx_queue_id is ignored.
+ *
+ * @param conf
+ * Array of additional configuration structures of type
+ * *rte_event_eth_rx_adapter_queue_conf*. conf[i] is used for rx_queue_id[i].
+ * If nb_rx_queues is 0, then conf[0] is used for all Rx queues.
+ *
+ * @param nb_rx_queues
+ * Number of receive queues to add.
+ * If nb_rx_queues is 0, then all Rx queues configured for
+ * the device are added with the same configuration in conf[0].
+ * @see RTE_EVENT_ETH_RX_ADAPTER_CAP_MULTI_EVENTQ
+ *
+ * @return
+ * - 0: Success, Receive queues added correctly.
+ * - <0: Error code on failure.
+ */
+__rte_experimental
+int rte_event_eth_rx_adapter_queues_add(
+ uint8_t id, uint16_t eth_dev_id, int32_t rx_queue_id[],
+ const struct rte_event_eth_rx_adapter_queue_conf conf[],
+ uint16_t nb_rx_queues);
+
RE: [RFC PATCH] eventdev: adapter API to configure multiple Rx queues
>This RFC introduces a new API, rte_event_eth_rx_adapter_queues_add(),
>designed to enhance the flexibility of configuring multiple Rx queues
>in eventdev Rx adapter.
>
>The existing rte_event_eth_rx_adapter_queue_add() API supports adding
>multiple queues by specifying rx_queue_id = -1, but it lacks the ability
>to apply specific configurations to each of the added queues.
>
>The proposed API, rte_event_eth_rx_adapter_queues_add, addresses this
>limitation by:
>
>- Enabling users to specify an array of rx_queue_id values alongside
> individual configurations for each queue.
>
>- Supporting a nb_rx_queues argument to define the number of queues to
> configure. When set to 0, the API applies a common configuration to
> all queues, similar to the existing rx_queue_id = -1 behavior.
>
>This enhancement allows for more granular control when configuring
>multiple Rx queues. Additionally, the API can act as a replacement for
>the older API, offering both flexibility and improved functionality.
>
>Signed-off-by: Shijith Thotton
>---
> lib/eventdev/eventdev_pmd.h | 34 +
> lib/eventdev/rte_event_eth_rx_adapter.h | 34
>+
> 2 files changed, 68 insertions(+)
>
>diff --git a/lib/eventdev/eventdev_pmd.h b/lib/eventdev/eventdev_pmd.h
>index 36148f8d86..2e458a9779 100644
>--- a/lib/eventdev/eventdev_pmd.h
>+++ b/lib/eventdev/eventdev_pmd.h
>@@ -25,6 +25,7 @@
> #include
>
> #include "event_timer_adapter_pmd.h"
>+#include "rte_event_eth_rx_adapter.h"
> #include "rte_eventdev.h"
>
> #ifdef __cplusplus
>@@ -708,6 +709,37 @@ typedef int
>(*eventdev_eth_rx_adapter_queue_add_t)(
> int32_t rx_queue_id,
> const struct rte_event_eth_rx_adapter_queue_conf
>*queue_conf);
>
>+/**
>+ * Add ethernet Rx queues to event device. This callback is invoked if
>+ * the caps returned from rte_eventdev_eth_rx_adapter_caps_get(,
>eth_port_id)
>+ * has RTE_EVENT_ETH_RX_ADAPTER_CAP_INTERNAL_PORT set.
>+ *
>+ * @param dev
>+ * Event device pointer
>+ *
>+ * @param eth_dev
>+ * Ethernet device pointer
>+ *
>+ * @param rx_queue_id
>+ * Ethernet device receive queue index array
>+ *
>+ * @param queue_conf
>+ * Additional configuration structure array
>+ *
>+ * @param nb_rx_queues
>+ * Number of ethernet device receive queues
>+ *
>+ * @return
>+ * - 0: Success, ethernet receive queues added successfully.
>+ * - <0: Error code returned by the driver function.
>+ */
>+typedef int (*eventdev_eth_rx_adapter_queues_add_t)(
>+ const struct rte_eventdev *dev,
>+ const struct rte_eth_dev *eth_dev,
>+ int32_t rx_queue_id[],
>+ const struct rte_event_eth_rx_adapter_queue_conf
>queue_conf[],
>+ uint16_t nb_rx_queues);
>+
> /**
> * Delete ethernet Rx queues from event device. This callback is invoked if
> * the caps returned from eventdev_eth_rx_adapter_caps_get(, eth_port_id)
>@@ -1578,6 +1610,8 @@ struct eventdev_ops {
> /**< Get ethernet Rx adapter capabilities */
> eventdev_eth_rx_adapter_queue_add_t eth_rx_adapter_queue_add;
> /**< Add Rx queues to ethernet Rx adapter */
>+ eventdev_eth_rx_adapter_queues_add_t eth_rx_adapter_queues_add;
>+ /**< Add Rx queues to ethernet Rx adapter */
> eventdev_eth_rx_adapter_queue_del_t eth_rx_adapter_queue_del;
> /**< Delete Rx queues from ethernet Rx adapter */
> eventdev_eth_rx_adapter_queue_conf_get_t
>eth_rx_adapter_queue_conf_get;
>diff --git a/lib/eventdev/rte_event_eth_rx_adapter.h
>b/lib/eventdev/rte_event_eth_rx_adapter.h
>index 9237e198a7..9a5c560b67 100644
>--- a/lib/eventdev/rte_event_eth_rx_adapter.h
>+++ b/lib/eventdev/rte_event_eth_rx_adapter.h
>@@ -553,6 +553,40 @@ int rte_event_eth_rx_adapter_queue_add(uint8_t id,
> int32_t rx_queue_id,
> const struct rte_event_eth_rx_adapter_queue_conf
>*conf);
>
>+/**
>+ * Add multiple receive queues to an event adapter.
>+ *
>+ * @param id
>+ * Adapter identifier.
>+ *
>+ * @param eth_dev_id
>+ * Port identifier of Ethernet device.
>+ *
>+ * @param rx_queue_id
>+ * Array of Ethernet device receive queue indices.
>+ * If nb_rx_queues is 0, then rx_queue_id is ignored.
>+ *
>+ * @param conf
>+ * Array of additional configuration structures of type
>+ * *rte_event_eth_rx_adapter_queue_conf*. conf[i] is used for
>rx_queue_id[i].
>+ * If nb_rx_queues is 0, then conf[0] is used for all Rx queues.
>+ *
>+ * @param nb_rx_queues
>+ * Number of receive queues to add.
>+ * If nb_rx_queues is 0, then all Rx queues configured for
>+ * the device are added with the same configuration in conf[0].
>+ * @see RTE_EVENT_ETH_RX_ADAPTER_CAP_MULTI_EVENTQ
>+ *
>+ * @return
>+ * - 0: Success, Receive queues added correctly.
>+ * - <0: Error code on failure.
>+ */
>+__rte_experimental
>+int rte_event_eth_rx_adapter_queues_add(
>+ uint8_t id, uint16_t eth_dev_id, int32_t rx_queue_id[],
>+
[PATCH] net/gve: Allocate qpl pages using malloc if memzone allocation fails
From: Praveen Kaligineedi
Allocating QPL for an RX queue might fail if enough contiguous IOVA
memory cannot be allocated. However, the only requirement for QPL
for RX is that each 4K buffer be IOVA contiguous, not the entire
QPL. Therefore, use malloc to allocate 4K buffers if the allocation
using memzone fails.
Signed-off-by: Praveen Kaligineedi
Reviewed-by: Joshua Washington
---
drivers/net/gve/gve_ethdev.c | 105
++
drivers/net/gve/gve_ethdev.h | 1 +
drivers/net/gve/gve_rx.c | 2 +-
3 files changed, 89 insertions(+), 19 deletions(-)
diff --git a/drivers/net/gve/gve_ethdev.c b/drivers/net/gve/gve_ethdev.c
index db4ebe7..09304ef 100644
--- a/drivers/net/gve/gve_ethdev.c
+++ b/drivers/net/gve/gve_ethdev.c
@@ -22,42 +22,100 @@ gve_write_version(uint8_t *driver_version_register)
writeb('\n', driver_version_register);
}
+static const struct rte_memzone *
+gve_alloc_using_mz(const char *name, uint32_t num_pages)
+{
+ const struct rte_memzone *mz;
+ mz = rte_memzone_reserve_aligned(name, num_pages * PAGE_SIZE,
+rte_socket_id(),
+RTE_MEMZONE_IOVA_CONTIG, PAGE_SIZE);
+ if (mz == NULL)
+ PMD_DRV_LOG(ERR, "Failed to alloc memzone %s.", name);
+ return mz;
+}
+
+static int
+gve_alloc_using_malloc(void **bufs, uint32_t num_entries)
+{
+ uint32_t i;
+
+ for (i = 0; i < num_entries; i++) {
+ bufs[i] = rte_malloc_socket(NULL, PAGE_SIZE, PAGE_SIZE,
rte_socket_id());
+ if (bufs[i] == NULL) {
+ PMD_DRV_LOG(ERR, "Failed to malloc");
+ goto free_bufs;
+ }
+ }
+ return 0;
+
+free_bufs:
+ while (i > 0)
+ rte_free(bufs[--i]);
+
+ return -ENOMEM;
+}
+
static struct gve_queue_page_list *
-gve_alloc_queue_page_list(const char *name, uint32_t num_pages)
+gve_alloc_queue_page_list(const char *name, uint32_t num_pages, bool is_rx)
{
struct gve_queue_page_list *qpl;
const struct rte_memzone *mz;
- dma_addr_t page_bus;
uint32_t i;
qpl = rte_zmalloc("qpl struct", sizeof(struct gve_queue_page_list), 0);
if (!qpl)
return NULL;
- mz = rte_memzone_reserve_aligned(name, num_pages * PAGE_SIZE,
-rte_socket_id(),
-RTE_MEMZONE_IOVA_CONTIG, PAGE_SIZE);
- if (mz == NULL) {
- PMD_DRV_LOG(ERR, "Failed to alloc %s.", name);
- goto free_qpl_struct;
- }
qpl->page_buses = rte_zmalloc("qpl page buses",
num_pages * sizeof(dma_addr_t), 0);
if (qpl->page_buses == NULL) {
PMD_DRV_LOG(ERR, "Failed to alloc qpl page buses");
- goto free_qpl_memzone;
+ goto free_qpl_struct;
}
- page_bus = mz->iova;
- for (i = 0; i < num_pages; i++) {
- qpl->page_buses[i] = page_bus;
- page_bus += PAGE_SIZE;
+
+ qpl->qpl_bufs = rte_zmalloc("qpl bufs",
+ num_pages * sizeof(void *), 0);
+ if (qpl->qpl_bufs == NULL) {
+ PMD_DRV_LOG(ERR, "Failed to alloc qpl bufs");
+ goto free_qpl_page_buses;
}
- qpl->mz = mz;
+
+ mz = gve_alloc_using_mz(name, num_pages);
+ if (mz) {
+ qpl->mz = mz;
+
+ /* Populate the buffer addresses */
+ for (i = 0; i < num_pages; i++)
+ qpl->qpl_bufs[i] =
+ (void *)((uint64_t)(mz->addr) + i * PAGE_SIZE);
+
+ /* Populate the IOVA addresses */
+ for (i = 0; i < num_pages; i++)
+ qpl->page_buses[i] = mz->iova + i * PAGE_SIZE;
+ } else {
+ /* Allocate 4K size buffers.
+* Cannot use non-contiguous approach for tx fifo.
+*/
+ if (!is_rx)
+ goto free_qpl_page_bufs;
+
+ PMD_DRV_LOG(ERR, "Allocating bufs using malloc for %s ", name);
+ if (gve_alloc_using_malloc(qpl->qpl_bufs, num_pages))
+ goto free_qpl_page_bufs;
+
+ /* Populate the IOVA addresses */
+ for (i = 0; i < num_pages; i++)
+ qpl->page_buses[i] =
+ rte_malloc_virt2iova(qpl->qpl_bufs[i]);
+ }
+
qpl->num_entries = num_pages;
return qpl;
-free_qpl_memzone:
- rte_memzone_free(qpl->mz);
+free_qpl_page_bufs:
+ rte_free(qpl->qpl_bufs);
+free_qpl_page_buses:
+ rte_free(qpl->page_buses);
free_qpl_struct:
rte_free(qpl);
return NULL;
@@ -69,7 +127,18 @@ gve_free_queue_page_list(struct gve_queue_page_list *qpl)
if (qpl->mz) {
rte_memzone_free(qpl->mz);
qpl->mz
