* [net-next 10/13] net/mlx5e: Add ethtool control of ring params to VF representors
From: Saeed Mahameed @ 2018-10-01 18:56 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Gavi Teitz, Saeed Mahameed
In-Reply-To: <20181001185701.2944-1-saeedm@mellanox.com>
From: Gavi Teitz <gavi@mellanox.com>
Added ethtool control to the representors for setting and querying
the ring params.
Signed-off-by: Gavi Teitz <gavi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/en_rep.c | 18 ++++++++++++++++++
1 file changed, 18 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index be435e76d316..9264c3332aa6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -180,6 +180,22 @@ static int mlx5e_rep_get_sset_count(struct net_device *dev, int sset)
}
}
+static void mlx5e_rep_get_ringparam(struct net_device *dev,
+ struct ethtool_ringparam *param)
+{
+ struct mlx5e_priv *priv = netdev_priv(dev);
+
+ mlx5e_ethtool_get_ringparam(priv, param);
+}
+
+static int mlx5e_rep_set_ringparam(struct net_device *dev,
+ struct ethtool_ringparam *param)
+{
+ struct mlx5e_priv *priv = netdev_priv(dev);
+
+ return mlx5e_ethtool_set_ringparam(priv, param);
+}
+
static int mlx5e_replace_rep_vport_rx_rule(struct mlx5e_priv *priv,
struct mlx5_flow_destination *dest)
{
@@ -260,6 +276,8 @@ static const struct ethtool_ops mlx5e_rep_ethtool_ops = {
.get_strings = mlx5e_rep_get_strings,
.get_sset_count = mlx5e_rep_get_sset_count,
.get_ethtool_stats = mlx5e_rep_get_ethtool_stats,
+ .get_ringparam = mlx5e_rep_get_ringparam,
+ .set_ringparam = mlx5e_rep_set_ringparam,
.get_channels = mlx5e_rep_get_channels,
.set_channels = mlx5e_rep_set_channels,
.get_rxfh_key_size = mlx5e_rep_get_rxfh_key_size,
--
2.17.1
^ permalink raw reply related
* [net-next 12/13] net/mlx5e: Allow reporting of checksum unnecessary
From: Saeed Mahameed @ 2018-10-01 18:57 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20181001185701.2944-1-saeedm@mellanox.com>
From: Or Gerlitz <ogerlitz@mellanox.com>
Currently we practically never report checksum unnecessary, because
for all IP packets we take the checksum complete path.
Enable non-default runs with reprorting checksum unnecessary, using
an ethtool private flag. This can be useful for performance evals
and other explorations.
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 2 ++
.../ethernet/mellanox/mlx5/core/en_ethtool.c | 28 +++++++++++++++++++
.../net/ethernet/mellanox/mlx5/core/en_main.c | 4 +++
.../net/ethernet/mellanox/mlx5/core/en_rx.c | 3 ++
4 files changed, 37 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index b3bd79833517..ef7a44eb9adb 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -209,6 +209,7 @@ enum mlx5e_priv_flag {
MLX5E_PFLAG_TX_CQE_BASED_MODER = (1 << 1),
MLX5E_PFLAG_RX_CQE_COMPRESS = (1 << 2),
MLX5E_PFLAG_RX_STRIDING_RQ = (1 << 3),
+ MLX5E_PFLAG_RX_NO_CSUM_COMPLETE = (1 << 4),
};
#define MLX5E_SET_PFLAG(params, pflag, enable) \
@@ -290,6 +291,7 @@ struct mlx5e_dcbx_dp {
enum {
MLX5E_RQ_STATE_ENABLED,
MLX5E_RQ_STATE_AM,
+ MLX5E_RQ_STATE_NO_CSUM_COMPLETE,
};
struct mlx5e_cq {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index 33dafd8638b1..c86fd770c463 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -140,6 +140,7 @@ static const char mlx5e_priv_flags[][ETH_GSTRING_LEN] = {
"tx_cqe_moder",
"rx_cqe_compress",
"rx_striding_rq",
+ "rx_no_csum_complete",
};
int mlx5e_ethtool_get_sset_count(struct mlx5e_priv *priv, int sset)
@@ -1531,6 +1532,27 @@ static int set_pflag_rx_striding_rq(struct net_device *netdev, bool enable)
return 0;
}
+static int set_pflag_rx_no_csum_complete(struct net_device *netdev, bool enable)
+{
+ struct mlx5e_priv *priv = netdev_priv(netdev);
+ struct mlx5e_channels *channels = &priv->channels;
+ struct mlx5e_channel *c;
+ int i;
+
+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+ return 0;
+
+ for (i = 0; i < channels->num; i++) {
+ c = channels->c[i];
+ if (enable)
+ __set_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &c->rq.state);
+ else
+ __clear_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &c->rq.state);
+ }
+
+ return 0;
+}
+
static int mlx5e_handle_pflag(struct net_device *netdev,
u32 wanted_flags,
enum mlx5e_priv_flag flag,
@@ -1582,6 +1604,12 @@ static int mlx5e_set_priv_flags(struct net_device *netdev, u32 pflags)
err = mlx5e_handle_pflag(netdev, pflags,
MLX5E_PFLAG_RX_STRIDING_RQ,
set_pflag_rx_striding_rq);
+ if (err)
+ goto out;
+
+ err = mlx5e_handle_pflag(netdev, pflags,
+ MLX5E_PFLAG_RX_NO_CSUM_COMPLETE,
+ set_pflag_rx_no_csum_complete);
out:
mutex_unlock(&priv->state_lock);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 6086f874c7bf..35aca9a8e3d6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -929,6 +929,9 @@ static int mlx5e_open_rq(struct mlx5e_channel *c,
if (params->rx_dim_enabled)
__set_bit(MLX5E_RQ_STATE_AM, &c->rq.state);
+ if (params->pflags & MLX5E_PFLAG_RX_NO_CSUM_COMPLETE)
+ __set_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &c->rq.state);
+
return 0;
err_destroy_rq:
@@ -4528,6 +4531,7 @@ void mlx5e_build_nic_params(struct mlx5_core_dev *mdev,
params->rx_cqe_compress_def = slow_pci_heuristic(mdev);
MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_CQE_COMPRESS, params->rx_cqe_compress_def);
+ MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_NO_CSUM_COMPLETE, false);
/* RQ */
mlx5e_build_rq_params(mdev, params);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index 5a43cbf9103f..f19067c94272 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -782,6 +782,9 @@ static inline void mlx5e_handle_csum(struct net_device *netdev,
return;
}
+ if (unlikely(test_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &rq->state)))
+ goto csum_unnecessary;
+
if (likely(is_last_ethertype_ip(skb, &network_depth, &proto))) {
if (unlikely(get_ip_proto(skb, proto) == IPPROTO_SCTP))
goto csum_unnecessary;
--
2.17.1
^ permalink raw reply related
* [net-next 13/13] net/mlx5: Cache the system image guid
From: Saeed Mahameed @ 2018-10-01 18:57 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Alaa Hleihel, Saeed Mahameed
In-Reply-To: <20181001185701.2944-1-saeedm@mellanox.com>
From: Alaa Hleihel <alaa@mellanox.com>
The system image guid is a read-only field which is used by the TC
offloads code to determine if two mlx5 devices belong to the same
ASIC while adding flows.
Read this once and save it on the core device rather than querying each
time an offloaded flow is added.
Signed-off-by: Alaa Hleihel <alaa@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 4 ++--
drivers/net/ethernet/mellanox/mlx5/core/vport.c | 9 +++++++++
include/linux/mlx5/driver.h | 1 +
include/linux/mlx5/vport.h | 2 ++
4 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 9fed54017659..82723a0e509a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -2040,8 +2040,8 @@ static bool same_hw_devs(struct mlx5e_priv *priv, struct mlx5e_priv *peer_priv)
fmdev = priv->mdev;
pmdev = peer_priv->mdev;
- mlx5_query_nic_vport_system_image_guid(fmdev, &fsystem_guid);
- mlx5_query_nic_vport_system_image_guid(pmdev, &psystem_guid);
+ fsystem_guid = mlx5_query_nic_system_image_guid(fmdev);
+ psystem_guid = mlx5_query_nic_system_image_guid(pmdev);
return (fsystem_guid == psystem_guid);
}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/vport.c b/drivers/net/ethernet/mellanox/mlx5/core/vport.c
index b02af317c125..cfbea66b4879 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/vport.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/vport.c
@@ -1201,3 +1201,12 @@ int mlx5_nic_vport_unaffiliate_multiport(struct mlx5_core_dev *port_mdev)
return err;
}
EXPORT_SYMBOL_GPL(mlx5_nic_vport_unaffiliate_multiport);
+
+u64 mlx5_query_nic_system_image_guid(struct mlx5_core_dev *mdev)
+{
+ if (!mdev->sys_image_guid)
+ mlx5_query_nic_vport_system_image_guid(mdev, &mdev->sys_image_guid);
+
+ return mdev->sys_image_guid;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_system_image_guid);
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index ed73b51f6697..26a92462f4ce 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -838,6 +838,7 @@ struct mlx5_core_dev {
u32 fpga[MLX5_ST_SZ_DW(fpga_cap)];
u32 qcam[MLX5_ST_SZ_DW(qcam_reg)];
} caps;
+ u64 sys_image_guid;
phys_addr_t iseg_base;
struct mlx5_init_seg __iomem *iseg;
enum mlx5_device_state state;
diff --git a/include/linux/mlx5/vport.h b/include/linux/mlx5/vport.h
index 7e7c6dfcfb09..9c694808c212 100644
--- a/include/linux/mlx5/vport.h
+++ b/include/linux/mlx5/vport.h
@@ -121,4 +121,6 @@ int mlx5_nic_vport_query_local_lb(struct mlx5_core_dev *mdev, bool *status);
int mlx5_nic_vport_affiliate_multiport(struct mlx5_core_dev *master_mdev,
struct mlx5_core_dev *port_mdev);
int mlx5_nic_vport_unaffiliate_multiport(struct mlx5_core_dev *port_mdev);
+
+u64 mlx5_query_nic_system_image_guid(struct mlx5_core_dev *mdev);
#endif /* __MLX5_VPORT_H__ */
--
2.17.1
^ permalink raw reply related
* [PATCH bpf] bpf: don't accept cgroup local storage with zero value size
From: Roman Gushchin @ 2018-10-02 2:41 UTC (permalink / raw)
To: netdev@vger.kernel.org
Cc: linux-kernel@vger.kernel.org, Kernel Team, Roman Gushchin,
Alexei Starovoitov, Daniel Borkmann
Explicitly forbid creating cgroup local storage maps with zero value
size, as it makes no sense and might even cause a panic.
Reported-by: syzbot+18628320d3b14a5c459c@syzkaller.appspotmail.com
Signed-off-by: Roman Gushchin <guro@fb.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
---
kernel/bpf/local_storage.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/kernel/bpf/local_storage.c b/kernel/bpf/local_storage.c
index 944eb297465f..9f79956bff7d 100644
--- a/kernel/bpf/local_storage.c
+++ b/kernel/bpf/local_storage.c
@@ -261,6 +261,9 @@ static struct bpf_map *cgroup_storage_map_alloc(union bpf_attr *attr)
if (attr->key_size != sizeof(struct bpf_cgroup_storage_key))
return ERR_PTR(-EINVAL);
+ if (attr->value_size == 0)
+ return ERR_PTR(-EINVAL);
+
if (attr->value_size > PAGE_SIZE)
return ERR_PTR(-E2BIG);
--
2.17.1
^ permalink raw reply related
* Re: Query in Crypto framework
From: Herbert Xu @ 2018-10-02 2:42 UTC (permalink / raw)
To: Kalyani Akula
Cc: davem@davemloft.net, kstewart@linuxfoundation.org,
gregkh@linuxfoundation.org, tglx@linutronix.de,
pombredanne@nexb.com, linux-crypto@vger.kernel.org,
linux-kernel@vger.kernel.org, netdev@vger.kernel.org
In-Reply-To: <BN7PR02MB51249DB95E4CCBC5257C1F45AFEF0@BN7PR02MB5124.namprd02.prod.outlook.com>
On Mon, Oct 01, 2018 at 09:30:35AM +0000, Kalyani Akula wrote:
> Hi ,
>
> We wrote a Xilinx specific AES driver in the crypto framework, while testing we have seen one issue.
> We are trying to send a input data of size 1MB using sendmsg system call from the application, and we could see it is hanging in the af_alg_sendmsg.
> Our analysis is, it is waiting for the availability of the writable memory inside af_alg_wait_for_wmem (entered into infinite loop).
>
> Is there any limitation on the maximum data that can sent via AF_ALG socket ?
When you make a send(2) system call, you're essentially allocating
kernel memory. For obvious reasons we don't allow unbounded
allocation of kernel memory from user-space.
In order to process a large amount of data, you should use chaining
for algorithms that support it. So instead of writing everything
at once, write a certain multiple of blocks and then read the partial
result before writing more.
For AEAD algorithms which do not support chaining, you can work
around this by adjusting the appropriate socket buffer limits.
Note that if you increase the socket buffer limits too much you will
run into system-wide limits which can also be adjusted under /proc.
Cheers,
--
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* Re: [PATCH bpf-next v2 0/5] xsk: fix bug when trying to use both copy and zero-copy mode
From: Jakub Kicinski @ 2018-10-01 20:31 UTC (permalink / raw)
To: Magnus Karlsson; +Cc: bjorn.topel, ast, daniel, netdev, Jesper Dangaard Brouer
In-Reply-To: <1538398297-14862-1-git-send-email-magnus.karlsson@intel.com>
On Mon, 1 Oct 2018 14:51:32 +0200, Magnus Karlsson wrote:
> Jakub, please take a look at your patches. The last one I had to
> change slightly to make it fit with the new interface
> xdp_get_umem_from_qid(). An added bonus with this function is that we,
> in the future, can also use it from the driver to get a umem, thus
> simplifying driver implementations (and later remove the umem from the
> NDO completely). Björn will mail patches, at a later point in time,
> using this in the i40e and ixgbe drivers, that removes a good chunk of
> code from the ZC implementations.
Nice, drivers which don't follow the prepare/commit model of handling
reconfigurations will benefit!
> I also made your code aware of Tx queues. If we create a socket that
> only has a Tx queue, then the queue id will refer to a Tx queue id
> only and could be larger than the available amount of Rx queues.
> Please take a look at it.
The semantics of Tx queue id are slightly unclear. To me XDP is
associated with Rx, so the qid in driver context can only refer to
Rx queue and its associated XDP Tx queue. It does not mean the Tx
queue stack uses, like it does for copy fallback. If one doesn't have
a Rx queue $id, there will be no associated XDP Tx queue $id (in all
drivers but Intel, and virtio, which use per-CPU Tx queues making TX
queue even more meaningless).
Its to be seen how others implement AF_XDP. My general feeling is
that we should only talk about Rx queues in context of driver XDP.
^ permalink raw reply
* Re: [PATCH net-next v6 03/23] zinc: ChaCha20 generic C implementation and selftest
From: Jason A. Donenfeld @ 2018-10-02 3:18 UTC (permalink / raw)
To: Herbert Xu
Cc: LKML, Netdev, Linux Crypto Mailing List, David Miller,
Greg Kroah-Hartman, Samuel Neves, Andrew Lutomirski,
Jean-Philippe Aumasson
In-Reply-To: <20181002031542.dbxu44gvqrughlgw@gondor.apana.org.au>
Hi Herbert,
On Tue, Oct 2, 2018 at 5:15 AM Herbert Xu <herbert@gondor.apana.org.au> wrote:
> I don't have any problems with this interface.
>
> However, I'm very unhappy about the fact that this adds another
> copy of the chacha20 implementation to the kernel. Either this
> is better than the one that we already have or it isn't.
>
> So if it's better please replace the existing chacha20 C
> implementation with this or otherwise just adapt the existing
> implementation to this interface.
>
> If we're replacing the existing one then you can either replace
> it directly within this patch or in a subsequent patch but please
> do it in the same series because otherwise we could wind up with
> two implementations for years to come.
Indeed this series takes care of removing the other implementation later on.
In case it helps to see things visualized, here's the layout of the
whole series:
https://git.kernel.org/pub/scm/linux/kernel/git/zx2c4/linux.git/log/?h=jd/wireguard
Note that comments on this are probably a bit more useful than on the
version posted to the mailing list, since I've already started
incorporating quite a bit of feedback from v6, and so v7 will be
posted sometime later containing these.
Jason
^ permalink raw reply
* Re: [PATCH net-next v6 04/23] zinc: ChaCha20 x86_64 implementation
From: Herbert Xu @ 2018-10-02 3:18 UTC (permalink / raw)
To: Jason A. Donenfeld
Cc: linux-kernel, netdev, linux-crypto, davem, gregkh, Jason, sneves,
luto, jeanphilippe.aumasson, appro, tglx, mingo, x86
In-Reply-To: <20180925145622.29959-5-Jason@zx2c4.com>
Jason A. Donenfeld <Jason@zx2c4.com> wrote:
> This provides SSSE3, AVX-2, AVX-512F, and AVX-512VL implementations for
> ChaCha20. The AVX-512F implementation is disabled on Skylake, due to
> throttling, and the VL ymm implementation is used instead. These come
> from Andy Polyakov's implementation, with the following modifications
> from Samuel Neves:
>
> - Some cosmetic changes, like renaming labels to .Lname, constants,
> and other Linux conventions.
>
> - CPU feature checking is done in C by the glue code, so that has been
> removed from the assembly.
>
> - Eliminate translating certain instructions, such as pshufb, palignr,
> vprotd, etc, to .byte directives. This is meant for compatibility
> with ancient toolchains, but presumably it is unnecessary here,
> since the build system already does checks on what GNU as can
> assemble.
>
> - When aligning the stack, the original code was saving %rsp to %r9.
> To keep objtool happy, we use instead the DRAP idiom to save %rsp
> to %r10:
>
> leaq 8(%rsp),%r10
> ... code here ...
> leaq -8(%r10),%rsp
>
> - The original code assumes the stack comes aligned to 16 bytes. This
> is not necessarily the case, and to avoid crashes,
> `andq $-alignment, %rsp` was added in the prolog of a few functions.
>
> - The original hardcodes returns as .byte 0xf3,0xc3, aka "rep ret".
> We replace this by "ret". "rep ret" was meant to help with AMD K8
> chips, cf. http://repzret.org/p/repzret. It makes no sense to
> continue to use this kludge for code that won't even run on ancient
> AMD chips.
>
> While this is CRYPTOGAMS code, the originating code for this happens to
> be the same as OpenSSL's commit cded951378069a478391843f5f8653c1eb5128da
>
> Cycle counts on a Core i7 6700HQ using the AVX-2 codepath:
>
> size old new
> ---- ---- ----
> 0 62 52
What is the old column? Is it the existing x86-64 implementation
in the kernel or something else? This needs to be made clear in the
patch description.
The same comment from the previous patch also stands here. Please
ensure that we only have one copy of chacha x86-64 code in the kernel.
Either replace the existing one right here or in a follow-up patch
in the same series.
Thanks,
--
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* Re: [PATCH net-next v6 04/23] zinc: ChaCha20 x86_64 implementation
From: Jason A. Donenfeld @ 2018-10-02 3:20 UTC (permalink / raw)
To: Herbert Xu
Cc: LKML, Netdev, Linux Crypto Mailing List, David Miller,
Greg Kroah-Hartman, Samuel Neves, Andrew Lutomirski,
Jean-Philippe Aumasson, Andy Polyakov, Thomas Gleixner, mingo,
X86 ML
In-Reply-To: <20181002031848.lyxrergyyxyc5cwi@gondor.apana.org.au>
Hi Herbert,
On Tue, Oct 2, 2018 at 5:18 AM Herbert Xu <herbert@gondor.apana.org.au> wrote:
> > size old new
> > ---- ---- ----
> > 0 62 52
>
> What is the old column? Is it the existing x86-64 implementation
> in the kernel or something else? This needs to be made clear in the
> patch description.
Good idea. I'll make that more clear.
> The same comment from the previous patch also stands here. Please
> ensure that we only have one copy of chacha x86-64 code in the kernel.
> Either replace the existing one right here or in a follow-up patch
> in the same series.
Yes indeed this happens within this same series.
Jason
^ permalink raw reply
* Re: [PATCH net-next v6 21/23] crypto: port ChaCha20 to Zinc
From: Herbert Xu @ 2018-10-02 3:26 UTC (permalink / raw)
To: Jason A. Donenfeld
Cc: linux-kernel, netdev, linux-crypto, davem, gregkh, Jason, sneves,
luto, jeanphilippe.aumasson, ebiggers
In-Reply-To: <20180925145622.29959-22-Jason@zx2c4.com>
Jason A. Donenfeld <Jason@zx2c4.com> wrote:
> Now that ChaCha20 is in Zinc, we can have the crypto API code simply
> call into it. The crypto API expects to have a stored key per instance
> and independent nonces, so we follow suite and store the key and
> initialize the nonce independently.
>
> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
> Cc: Samuel Neves <sneves@dei.uc.pt>
> Cc: Andy Lutomirski <luto@kernel.org>
> Cc: Greg KH <gregkh@linuxfoundation.org>
> Cc: Jean-Philippe Aumasson <jeanphilippe.aumasson@gmail.com>
> Cc: Eric Biggers <ebiggers@google.com>
> ---
> arch/arm/configs/exynos_defconfig | 1 -
> arch/arm/configs/multi_v7_defconfig | 1 -
> arch/arm/configs/omap2plus_defconfig | 1 -
> arch/arm/crypto/Kconfig | 6 -
> arch/arm/crypto/Makefile | 2 -
> arch/arm/crypto/chacha20-neon-core.S | 521 --------------------
> arch/arm/crypto/chacha20-neon-glue.c | 127 -----
> arch/arm64/configs/defconfig | 1 -
> arch/arm64/crypto/Kconfig | 6 -
> arch/arm64/crypto/Makefile | 3 -
> arch/arm64/crypto/chacha20-neon-core.S | 450 -----------------
> arch/arm64/crypto/chacha20-neon-glue.c | 133 -----
> arch/x86/crypto/Makefile | 3 -
> arch/x86/crypto/chacha20-avx2-x86_64.S | 448 -----------------
> arch/x86/crypto/chacha20-ssse3-x86_64.S | 630 ------------------------
> arch/x86/crypto/chacha20_glue.c | 146 ------
> crypto/Kconfig | 17 +-
> crypto/Makefile | 2 +-
> crypto/chacha20_generic.c | 136 -----
> crypto/chacha20_zinc.c | 90 ++++
> crypto/chacha20poly1305.c | 8 +-
> include/crypto/chacha20.h | 12 -
> 22 files changed, 96 insertions(+), 2648 deletions(-)
> delete mode 100644 arch/arm/crypto/chacha20-neon-core.S
> delete mode 100644 arch/arm/crypto/chacha20-neon-glue.c
> delete mode 100644 arch/arm64/crypto/chacha20-neon-core.S
> delete mode 100644 arch/arm64/crypto/chacha20-neon-glue.c
> delete mode 100644 arch/x86/crypto/chacha20-avx2-x86_64.S
> delete mode 100644 arch/x86/crypto/chacha20-ssse3-x86_64.S
> delete mode 100644 arch/x86/crypto/chacha20_glue.c
> delete mode 100644 crypto/chacha20_generic.c
> create mode 100644 crypto/chacha20_zinc.c
Oh nice, so you did the conversion of the existing crypto code.
I presume someone has done the numbers and verified that there
is no performance regression?
If so it would be good to include those numbers somewhere in this
submission (within a patch description so that it goes into git).
Thanks,
--
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* Re: [PATCH net-next v6 00/23] WireGuard: Secure Network Tunnel
From: Herbert Xu @ 2018-10-02 3:39 UTC (permalink / raw)
To: Ard Biesheuvel
Cc: Jason, ebiggers, linux-kernel, netdev, linux-crypto, davem,
gregkh
In-Reply-To: <CAKv+Gu-ZMR7zEAKv6_2tyD3OqnjVugOzKHzBEXW3pQs5W-SvLw@mail.gmail.com>
Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
> On 28 September 2018 at 07:46, Jason A. Donenfeld <Jason@zx2c4.com> wrote:
>
>> If an implementation enters Zinc, it will go through my tree. If it
>> enters the crypto API, it will go through Herbert's tree. If there
>> wind up being messy tree dependency and merge timing issues, I can
>> work this out in the usual way with Herbert. I'll be sure to discuss
>> these edge cases with him when we discuss. I think there's a way to
>> handle that with minimum friction.
>
> I would also strongly prefer that all crypto work is taken through
> Herbert's tree, so we have a coherent view of it before it goes
> upstream.
I agree. I don't have any problems with the zinc code living in
its own git tree. But any upstream merges should definitely go
through the crypto tree because the inherent ties between the two
code-base.
Thanks,
--
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* Re: [PATCH net-next v6 00/23] WireGuard: Secure Network Tunnel
From: Jason A. Donenfeld @ 2018-10-02 3:45 UTC (permalink / raw)
To: Herbert Xu
Cc: Ard Biesheuvel, Eric Biggers, LKML, Netdev,
Linux Crypto Mailing List, David Miller, Greg Kroah-Hartman
In-Reply-To: <20181002033908.324yhwqaohfsq65d@gondor.apana.org.au>
Hi Herbert,
On Tue, Oct 2, 2018 at 5:39 AM Herbert Xu <herbert@gondor.apana.org.au> wrote:
> > I would also strongly prefer that all crypto work is taken through
> > Herbert's tree, so we have a coherent view of it before it goes
> > upstream.
>
> I agree. I don't have any problems with the zinc code living in
> its own git tree. But any upstream merges should definitely go
> through the crypto tree because the inherent ties between the two
> code-base.
I can send you pull requests then if there are development cycles when
there are in fact relations between the two trees. I'll update the
commit message describing Zinc to include this.
Jason
^ permalink raw reply
* Re: [PATCH net-next v6 00/23] WireGuard: Secure Network Tunnel
From: Herbert Xu @ 2018-10-02 3:49 UTC (permalink / raw)
To: Jason A. Donenfeld
Cc: Ard Biesheuvel, Eric Biggers, LKML, Netdev,
Linux Crypto Mailing List, David Miller, Greg Kroah-Hartman
In-Reply-To: <CAHmME9piX1kbUhOBcUKVdecr45EjxuruxYuMTepfXVvRrkVmyg@mail.gmail.com>
Hi Jason:
On Tue, Oct 02, 2018 at 05:45:20AM +0200, Jason A. Donenfeld wrote:
> Hi Herbert,
>
> > I agree. I don't have any problems with the zinc code living in
> > its own git tree. But any upstream merges should definitely go
> > through the crypto tree because the inherent ties between the two
> > code-base.
>
> I can send you pull requests then if there are development cycles when
> there are in fact relations between the two trees. I'll update the
> commit message describing Zinc to include this.
I think all merges should go through the crypto tree so that the
kernel crypto community has oversight over the crypto implementations.
Cheers,
--
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* [net-next 0/8][pull request] 100GbE Intel Wired LAN Driver Updates 2018-10-01
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem; +Cc: Jeff Kirsher, netdev, nhorman, sassmann, jogreene
This series contains updates to ice driver only.
Anirudh provides several changes to "prep" the driver for upcoming
features. Specifically, the functions that are used for PF VSI/netdev
setup will also be used in SR-IOV support and to allow the reuse of
these functions, code needs to move.
Dave provides the only other change in the series, updates the driver to
protect the reset patch in its entirety. This is done by adding the
various bit checks to determine if a reset is scheduled/initiated and
whether it came from the software or firmware.
The following are changes since commit 804fe108fc92e591ddfe9447e7fb4691ed16daee:
openvswitch: Use correct reply values in datapath and vport ops
and are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue 100GbE
Anirudh Venkataramanan (7):
ice: Move common functions out of ice_main.c part 1/7
ice: Move common functions out of ice_main.c part 2/7
ice: Move common functions out of ice_main.c part 3/7
ice: Move common functions out of ice_main.c part 4/7
ice: Move common functions out of ice_main.c part 5/7
ice: Move common functions out of ice_main.c part 6/7
ice: Move common functions out of ice_main.c part 7/7
Dave Ertman (1):
ice: Change pf state behavior to protect reset path
drivers/net/ethernet/intel/ice/Makefile | 1 +
drivers/net/ethernet/intel/ice/ice.h | 2 +-
drivers/net/ethernet/intel/ice/ice_common.c | 61 +
drivers/net/ethernet/intel/ice/ice_common.h | 4 +
drivers/net/ethernet/intel/ice/ice_lib.c | 2379 +++++++++++++++
drivers/net/ethernet/intel/ice/ice_lib.h | 74 +
drivers/net/ethernet/intel/ice/ice_main.c | 2901 ++-----------------
7 files changed, 2773 insertions(+), 2649 deletions(-)
create mode 100644 drivers/net/ethernet/intel/ice/ice_lib.c
create mode 100644 drivers/net/ethernet/intel/ice/ice_lib.h
--
2.17.1
^ permalink raw reply
* [net-next 2/8] ice: Move common functions out of ice_main.c part 2/7
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem
Cc: Anirudh Venkataramanan, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <20181001211431.14325-1-jeffrey.t.kirsher@intel.com>
From: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
This patch continues the code move out of ice_main.c
The following top level functions (and related dependency functions) were
moved to ice_lib.c:
ice_vsi_start_rx_rings
ice_vsi_stop_rx_rings
ice_vsi_stop_tx_rings
ice_vsi_cfg_rxqs
ice_vsi_cfg_txqs
ice_vsi_cfg_msix
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_lib.c | 491 ++++++++++++++++++++
drivers/net/ethernet/intel/ice/ice_lib.h | 13 +
drivers/net/ethernet/intel/ice/ice_main.c | 541 +---------------------
3 files changed, 526 insertions(+), 519 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 1cf4dca12495..06a54d79fba8 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -4,6 +4,227 @@
#include "ice.h"
#include "ice_lib.h"
+/**
+ * ice_setup_rx_ctx - Configure a receive ring context
+ * @ring: The Rx ring to configure
+ *
+ * Configure the Rx descriptor ring in RLAN context.
+ */
+static int ice_setup_rx_ctx(struct ice_ring *ring)
+{
+ struct ice_vsi *vsi = ring->vsi;
+ struct ice_hw *hw = &vsi->back->hw;
+ u32 rxdid = ICE_RXDID_FLEX_NIC;
+ struct ice_rlan_ctx rlan_ctx;
+ u32 regval;
+ u16 pf_q;
+ int err;
+
+ /* what is RX queue number in global space of 2K Rx queues */
+ pf_q = vsi->rxq_map[ring->q_index];
+
+ /* clear the context structure first */
+ memset(&rlan_ctx, 0, sizeof(rlan_ctx));
+
+ rlan_ctx.base = ring->dma >> 7;
+
+ rlan_ctx.qlen = ring->count;
+
+ /* Receive Packet Data Buffer Size.
+ * The Packet Data Buffer Size is defined in 128 byte units.
+ */
+ rlan_ctx.dbuf = vsi->rx_buf_len >> ICE_RLAN_CTX_DBUF_S;
+
+ /* use 32 byte descriptors */
+ rlan_ctx.dsize = 1;
+
+ /* Strip the Ethernet CRC bytes before the packet is posted to host
+ * memory.
+ */
+ rlan_ctx.crcstrip = 1;
+
+ /* L2TSEL flag defines the reported L2 Tags in the receive descriptor */
+ rlan_ctx.l2tsel = 1;
+
+ rlan_ctx.dtype = ICE_RX_DTYPE_NO_SPLIT;
+ rlan_ctx.hsplit_0 = ICE_RLAN_RX_HSPLIT_0_NO_SPLIT;
+ rlan_ctx.hsplit_1 = ICE_RLAN_RX_HSPLIT_1_NO_SPLIT;
+
+ /* This controls whether VLAN is stripped from inner headers
+ * The VLAN in the inner L2 header is stripped to the receive
+ * descriptor if enabled by this flag.
+ */
+ rlan_ctx.showiv = 0;
+
+ /* Max packet size for this queue - must not be set to a larger value
+ * than 5 x DBUF
+ */
+ rlan_ctx.rxmax = min_t(u16, vsi->max_frame,
+ ICE_MAX_CHAINED_RX_BUFS * vsi->rx_buf_len);
+
+ /* Rx queue threshold in units of 64 */
+ rlan_ctx.lrxqthresh = 1;
+
+ /* Enable Flexible Descriptors in the queue context which
+ * allows this driver to select a specific receive descriptor format
+ */
+ regval = rd32(hw, QRXFLXP_CNTXT(pf_q));
+ regval |= (rxdid << QRXFLXP_CNTXT_RXDID_IDX_S) &
+ QRXFLXP_CNTXT_RXDID_IDX_M;
+
+ /* increasing context priority to pick up profile id;
+ * default is 0x01; setting to 0x03 to ensure profile
+ * is programming if prev context is of same priority
+ */
+ regval |= (0x03 << QRXFLXP_CNTXT_RXDID_PRIO_S) &
+ QRXFLXP_CNTXT_RXDID_PRIO_M;
+
+ wr32(hw, QRXFLXP_CNTXT(pf_q), regval);
+
+ /* Absolute queue number out of 2K needs to be passed */
+ err = ice_write_rxq_ctx(hw, &rlan_ctx, pf_q);
+ if (err) {
+ dev_err(&vsi->back->pdev->dev,
+ "Failed to set LAN Rx queue context for absolute Rx queue %d error: %d\n",
+ pf_q, err);
+ return -EIO;
+ }
+
+ /* init queue specific tail register */
+ ring->tail = hw->hw_addr + QRX_TAIL(pf_q);
+ writel(0, ring->tail);
+ ice_alloc_rx_bufs(ring, ICE_DESC_UNUSED(ring));
+
+ return 0;
+}
+
+/**
+ * ice_setup_tx_ctx - setup a struct ice_tlan_ctx instance
+ * @ring: The Tx ring to configure
+ * @tlan_ctx: Pointer to the Tx LAN queue context structure to be initialized
+ * @pf_q: queue index in the PF space
+ *
+ * Configure the Tx descriptor ring in TLAN context.
+ */
+static void
+ice_setup_tx_ctx(struct ice_ring *ring, struct ice_tlan_ctx *tlan_ctx, u16 pf_q)
+{
+ struct ice_vsi *vsi = ring->vsi;
+ struct ice_hw *hw = &vsi->back->hw;
+
+ tlan_ctx->base = ring->dma >> ICE_TLAN_CTX_BASE_S;
+
+ tlan_ctx->port_num = vsi->port_info->lport;
+
+ /* Transmit Queue Length */
+ tlan_ctx->qlen = ring->count;
+
+ /* PF number */
+ tlan_ctx->pf_num = hw->pf_id;
+
+ /* queue belongs to a specific VSI type
+ * VF / VM index should be programmed per vmvf_type setting:
+ * for vmvf_type = VF, it is VF number between 0-256
+ * for vmvf_type = VM, it is VM number between 0-767
+ * for PF or EMP this field should be set to zero
+ */
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ tlan_ctx->vmvf_type = ICE_TLAN_CTX_VMVF_TYPE_PF;
+ break;
+ default:
+ return;
+ }
+
+ /* make sure the context is associated with the right VSI */
+ tlan_ctx->src_vsi = vsi->vsi_num;
+
+ tlan_ctx->tso_ena = ICE_TX_LEGACY;
+ tlan_ctx->tso_qnum = pf_q;
+
+ /* Legacy or Advanced Host Interface:
+ * 0: Advanced Host Interface
+ * 1: Legacy Host Interface
+ */
+ tlan_ctx->legacy_int = ICE_TX_LEGACY;
+}
+
+/**
+ * ice_pf_rxq_wait - Wait for a PF's Rx queue to be enabled or disabled
+ * @pf: the PF being configured
+ * @pf_q: the PF queue
+ * @ena: enable or disable state of the queue
+ *
+ * This routine will wait for the given Rx queue of the PF to reach the
+ * enabled or disabled state.
+ * Returns -ETIMEDOUT in case of failing to reach the requested state after
+ * multiple retries; else will return 0 in case of success.
+ */
+static int ice_pf_rxq_wait(struct ice_pf *pf, int pf_q, bool ena)
+{
+ int i;
+
+ for (i = 0; i < ICE_Q_WAIT_RETRY_LIMIT; i++) {
+ u32 rx_reg = rd32(&pf->hw, QRX_CTRL(pf_q));
+
+ if (ena == !!(rx_reg & QRX_CTRL_QENA_STAT_M))
+ break;
+
+ usleep_range(10, 20);
+ }
+ if (i >= ICE_Q_WAIT_RETRY_LIMIT)
+ return -ETIMEDOUT;
+
+ return 0;
+}
+
+/**
+ * ice_vsi_ctrl_rx_rings - Start or stop a VSI's Rx rings
+ * @vsi: the VSI being configured
+ * @ena: start or stop the Rx rings
+ */
+static int ice_vsi_ctrl_rx_rings(struct ice_vsi *vsi, bool ena)
+{
+ struct ice_pf *pf = vsi->back;
+ struct ice_hw *hw = &pf->hw;
+ int i, j, ret = 0;
+
+ for (i = 0; i < vsi->num_rxq; i++) {
+ int pf_q = vsi->rxq_map[i];
+ u32 rx_reg;
+
+ for (j = 0; j < ICE_Q_WAIT_MAX_RETRY; j++) {
+ rx_reg = rd32(hw, QRX_CTRL(pf_q));
+ if (((rx_reg >> QRX_CTRL_QENA_REQ_S) & 1) ==
+ ((rx_reg >> QRX_CTRL_QENA_STAT_S) & 1))
+ break;
+ usleep_range(1000, 2000);
+ }
+
+ /* Skip if the queue is already in the requested state */
+ if (ena == !!(rx_reg & QRX_CTRL_QENA_STAT_M))
+ continue;
+
+ /* turn on/off the queue */
+ if (ena)
+ rx_reg |= QRX_CTRL_QENA_REQ_M;
+ else
+ rx_reg &= ~QRX_CTRL_QENA_REQ_M;
+ wr32(hw, QRX_CTRL(pf_q), rx_reg);
+
+ /* wait for the change to finish */
+ ret = ice_pf_rxq_wait(pf, pf_q, ena);
+ if (ret) {
+ dev_err(&pf->pdev->dev,
+ "VSI idx %d Rx ring %d %sable timeout\n",
+ vsi->idx, pf_q, (ena ? "en" : "dis"));
+ break;
+ }
+ }
+
+ return ret;
+}
+
/**
* ice_add_mac_to_list - Add a mac address filter entry to the list
* @vsi: the VSI to be forwarded to
@@ -185,6 +406,174 @@ int ice_vsi_kill_vlan(struct ice_vsi *vsi, u16 vid)
return status;
}
+/**
+ * ice_vsi_cfg_rxqs - Configure the VSI for Rx
+ * @vsi: the VSI being configured
+ *
+ * Return 0 on success and a negative value on error
+ * Configure the Rx VSI for operation.
+ */
+int ice_vsi_cfg_rxqs(struct ice_vsi *vsi)
+{
+ int err = 0;
+ u16 i;
+
+ if (vsi->netdev && vsi->netdev->mtu > ETH_DATA_LEN)
+ vsi->max_frame = vsi->netdev->mtu +
+ ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN;
+ else
+ vsi->max_frame = ICE_RXBUF_2048;
+
+ vsi->rx_buf_len = ICE_RXBUF_2048;
+ /* set up individual rings */
+ for (i = 0; i < vsi->num_rxq && !err; i++)
+ err = ice_setup_rx_ctx(vsi->rx_rings[i]);
+
+ if (err) {
+ dev_err(&vsi->back->pdev->dev, "ice_setup_rx_ctx failed\n");
+ return -EIO;
+ }
+ return err;
+}
+
+/**
+ * ice_vsi_cfg_txqs - Configure the VSI for Tx
+ * @vsi: the VSI being configured
+ *
+ * Return 0 on success and a negative value on error
+ * Configure the Tx VSI for operation.
+ */
+int ice_vsi_cfg_txqs(struct ice_vsi *vsi)
+{
+ struct ice_aqc_add_tx_qgrp *qg_buf;
+ struct ice_aqc_add_txqs_perq *txq;
+ struct ice_pf *pf = vsi->back;
+ enum ice_status status;
+ u16 buf_len, i, pf_q;
+ int err = 0, tc = 0;
+ u8 num_q_grps;
+
+ buf_len = sizeof(struct ice_aqc_add_tx_qgrp);
+ qg_buf = devm_kzalloc(&pf->pdev->dev, buf_len, GFP_KERNEL);
+ if (!qg_buf)
+ return -ENOMEM;
+
+ if (vsi->num_txq > ICE_MAX_TXQ_PER_TXQG) {
+ err = -EINVAL;
+ goto err_cfg_txqs;
+ }
+ qg_buf->num_txqs = 1;
+ num_q_grps = 1;
+
+ /* set up and configure the Tx queues */
+ ice_for_each_txq(vsi, i) {
+ struct ice_tlan_ctx tlan_ctx = { 0 };
+
+ pf_q = vsi->txq_map[i];
+ ice_setup_tx_ctx(vsi->tx_rings[i], &tlan_ctx, pf_q);
+ /* copy context contents into the qg_buf */
+ qg_buf->txqs[0].txq_id = cpu_to_le16(pf_q);
+ ice_set_ctx((u8 *)&tlan_ctx, qg_buf->txqs[0].txq_ctx,
+ ice_tlan_ctx_info);
+
+ /* init queue specific tail reg. It is referred as transmit
+ * comm scheduler queue doorbell.
+ */
+ vsi->tx_rings[i]->tail = pf->hw.hw_addr + QTX_COMM_DBELL(pf_q);
+ status = ice_ena_vsi_txq(vsi->port_info, vsi->vsi_num, tc,
+ num_q_grps, qg_buf, buf_len, NULL);
+ if (status) {
+ dev_err(&vsi->back->pdev->dev,
+ "Failed to set LAN Tx queue context, error: %d\n",
+ status);
+ err = -ENODEV;
+ goto err_cfg_txqs;
+ }
+
+ /* Add Tx Queue TEID into the VSI Tx ring from the response
+ * This will complete configuring and enabling the queue.
+ */
+ txq = &qg_buf->txqs[0];
+ if (pf_q == le16_to_cpu(txq->txq_id))
+ vsi->tx_rings[i]->txq_teid =
+ le32_to_cpu(txq->q_teid);
+ }
+err_cfg_txqs:
+ devm_kfree(&pf->pdev->dev, qg_buf);
+ return err;
+}
+
+/**
+ * ice_vsi_cfg_msix - MSIX mode Interrupt Config in the HW
+ * @vsi: the VSI being configured
+ */
+void ice_vsi_cfg_msix(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ u16 vector = vsi->base_vector;
+ struct ice_hw *hw = &pf->hw;
+ u32 txq = 0, rxq = 0;
+ int i, q, itr;
+ u8 itr_gran;
+
+ for (i = 0; i < vsi->num_q_vectors; i++, vector++) {
+ struct ice_q_vector *q_vector = vsi->q_vectors[i];
+
+ itr_gran = hw->itr_gran_200;
+
+ if (q_vector->num_ring_rx) {
+ q_vector->rx.itr =
+ ITR_TO_REG(vsi->rx_rings[rxq]->rx_itr_setting,
+ itr_gran);
+ q_vector->rx.latency_range = ICE_LOW_LATENCY;
+ }
+
+ if (q_vector->num_ring_tx) {
+ q_vector->tx.itr =
+ ITR_TO_REG(vsi->tx_rings[txq]->tx_itr_setting,
+ itr_gran);
+ q_vector->tx.latency_range = ICE_LOW_LATENCY;
+ }
+ wr32(hw, GLINT_ITR(ICE_RX_ITR, vector), q_vector->rx.itr);
+ wr32(hw, GLINT_ITR(ICE_TX_ITR, vector), q_vector->tx.itr);
+
+ /* Both Transmit Queue Interrupt Cause Control register
+ * and Receive Queue Interrupt Cause control register
+ * expects MSIX_INDX field to be the vector index
+ * within the function space and not the absolute
+ * vector index across PF or across device.
+ * For SR-IOV VF VSIs queue vector index always starts
+ * with 1 since first vector index(0) is used for OICR
+ * in VF space. Since VMDq and other PF VSIs are within
+ * the PF function space, use the vector index that is
+ * tracked for this PF.
+ */
+ for (q = 0; q < q_vector->num_ring_tx; q++) {
+ u32 val;
+
+ itr = ICE_ITR_NONE;
+ val = QINT_TQCTL_CAUSE_ENA_M |
+ (itr << QINT_TQCTL_ITR_INDX_S) |
+ (vector << QINT_TQCTL_MSIX_INDX_S);
+ wr32(hw, QINT_TQCTL(vsi->txq_map[txq]), val);
+ txq++;
+ }
+
+ for (q = 0; q < q_vector->num_ring_rx; q++) {
+ u32 val;
+
+ itr = ICE_ITR_NONE;
+ val = QINT_RQCTL_CAUSE_ENA_M |
+ (itr << QINT_RQCTL_ITR_INDX_S) |
+ (vector << QINT_RQCTL_MSIX_INDX_S);
+ wr32(hw, QINT_RQCTL(vsi->rxq_map[rxq]), val);
+ rxq++;
+ }
+ }
+
+ ice_flush(hw);
+}
+
/**
* ice_vsi_manage_vlan_insertion - Manage VLAN insertion for the VSI for Tx
* @vsi: the VSI being changed
@@ -256,3 +645,105 @@ int ice_vsi_manage_vlan_stripping(struct ice_vsi *vsi, bool ena)
vsi->info.vlan_flags = ctxt.info.vlan_flags;
return 0;
}
+
+/**
+ * ice_vsi_start_rx_rings - start VSI's Rx rings
+ * @vsi: the VSI whose rings are to be started
+ *
+ * Returns 0 on success and a negative value on error
+ */
+int ice_vsi_start_rx_rings(struct ice_vsi *vsi)
+{
+ return ice_vsi_ctrl_rx_rings(vsi, true);
+}
+
+/**
+ * ice_vsi_stop_rx_rings - stop VSI's Rx rings
+ * @vsi: the VSI
+ *
+ * Returns 0 on success and a negative value on error
+ */
+int ice_vsi_stop_rx_rings(struct ice_vsi *vsi)
+{
+ return ice_vsi_ctrl_rx_rings(vsi, false);
+}
+
+/**
+ * ice_vsi_stop_tx_rings - Disable Tx rings
+ * @vsi: the VSI being configured
+ */
+int ice_vsi_stop_tx_rings(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ struct ice_hw *hw = &pf->hw;
+ enum ice_status status;
+ u32 *q_teids, val;
+ u16 *q_ids, i;
+ int err = 0;
+
+ if (vsi->num_txq > ICE_LAN_TXQ_MAX_QDIS)
+ return -EINVAL;
+
+ q_teids = devm_kcalloc(&pf->pdev->dev, vsi->num_txq, sizeof(*q_teids),
+ GFP_KERNEL);
+ if (!q_teids)
+ return -ENOMEM;
+
+ q_ids = devm_kcalloc(&pf->pdev->dev, vsi->num_txq, sizeof(*q_ids),
+ GFP_KERNEL);
+ if (!q_ids) {
+ err = -ENOMEM;
+ goto err_alloc_q_ids;
+ }
+
+ /* set up the Tx queue list to be disabled */
+ ice_for_each_txq(vsi, i) {
+ u16 v_idx;
+
+ if (!vsi->tx_rings || !vsi->tx_rings[i]) {
+ err = -EINVAL;
+ goto err_out;
+ }
+
+ q_ids[i] = vsi->txq_map[i];
+ q_teids[i] = vsi->tx_rings[i]->txq_teid;
+
+ /* clear cause_ena bit for disabled queues */
+ val = rd32(hw, QINT_TQCTL(vsi->tx_rings[i]->reg_idx));
+ val &= ~QINT_TQCTL_CAUSE_ENA_M;
+ wr32(hw, QINT_TQCTL(vsi->tx_rings[i]->reg_idx), val);
+
+ /* software is expected to wait for 100 ns */
+ ndelay(100);
+
+ /* trigger a software interrupt for the vector associated to
+ * the queue to schedule NAPI handler
+ */
+ v_idx = vsi->tx_rings[i]->q_vector->v_idx;
+ wr32(hw, GLINT_DYN_CTL(vsi->base_vector + v_idx),
+ GLINT_DYN_CTL_SWINT_TRIG_M | GLINT_DYN_CTL_INTENA_MSK_M);
+ }
+ status = ice_dis_vsi_txq(vsi->port_info, vsi->num_txq, q_ids, q_teids,
+ NULL);
+ /* if the disable queue command was exercised during an active reset
+ * flow, ICE_ERR_RESET_ONGOING is returned. This is not an error as
+ * the reset operation disables queues at the hardware level anyway.
+ */
+ if (status == ICE_ERR_RESET_ONGOING) {
+ dev_info(&pf->pdev->dev,
+ "Reset in progress. LAN Tx queues already disabled\n");
+ } else if (status) {
+ dev_err(&pf->pdev->dev,
+ "Failed to disable LAN Tx queues, error: %d\n",
+ status);
+ err = -ENODEV;
+ }
+
+err_out:
+ devm_kfree(&pf->pdev->dev, q_ids);
+
+err_alloc_q_ids:
+ devm_kfree(&pf->pdev->dev, q_teids);
+
+ return err;
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
index c10874d26eee..ad4257929b9b 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -13,6 +13,12 @@ void ice_free_fltr_list(struct device *dev, struct list_head *h);
void ice_update_eth_stats(struct ice_vsi *vsi);
+int ice_vsi_cfg_rxqs(struct ice_vsi *vsi);
+
+int ice_vsi_cfg_txqs(struct ice_vsi *vsi);
+
+void ice_vsi_cfg_msix(struct ice_vsi *vsi);
+
int ice_vsi_add_vlan(struct ice_vsi *vsi, u16 vid);
int ice_vsi_kill_vlan(struct ice_vsi *vsi, u16 vid);
@@ -20,4 +26,11 @@ int ice_vsi_kill_vlan(struct ice_vsi *vsi, u16 vid);
int ice_vsi_manage_vlan_insertion(struct ice_vsi *vsi);
int ice_vsi_manage_vlan_stripping(struct ice_vsi *vsi, bool ena);
+
+int ice_vsi_start_rx_rings(struct ice_vsi *vsi);
+
+int ice_vsi_stop_rx_rings(struct ice_vsi *vsi);
+
+int ice_vsi_stop_tx_rings(struct ice_vsi *vsi);
+
#endif /* !_ICE_LIB_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 2673430ec2e9..ececf3dabf7e 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1844,77 +1844,6 @@ static void ice_vsi_free_irq(struct ice_vsi *vsi)
}
}
-/**
- * ice_vsi_cfg_msix - MSIX mode Interrupt Config in the HW
- * @vsi: the VSI being configured
- */
-static void ice_vsi_cfg_msix(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- u16 vector = vsi->base_vector;
- struct ice_hw *hw = &pf->hw;
- u32 txq = 0, rxq = 0;
- int i, q, itr;
- u8 itr_gran;
-
- for (i = 0; i < vsi->num_q_vectors; i++, vector++) {
- struct ice_q_vector *q_vector = vsi->q_vectors[i];
-
- itr_gran = hw->itr_gran_200;
-
- if (q_vector->num_ring_rx) {
- q_vector->rx.itr =
- ITR_TO_REG(vsi->rx_rings[rxq]->rx_itr_setting,
- itr_gran);
- q_vector->rx.latency_range = ICE_LOW_LATENCY;
- }
-
- if (q_vector->num_ring_tx) {
- q_vector->tx.itr =
- ITR_TO_REG(vsi->tx_rings[txq]->tx_itr_setting,
- itr_gran);
- q_vector->tx.latency_range = ICE_LOW_LATENCY;
- }
- wr32(hw, GLINT_ITR(ICE_RX_ITR, vector), q_vector->rx.itr);
- wr32(hw, GLINT_ITR(ICE_TX_ITR, vector), q_vector->tx.itr);
-
- /* Both Transmit Queue Interrupt Cause Control register
- * and Receive Queue Interrupt Cause control register
- * expects MSIX_INDX field to be the vector index
- * within the function space and not the absolute
- * vector index across PF or across device.
- * For SR-IOV VF VSIs queue vector index always starts
- * with 1 since first vector index(0) is used for OICR
- * in VF space. Since VMDq and other PF VSIs are withtin
- * the PF function space, use the vector index thats
- * tracked for this PF.
- */
- for (q = 0; q < q_vector->num_ring_tx; q++) {
- u32 val;
-
- itr = ICE_TX_ITR;
- val = QINT_TQCTL_CAUSE_ENA_M |
- (itr << QINT_TQCTL_ITR_INDX_S) |
- (vector << QINT_TQCTL_MSIX_INDX_S);
- wr32(hw, QINT_TQCTL(vsi->txq_map[txq]), val);
- txq++;
- }
-
- for (q = 0; q < q_vector->num_ring_rx; q++) {
- u32 val;
-
- itr = ICE_RX_ITR;
- val = QINT_RQCTL_CAUSE_ENA_M |
- (itr << QINT_RQCTL_ITR_INDX_S) |
- (vector << QINT_RQCTL_MSIX_INDX_S);
- wr32(hw, QINT_RQCTL(vsi->rxq_map[rxq]), val);
- rxq++;
- }
- }
-
- ice_flush(hw);
-}
-
/**
* ice_ena_misc_vector - enable the non-queue interrupts
* @pf: board private structure
@@ -3966,248 +3895,6 @@ static int ice_restore_vlan(struct ice_vsi *vsi)
return err;
}
-/**
- * ice_setup_tx_ctx - setup a struct ice_tlan_ctx instance
- * @ring: The Tx ring to configure
- * @tlan_ctx: Pointer to the Tx LAN queue context structure to be initialized
- * @pf_q: queue index in the PF space
- *
- * Configure the Tx descriptor ring in TLAN context.
- */
-static void
-ice_setup_tx_ctx(struct ice_ring *ring, struct ice_tlan_ctx *tlan_ctx, u16 pf_q)
-{
- struct ice_vsi *vsi = ring->vsi;
- struct ice_hw *hw = &vsi->back->hw;
-
- tlan_ctx->base = ring->dma >> ICE_TLAN_CTX_BASE_S;
-
- tlan_ctx->port_num = vsi->port_info->lport;
-
- /* Transmit Queue Length */
- tlan_ctx->qlen = ring->count;
-
- /* PF number */
- tlan_ctx->pf_num = hw->pf_id;
-
- /* queue belongs to a specific VSI type
- * VF / VM index should be programmed per vmvf_type setting:
- * for vmvf_type = VF, it is VF number between 0-256
- * for vmvf_type = VM, it is VM number between 0-767
- * for PF or EMP this field should be set to zero
- */
- switch (vsi->type) {
- case ICE_VSI_PF:
- tlan_ctx->vmvf_type = ICE_TLAN_CTX_VMVF_TYPE_PF;
- break;
- default:
- return;
- }
-
- /* make sure the context is associated with the right VSI */
- tlan_ctx->src_vsi = vsi->vsi_num;
-
- tlan_ctx->tso_ena = ICE_TX_LEGACY;
- tlan_ctx->tso_qnum = pf_q;
-
- /* Legacy or Advanced Host Interface:
- * 0: Advanced Host Interface
- * 1: Legacy Host Interface
- */
- tlan_ctx->legacy_int = ICE_TX_LEGACY;
-}
-
-/**
- * ice_vsi_cfg_txqs - Configure the VSI for Tx
- * @vsi: the VSI being configured
- *
- * Return 0 on success and a negative value on error
- * Configure the Tx VSI for operation.
- */
-static int ice_vsi_cfg_txqs(struct ice_vsi *vsi)
-{
- struct ice_aqc_add_tx_qgrp *qg_buf;
- struct ice_aqc_add_txqs_perq *txq;
- struct ice_pf *pf = vsi->back;
- enum ice_status status;
- u16 buf_len, i, pf_q;
- int err = 0, tc = 0;
- u8 num_q_grps;
-
- buf_len = sizeof(struct ice_aqc_add_tx_qgrp);
- qg_buf = devm_kzalloc(&pf->pdev->dev, buf_len, GFP_KERNEL);
- if (!qg_buf)
- return -ENOMEM;
-
- if (vsi->num_txq > ICE_MAX_TXQ_PER_TXQG) {
- err = -EINVAL;
- goto err_cfg_txqs;
- }
- qg_buf->num_txqs = 1;
- num_q_grps = 1;
-
- /* set up and configure the tx queues */
- ice_for_each_txq(vsi, i) {
- struct ice_tlan_ctx tlan_ctx = { 0 };
-
- pf_q = vsi->txq_map[i];
- ice_setup_tx_ctx(vsi->tx_rings[i], &tlan_ctx, pf_q);
- /* copy context contents into the qg_buf */
- qg_buf->txqs[0].txq_id = cpu_to_le16(pf_q);
- ice_set_ctx((u8 *)&tlan_ctx, qg_buf->txqs[0].txq_ctx,
- ice_tlan_ctx_info);
-
- /* init queue specific tail reg. It is referred as transmit
- * comm scheduler queue doorbell.
- */
- vsi->tx_rings[i]->tail = pf->hw.hw_addr + QTX_COMM_DBELL(pf_q);
- status = ice_ena_vsi_txq(vsi->port_info, vsi->vsi_num, tc,
- num_q_grps, qg_buf, buf_len, NULL);
- if (status) {
- dev_err(&vsi->back->pdev->dev,
- "Failed to set LAN Tx queue context, error: %d\n",
- status);
- err = -ENODEV;
- goto err_cfg_txqs;
- }
-
- /* Add Tx Queue TEID into the VSI tx ring from the response
- * This will complete configuring and enabling the queue.
- */
- txq = &qg_buf->txqs[0];
- if (pf_q == le16_to_cpu(txq->txq_id))
- vsi->tx_rings[i]->txq_teid =
- le32_to_cpu(txq->q_teid);
- }
-err_cfg_txqs:
- devm_kfree(&pf->pdev->dev, qg_buf);
- return err;
-}
-
-/**
- * ice_setup_rx_ctx - Configure a receive ring context
- * @ring: The Rx ring to configure
- *
- * Configure the Rx descriptor ring in RLAN context.
- */
-static int ice_setup_rx_ctx(struct ice_ring *ring)
-{
- struct ice_vsi *vsi = ring->vsi;
- struct ice_hw *hw = &vsi->back->hw;
- u32 rxdid = ICE_RXDID_FLEX_NIC;
- struct ice_rlan_ctx rlan_ctx;
- u32 regval;
- u16 pf_q;
- int err;
-
- /* what is RX queue number in global space of 2K rx queues */
- pf_q = vsi->rxq_map[ring->q_index];
-
- /* clear the context structure first */
- memset(&rlan_ctx, 0, sizeof(rlan_ctx));
-
- rlan_ctx.base = ring->dma >> ICE_RLAN_BASE_S;
-
- rlan_ctx.qlen = ring->count;
-
- /* Receive Packet Data Buffer Size.
- * The Packet Data Buffer Size is defined in 128 byte units.
- */
- rlan_ctx.dbuf = vsi->rx_buf_len >> ICE_RLAN_CTX_DBUF_S;
-
- /* use 32 byte descriptors */
- rlan_ctx.dsize = 1;
-
- /* Strip the Ethernet CRC bytes before the packet is posted to host
- * memory.
- */
- rlan_ctx.crcstrip = 1;
-
- /* L2TSEL flag defines the reported L2 Tags in the receive descriptor */
- rlan_ctx.l2tsel = 1;
-
- rlan_ctx.dtype = ICE_RX_DTYPE_NO_SPLIT;
- rlan_ctx.hsplit_0 = ICE_RLAN_RX_HSPLIT_0_NO_SPLIT;
- rlan_ctx.hsplit_1 = ICE_RLAN_RX_HSPLIT_1_NO_SPLIT;
-
- /* This controls whether VLAN is stripped from inner headers
- * The VLAN in the inner L2 header is stripped to the receive
- * descriptor if enabled by this flag.
- */
- rlan_ctx.showiv = 0;
-
- /* Max packet size for this queue - must not be set to a larger value
- * than 5 x DBUF
- */
- rlan_ctx.rxmax = min_t(u16, vsi->max_frame,
- ICE_MAX_CHAINED_RX_BUFS * vsi->rx_buf_len);
-
- /* Rx queue threshold in units of 64 */
- rlan_ctx.lrxqthresh = 1;
-
- /* Enable Flexible Descriptors in the queue context which
- * allows this driver to select a specific receive descriptor format
- */
- regval = rd32(hw, QRXFLXP_CNTXT(pf_q));
- regval |= (rxdid << QRXFLXP_CNTXT_RXDID_IDX_S) &
- QRXFLXP_CNTXT_RXDID_IDX_M;
-
- /* increasing context priority to pick up profile id;
- * default is 0x01; setting to 0x03 to ensure profile
- * is programming if prev context is of same priority
- */
- regval |= (0x03 << QRXFLXP_CNTXT_RXDID_PRIO_S) &
- QRXFLXP_CNTXT_RXDID_PRIO_M;
-
- wr32(hw, QRXFLXP_CNTXT(pf_q), regval);
-
- /* Absolute queue number out of 2K needs to be passed */
- err = ice_write_rxq_ctx(hw, &rlan_ctx, pf_q);
- if (err) {
- dev_err(&vsi->back->pdev->dev,
- "Failed to set LAN Rx queue context for absolute Rx queue %d error: %d\n",
- pf_q, err);
- return -EIO;
- }
-
- /* init queue specific tail register */
- ring->tail = hw->hw_addr + QRX_TAIL(pf_q);
- writel(0, ring->tail);
- ice_alloc_rx_bufs(ring, ICE_DESC_UNUSED(ring));
-
- return 0;
-}
-
-/**
- * ice_vsi_cfg_rxqs - Configure the VSI for Rx
- * @vsi: the VSI being configured
- *
- * Return 0 on success and a negative value on error
- * Configure the Rx VSI for operation.
- */
-static int ice_vsi_cfg_rxqs(struct ice_vsi *vsi)
-{
- int err = 0;
- u16 i;
-
- if (vsi->netdev && vsi->netdev->mtu > ETH_DATA_LEN)
- vsi->max_frame = vsi->netdev->mtu +
- ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN;
- else
- vsi->max_frame = ICE_RXBUF_2048;
-
- vsi->rx_buf_len = ICE_RXBUF_2048;
- /* set up individual rings */
- for (i = 0; i < vsi->num_rxq && !err; i++)
- err = ice_setup_rx_ctx(vsi->rx_rings[i]);
-
- if (err) {
- dev_err(&vsi->back->pdev->dev, "ice_setup_rx_ctx failed\n");
- return -EIO;
- }
- return err;
-}
-
/**
* ice_vsi_cfg - Setup the VSI
* @vsi: the VSI being configured
@@ -4232,207 +3919,6 @@ static int ice_vsi_cfg(struct ice_vsi *vsi)
return err;
}
-/**
- * ice_vsi_stop_tx_rings - Disable Tx rings
- * @vsi: the VSI being configured
- */
-static int ice_vsi_stop_tx_rings(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- struct ice_hw *hw = &pf->hw;
- enum ice_status status;
- u32 *q_teids, val;
- u16 *q_ids, i;
- int err = 0;
-
- if (vsi->num_txq > ICE_LAN_TXQ_MAX_QDIS)
- return -EINVAL;
-
- q_teids = devm_kcalloc(&pf->pdev->dev, vsi->num_txq, sizeof(*q_teids),
- GFP_KERNEL);
- if (!q_teids)
- return -ENOMEM;
-
- q_ids = devm_kcalloc(&pf->pdev->dev, vsi->num_txq, sizeof(*q_ids),
- GFP_KERNEL);
- if (!q_ids) {
- err = -ENOMEM;
- goto err_alloc_q_ids;
- }
-
- /* set up the tx queue list to be disabled */
- ice_for_each_txq(vsi, i) {
- u16 v_idx;
-
- if (!vsi->tx_rings || !vsi->tx_rings[i]) {
- err = -EINVAL;
- goto err_out;
- }
-
- q_ids[i] = vsi->txq_map[i];
- q_teids[i] = vsi->tx_rings[i]->txq_teid;
-
- /* clear cause_ena bit for disabled queues */
- val = rd32(hw, QINT_TQCTL(vsi->tx_rings[i]->reg_idx));
- val &= ~QINT_TQCTL_CAUSE_ENA_M;
- wr32(hw, QINT_TQCTL(vsi->tx_rings[i]->reg_idx), val);
-
- /* software is expected to wait for 100 ns */
- ndelay(100);
-
- /* trigger a software interrupt for the vector associated to
- * the queue to schedule napi handler
- */
- v_idx = vsi->tx_rings[i]->q_vector->v_idx;
- wr32(hw, GLINT_DYN_CTL(vsi->base_vector + v_idx),
- GLINT_DYN_CTL_SWINT_TRIG_M | GLINT_DYN_CTL_INTENA_MSK_M);
- }
- status = ice_dis_vsi_txq(vsi->port_info, vsi->num_txq, q_ids, q_teids,
- NULL);
- /* if the disable queue command was exercised during an active reset
- * flow, ICE_ERR_RESET_ONGOING is returned. This is not an error as
- * the reset operation disables queues at the hardware level anyway.
- */
- if (status == ICE_ERR_RESET_ONGOING) {
- dev_dbg(&pf->pdev->dev,
- "Reset in progress. LAN Tx queues already disabled\n");
- } else if (status) {
- dev_err(&pf->pdev->dev,
- "Failed to disable LAN Tx queues, error: %d\n",
- status);
- err = -ENODEV;
- }
-
-err_out:
- devm_kfree(&pf->pdev->dev, q_ids);
-
-err_alloc_q_ids:
- devm_kfree(&pf->pdev->dev, q_teids);
-
- return err;
-}
-
-/**
- * ice_pf_rxq_wait - Wait for a PF's Rx queue to be enabled or disabled
- * @pf: the PF being configured
- * @pf_q: the PF queue
- * @ena: enable or disable state of the queue
- *
- * This routine will wait for the given Rx queue of the PF to reach the
- * enabled or disabled state.
- * Returns -ETIMEDOUT in case of failing to reach the requested state after
- * multiple retries; else will return 0 in case of success.
- */
-static int ice_pf_rxq_wait(struct ice_pf *pf, int pf_q, bool ena)
-{
- int i;
-
- for (i = 0; i < ICE_Q_WAIT_RETRY_LIMIT; i++) {
- u32 rx_reg = rd32(&pf->hw, QRX_CTRL(pf_q));
-
- if (ena == !!(rx_reg & QRX_CTRL_QENA_STAT_M))
- break;
-
- usleep_range(10, 20);
- }
- if (i >= ICE_Q_WAIT_RETRY_LIMIT)
- return -ETIMEDOUT;
-
- return 0;
-}
-
-/**
- * ice_vsi_ctrl_rx_rings - Start or stop a VSI's rx rings
- * @vsi: the VSI being configured
- * @ena: start or stop the rx rings
- */
-static int ice_vsi_ctrl_rx_rings(struct ice_vsi *vsi, bool ena)
-{
- struct ice_pf *pf = vsi->back;
- struct ice_hw *hw = &pf->hw;
- int i, j, ret = 0;
-
- for (i = 0; i < vsi->num_rxq; i++) {
- int pf_q = vsi->rxq_map[i];
- u32 rx_reg;
-
- for (j = 0; j < ICE_Q_WAIT_MAX_RETRY; j++) {
- rx_reg = rd32(hw, QRX_CTRL(pf_q));
- if (((rx_reg >> QRX_CTRL_QENA_REQ_S) & 1) ==
- ((rx_reg >> QRX_CTRL_QENA_STAT_S) & 1))
- break;
- usleep_range(1000, 2000);
- }
-
- /* Skip if the queue is already in the requested state */
- if (ena == !!(rx_reg & QRX_CTRL_QENA_STAT_M))
- continue;
-
- /* turn on/off the queue */
- if (ena)
- rx_reg |= QRX_CTRL_QENA_REQ_M;
- else
- rx_reg &= ~QRX_CTRL_QENA_REQ_M;
- wr32(hw, QRX_CTRL(pf_q), rx_reg);
-
- /* wait for the change to finish */
- ret = ice_pf_rxq_wait(pf, pf_q, ena);
- if (ret) {
- dev_err(&pf->pdev->dev,
- "VSI idx %d Rx ring %d %sable timeout\n",
- vsi->idx, pf_q, (ena ? "en" : "dis"));
- break;
- }
- }
-
- return ret;
-}
-
-/**
- * ice_vsi_start_rx_rings - start VSI's rx rings
- * @vsi: the VSI whose rings are to be started
- *
- * Returns 0 on success and a negative value on error
- */
-static int ice_vsi_start_rx_rings(struct ice_vsi *vsi)
-{
- return ice_vsi_ctrl_rx_rings(vsi, true);
-}
-
-/**
- * ice_vsi_stop_rx_rings - stop VSI's rx rings
- * @vsi: the VSI
- *
- * Returns 0 on success and a negative value on error
- */
-static int ice_vsi_stop_rx_rings(struct ice_vsi *vsi)
-{
- return ice_vsi_ctrl_rx_rings(vsi, false);
-}
-
-/**
- * ice_vsi_stop_tx_rx_rings - stop VSI's tx and rx rings
- * @vsi: the VSI
- * Returns 0 on success and a negative value on error
- */
-static int ice_vsi_stop_tx_rx_rings(struct ice_vsi *vsi)
-{
- int err_tx, err_rx;
-
- err_tx = ice_vsi_stop_tx_rings(vsi);
- if (err_tx)
- dev_dbg(&vsi->back->pdev->dev, "Failed to disable Tx rings\n");
-
- err_rx = ice_vsi_stop_rx_rings(vsi);
- if (err_rx)
- dev_dbg(&vsi->back->pdev->dev, "Failed to disable Rx rings\n");
-
- if (err_tx || err_rx)
- return -EIO;
-
- return 0;
-}
-
/**
* ice_napi_enable_all - Enable NAPI for all q_vectors in the VSI
* @vsi: the VSI being configured
@@ -4822,7 +4308,7 @@ static void ice_napi_disable_all(struct ice_vsi *vsi)
*/
int ice_down(struct ice_vsi *vsi)
{
- int i, err;
+ int i, tx_err, rx_err;
/* Caller of this function is expected to set the
* vsi->state __ICE_DOWN bit
@@ -4833,7 +4319,18 @@ int ice_down(struct ice_vsi *vsi)
}
ice_vsi_dis_irq(vsi);
- err = ice_vsi_stop_tx_rx_rings(vsi);
+ tx_err = ice_vsi_stop_tx_rings(vsi);
+ if (tx_err)
+ netdev_err(vsi->netdev,
+ "Failed stop Tx rings, VSI %d error %d\n",
+ vsi->vsi_num, tx_err);
+
+ rx_err = ice_vsi_stop_rx_rings(vsi);
+ if (rx_err)
+ netdev_err(vsi->netdev,
+ "Failed stop Rx rings, VSI %d error %d\n",
+ vsi->vsi_num, rx_err);
+
ice_napi_disable_all(vsi);
ice_for_each_txq(vsi, i)
@@ -4842,10 +4339,14 @@ int ice_down(struct ice_vsi *vsi)
ice_for_each_rxq(vsi, i)
ice_clean_rx_ring(vsi->rx_rings[i]);
- if (err)
- netdev_err(vsi->netdev, "Failed to close VSI 0x%04X on switch 0x%04X\n",
+ if (tx_err || rx_err) {
+ netdev_err(vsi->netdev,
+ "Failed to close VSI 0x%04X on switch 0x%04X\n",
vsi->vsi_num, vsi->vsw->sw_id);
- return err;
+ return -EIO;
+ }
+
+ return 0;
}
/**
@@ -4865,6 +4366,7 @@ static int ice_vsi_setup_tx_rings(struct ice_vsi *vsi)
}
ice_for_each_txq(vsi, i) {
+ vsi->tx_rings[i]->netdev = vsi->netdev;
err = ice_setup_tx_ring(vsi->tx_rings[i]);
if (err)
break;
@@ -4890,6 +4392,7 @@ static int ice_vsi_setup_rx_rings(struct ice_vsi *vsi)
}
ice_for_each_rxq(vsi, i) {
+ vsi->rx_rings[i]->netdev = vsi->netdev;
err = ice_setup_rx_ring(vsi->rx_rings[i]);
if (err)
break;
--
2.17.1
^ permalink raw reply related
* [net-next 1/8] ice: Move common functions out of ice_main.c part 1/7
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem
Cc: Anirudh Venkataramanan, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <20181001211431.14325-1-jeffrey.t.kirsher@intel.com>
From: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
The functions that are used for PF VSI/netdev setup will also be used
for SR-IOV support. To allow reuse of these functions, move these
functions out of ice_main.c to ice_common.c/ice_lib.c
This move is done across multiple patches. Each patch moves a few
functions and may have minor adjustments. For example, a function that was
previously static in ice_main.c will be made non-static temporarily in
its new location to allow the driver to build cleanly. These adjustments
will be removed in subsequent patches where more code is moved out of
ice_main.c
In this particular patch, the following functions were moved out of
ice_main.c:
int ice_add_mac_to_list
ice_free_fltr_list
ice_stat_update40
ice_stat_update32
ice_update_eth_stats
ice_vsi_add_vlan
ice_vsi_kill_vlan
ice_vsi_manage_vlan_insertion
ice_vsi_manage_vlan_stripping
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/Makefile | 1 +
drivers/net/ethernet/intel/ice/ice_common.c | 61 ++++
drivers/net/ethernet/intel/ice/ice_common.h | 4 +
drivers/net/ethernet/intel/ice/ice_lib.c | 258 ++++++++++++++++
drivers/net/ethernet/intel/ice/ice_lib.h | 23 ++
drivers/net/ethernet/intel/ice/ice_main.c | 316 +-------------------
6 files changed, 348 insertions(+), 315 deletions(-)
create mode 100644 drivers/net/ethernet/intel/ice/ice_lib.c
create mode 100644 drivers/net/ethernet/intel/ice/ice_lib.h
diff --git a/drivers/net/ethernet/intel/ice/Makefile b/drivers/net/ethernet/intel/ice/Makefile
index 4058673fd853..45125bd074d9 100644
--- a/drivers/net/ethernet/intel/ice/Makefile
+++ b/drivers/net/ethernet/intel/ice/Makefile
@@ -13,5 +13,6 @@ ice-y := ice_main.o \
ice_nvm.o \
ice_switch.o \
ice_sched.o \
+ ice_lib.o \
ice_txrx.o \
ice_ethtool.o
diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index decfdb065a20..ef9229fa5510 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -2652,3 +2652,64 @@ ice_cfg_vsi_lan(struct ice_port_info *pi, u16 vsi_id, u8 tc_bitmap,
return ice_cfg_vsi_qs(pi, vsi_id, tc_bitmap, max_lanqs,
ICE_SCHED_NODE_OWNER_LAN);
}
+
+/**
+ * ice_stat_update40 - read 40 bit stat from the chip and update stat values
+ * @hw: ptr to the hardware info
+ * @hireg: high 32 bit HW register to read from
+ * @loreg: low 32 bit HW register to read from
+ * @prev_stat_loaded: bool to specify if previous stats are loaded
+ * @prev_stat: ptr to previous loaded stat value
+ * @cur_stat: ptr to current stat value
+ */
+void ice_stat_update40(struct ice_hw *hw, u32 hireg, u32 loreg,
+ bool prev_stat_loaded, u64 *prev_stat, u64 *cur_stat)
+{
+ u64 new_data;
+
+ new_data = rd32(hw, loreg);
+ new_data |= ((u64)(rd32(hw, hireg) & 0xFFFF)) << 32;
+
+ /* device stats are not reset at PFR, they likely will not be zeroed
+ * when the driver starts. So save the first values read and use them as
+ * offsets to be subtracted from the raw values in order to report stats
+ * that count from zero.
+ */
+ if (!prev_stat_loaded)
+ *prev_stat = new_data;
+ if (new_data >= *prev_stat)
+ *cur_stat = new_data - *prev_stat;
+ else
+ /* to manage the potential roll-over */
+ *cur_stat = (new_data + BIT_ULL(40)) - *prev_stat;
+ *cur_stat &= 0xFFFFFFFFFFULL;
+}
+
+/**
+ * ice_stat_update32 - read 32 bit stat from the chip and update stat values
+ * @hw: ptr to the hardware info
+ * @reg: HW register to read from
+ * @prev_stat_loaded: bool to specify if previous stats are loaded
+ * @prev_stat: ptr to previous loaded stat value
+ * @cur_stat: ptr to current stat value
+ */
+void ice_stat_update32(struct ice_hw *hw, u32 reg, bool prev_stat_loaded,
+ u64 *prev_stat, u64 *cur_stat)
+{
+ u32 new_data;
+
+ new_data = rd32(hw, reg);
+
+ /* device stats are not reset at PFR, they likely will not be zeroed
+ * when the driver starts. So save the first values read and use them as
+ * offsets to be subtracted from the raw values in order to report stats
+ * that count from zero.
+ */
+ if (!prev_stat_loaded)
+ *prev_stat = new_data;
+ if (new_data >= *prev_stat)
+ *cur_stat = new_data - *prev_stat;
+ else
+ /* to manage the potential roll-over */
+ *cur_stat = (new_data + BIT_ULL(32)) - *prev_stat;
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_common.h b/drivers/net/ethernet/intel/ice/ice_common.h
index aac2d6cadaaf..80d288a07731 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.h
+++ b/drivers/net/ethernet/intel/ice/ice_common.h
@@ -96,4 +96,8 @@ ice_ena_vsi_txq(struct ice_port_info *pi, u16 vsi_id, u8 tc, u8 num_qgrps,
struct ice_aqc_add_tx_qgrp *buf, u16 buf_size,
struct ice_sq_cd *cd);
void ice_output_fw_log(struct ice_hw *hw, struct ice_aq_desc *desc, void *buf);
+void ice_stat_update40(struct ice_hw *hw, u32 hireg, u32 loreg,
+ bool prev_stat_loaded, u64 *prev_stat, u64 *cur_stat);
+void ice_stat_update32(struct ice_hw *hw, u32 reg, bool prev_stat_loaded,
+ u64 *prev_stat, u64 *cur_stat);
#endif /* _ICE_COMMON_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
new file mode 100644
index 000000000000..1cf4dca12495
--- /dev/null
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -0,0 +1,258 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2018, Intel Corporation. */
+
+#include "ice.h"
+#include "ice_lib.h"
+
+/**
+ * ice_add_mac_to_list - Add a mac address filter entry to the list
+ * @vsi: the VSI to be forwarded to
+ * @add_list: pointer to the list which contains MAC filter entries
+ * @macaddr: the MAC address to be added.
+ *
+ * Adds mac address filter entry to the temp list
+ *
+ * Returns 0 on success or ENOMEM on failure.
+ */
+int ice_add_mac_to_list(struct ice_vsi *vsi, struct list_head *add_list,
+ const u8 *macaddr)
+{
+ struct ice_fltr_list_entry *tmp;
+ struct ice_pf *pf = vsi->back;
+
+ tmp = devm_kzalloc(&pf->pdev->dev, sizeof(*tmp), GFP_ATOMIC);
+ if (!tmp)
+ return -ENOMEM;
+
+ tmp->fltr_info.flag = ICE_FLTR_TX;
+ tmp->fltr_info.src = vsi->vsi_num;
+ tmp->fltr_info.lkup_type = ICE_SW_LKUP_MAC;
+ tmp->fltr_info.fltr_act = ICE_FWD_TO_VSI;
+ tmp->fltr_info.fwd_id.vsi_id = vsi->vsi_num;
+ ether_addr_copy(tmp->fltr_info.l_data.mac.mac_addr, macaddr);
+
+ INIT_LIST_HEAD(&tmp->list_entry);
+ list_add(&tmp->list_entry, add_list);
+
+ return 0;
+}
+
+/**
+ * ice_update_eth_stats - Update VSI-specific ethernet statistics counters
+ * @vsi: the VSI to be updated
+ */
+void ice_update_eth_stats(struct ice_vsi *vsi)
+{
+ struct ice_eth_stats *prev_es, *cur_es;
+ struct ice_hw *hw = &vsi->back->hw;
+ u16 vsi_num = vsi->vsi_num; /* HW absolute index of a VSI */
+
+ prev_es = &vsi->eth_stats_prev;
+ cur_es = &vsi->eth_stats;
+
+ ice_stat_update40(hw, GLV_GORCH(vsi_num), GLV_GORCL(vsi_num),
+ vsi->stat_offsets_loaded, &prev_es->rx_bytes,
+ &cur_es->rx_bytes);
+
+ ice_stat_update40(hw, GLV_UPRCH(vsi_num), GLV_UPRCL(vsi_num),
+ vsi->stat_offsets_loaded, &prev_es->rx_unicast,
+ &cur_es->rx_unicast);
+
+ ice_stat_update40(hw, GLV_MPRCH(vsi_num), GLV_MPRCL(vsi_num),
+ vsi->stat_offsets_loaded, &prev_es->rx_multicast,
+ &cur_es->rx_multicast);
+
+ ice_stat_update40(hw, GLV_BPRCH(vsi_num), GLV_BPRCL(vsi_num),
+ vsi->stat_offsets_loaded, &prev_es->rx_broadcast,
+ &cur_es->rx_broadcast);
+
+ ice_stat_update32(hw, GLV_RDPC(vsi_num), vsi->stat_offsets_loaded,
+ &prev_es->rx_discards, &cur_es->rx_discards);
+
+ ice_stat_update40(hw, GLV_GOTCH(vsi_num), GLV_GOTCL(vsi_num),
+ vsi->stat_offsets_loaded, &prev_es->tx_bytes,
+ &cur_es->tx_bytes);
+
+ ice_stat_update40(hw, GLV_UPTCH(vsi_num), GLV_UPTCL(vsi_num),
+ vsi->stat_offsets_loaded, &prev_es->tx_unicast,
+ &cur_es->tx_unicast);
+
+ ice_stat_update40(hw, GLV_MPTCH(vsi_num), GLV_MPTCL(vsi_num),
+ vsi->stat_offsets_loaded, &prev_es->tx_multicast,
+ &cur_es->tx_multicast);
+
+ ice_stat_update40(hw, GLV_BPTCH(vsi_num), GLV_BPTCL(vsi_num),
+ vsi->stat_offsets_loaded, &prev_es->tx_broadcast,
+ &cur_es->tx_broadcast);
+
+ ice_stat_update32(hw, GLV_TEPC(vsi_num), vsi->stat_offsets_loaded,
+ &prev_es->tx_errors, &cur_es->tx_errors);
+
+ vsi->stat_offsets_loaded = true;
+}
+
+/**
+ * ice_free_fltr_list - free filter lists helper
+ * @dev: pointer to the device struct
+ * @h: pointer to the list head to be freed
+ *
+ * Helper function to free filter lists previously created using
+ * ice_add_mac_to_list
+ */
+void ice_free_fltr_list(struct device *dev, struct list_head *h)
+{
+ struct ice_fltr_list_entry *e, *tmp;
+
+ list_for_each_entry_safe(e, tmp, h, list_entry) {
+ list_del(&e->list_entry);
+ devm_kfree(dev, e);
+ }
+}
+
+/**
+ * ice_vsi_add_vlan - Add VSI membership for given VLAN
+ * @vsi: the VSI being configured
+ * @vid: VLAN id to be added
+ */
+int ice_vsi_add_vlan(struct ice_vsi *vsi, u16 vid)
+{
+ struct ice_fltr_list_entry *tmp;
+ struct ice_pf *pf = vsi->back;
+ LIST_HEAD(tmp_add_list);
+ enum ice_status status;
+ int err = 0;
+
+ tmp = devm_kzalloc(&pf->pdev->dev, sizeof(*tmp), GFP_KERNEL);
+ if (!tmp)
+ return -ENOMEM;
+
+ tmp->fltr_info.lkup_type = ICE_SW_LKUP_VLAN;
+ tmp->fltr_info.fltr_act = ICE_FWD_TO_VSI;
+ tmp->fltr_info.flag = ICE_FLTR_TX;
+ tmp->fltr_info.src = vsi->vsi_num;
+ tmp->fltr_info.fwd_id.vsi_id = vsi->vsi_num;
+ tmp->fltr_info.l_data.vlan.vlan_id = vid;
+
+ INIT_LIST_HEAD(&tmp->list_entry);
+ list_add(&tmp->list_entry, &tmp_add_list);
+
+ status = ice_add_vlan(&pf->hw, &tmp_add_list);
+ if (status) {
+ err = -ENODEV;
+ dev_err(&pf->pdev->dev, "Failure Adding VLAN %d on VSI %i\n",
+ vid, vsi->vsi_num);
+ }
+
+ ice_free_fltr_list(&pf->pdev->dev, &tmp_add_list);
+ return err;
+}
+
+/**
+ * ice_vsi_kill_vlan - Remove VSI membership for a given VLAN
+ * @vsi: the VSI being configured
+ * @vid: VLAN id to be removed
+ *
+ * Returns 0 on success and negative on failure
+ */
+int ice_vsi_kill_vlan(struct ice_vsi *vsi, u16 vid)
+{
+ struct ice_fltr_list_entry *list;
+ struct ice_pf *pf = vsi->back;
+ LIST_HEAD(tmp_add_list);
+ int status = 0;
+
+ list = devm_kzalloc(&pf->pdev->dev, sizeof(*list), GFP_KERNEL);
+ if (!list)
+ return -ENOMEM;
+
+ list->fltr_info.lkup_type = ICE_SW_LKUP_VLAN;
+ list->fltr_info.fwd_id.vsi_id = vsi->vsi_num;
+ list->fltr_info.fltr_act = ICE_FWD_TO_VSI;
+ list->fltr_info.l_data.vlan.vlan_id = vid;
+ list->fltr_info.flag = ICE_FLTR_TX;
+ list->fltr_info.src = vsi->vsi_num;
+
+ INIT_LIST_HEAD(&list->list_entry);
+ list_add(&list->list_entry, &tmp_add_list);
+
+ if (ice_remove_vlan(&pf->hw, &tmp_add_list)) {
+ dev_err(&pf->pdev->dev, "Error removing VLAN %d on vsi %i\n",
+ vid, vsi->vsi_num);
+ status = -EIO;
+ }
+
+ ice_free_fltr_list(&pf->pdev->dev, &tmp_add_list);
+ return status;
+}
+
+/**
+ * ice_vsi_manage_vlan_insertion - Manage VLAN insertion for the VSI for Tx
+ * @vsi: the VSI being changed
+ */
+int ice_vsi_manage_vlan_insertion(struct ice_vsi *vsi)
+{
+ struct device *dev = &vsi->back->pdev->dev;
+ struct ice_hw *hw = &vsi->back->hw;
+ struct ice_vsi_ctx ctxt = { 0 };
+ enum ice_status status;
+
+ /* Here we are configuring the VSI to let the driver add VLAN tags by
+ * setting vlan_flags to ICE_AQ_VSI_VLAN_MODE_ALL. The actual VLAN tag
+ * insertion happens in the Tx hot path, in ice_tx_map.
+ */
+ ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_MODE_ALL;
+
+ ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
+ ctxt.vsi_num = vsi->vsi_num;
+
+ status = ice_aq_update_vsi(hw, &ctxt, NULL);
+ if (status) {
+ dev_err(dev, "update VSI for VLAN insert failed, err %d aq_err %d\n",
+ status, hw->adminq.sq_last_status);
+ return -EIO;
+ }
+
+ vsi->info.vlan_flags = ctxt.info.vlan_flags;
+ return 0;
+}
+
+/**
+ * ice_vsi_manage_vlan_stripping - Manage VLAN stripping for the VSI for Rx
+ * @vsi: the VSI being changed
+ * @ena: boolean value indicating if this is a enable or disable request
+ */
+int ice_vsi_manage_vlan_stripping(struct ice_vsi *vsi, bool ena)
+{
+ struct device *dev = &vsi->back->pdev->dev;
+ struct ice_hw *hw = &vsi->back->hw;
+ struct ice_vsi_ctx ctxt = { 0 };
+ enum ice_status status;
+
+ /* Here we are configuring what the VSI should do with the VLAN tag in
+ * the Rx packet. We can either leave the tag in the packet or put it in
+ * the Rx descriptor.
+ */
+ if (ena) {
+ /* Strip VLAN tag from Rx packet and put it in the desc */
+ ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_EMOD_STR_BOTH;
+ } else {
+ /* Disable stripping. Leave tag in packet */
+ ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_EMOD_NOTHING;
+ }
+
+ /* Allow all packets untagged/tagged */
+ ctxt.info.vlan_flags |= ICE_AQ_VSI_VLAN_MODE_ALL;
+
+ ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
+ ctxt.vsi_num = vsi->vsi_num;
+
+ status = ice_aq_update_vsi(hw, &ctxt, NULL);
+ if (status) {
+ dev_err(dev, "update VSI for VLAN strip failed, ena = %d err %d aq_err %d\n",
+ ena, status, hw->adminq.sq_last_status);
+ return -EIO;
+ }
+
+ vsi->info.vlan_flags = ctxt.info.vlan_flags;
+ return 0;
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
new file mode 100644
index 000000000000..c10874d26eee
--- /dev/null
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2018, Intel Corporation. */
+
+#ifndef _ICE_LIB_H_
+#define _ICE_LIB_H_
+
+#include "ice.h"
+
+int ice_add_mac_to_list(struct ice_vsi *vsi, struct list_head *add_list,
+ const u8 *macaddr);
+
+void ice_free_fltr_list(struct device *dev, struct list_head *h);
+
+void ice_update_eth_stats(struct ice_vsi *vsi);
+
+int ice_vsi_add_vlan(struct ice_vsi *vsi, u16 vid);
+
+int ice_vsi_kill_vlan(struct ice_vsi *vsi, u16 vid);
+
+int ice_vsi_manage_vlan_insertion(struct ice_vsi *vsi);
+
+int ice_vsi_manage_vlan_stripping(struct ice_vsi *vsi, bool ena);
+#endif /* !_ICE_LIB_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index b02942efcaea..2673430ec2e9 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -6,6 +6,7 @@
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include "ice.h"
+#include "ice_lib.h"
#define DRV_VERSION "0.7.1-k"
#define DRV_SUMMARY "Intel(R) Ethernet Connection E800 Series Linux Driver"
@@ -243,39 +244,6 @@ static int ice_free_res(struct ice_res_tracker *res, u16 index, u16 id)
return count;
}
-/**
- * ice_add_mac_to_list - Add a mac address filter entry to the list
- * @vsi: the VSI to be forwarded to
- * @add_list: pointer to the list which contains MAC filter entries
- * @macaddr: the MAC address to be added.
- *
- * Adds mac address filter entry to the temp list
- *
- * Returns 0 on success or ENOMEM on failure.
- */
-static int ice_add_mac_to_list(struct ice_vsi *vsi, struct list_head *add_list,
- const u8 *macaddr)
-{
- struct ice_fltr_list_entry *tmp;
- struct ice_pf *pf = vsi->back;
-
- tmp = devm_kzalloc(&pf->pdev->dev, sizeof(*tmp), GFP_ATOMIC);
- if (!tmp)
- return -ENOMEM;
-
- tmp->fltr_info.flag = ICE_FLTR_TX;
- tmp->fltr_info.src = vsi->vsi_num;
- tmp->fltr_info.lkup_type = ICE_SW_LKUP_MAC;
- tmp->fltr_info.fltr_act = ICE_FWD_TO_VSI;
- tmp->fltr_info.fwd_id.vsi_id = vsi->vsi_num;
- ether_addr_copy(tmp->fltr_info.l_data.mac.mac_addr, macaddr);
-
- INIT_LIST_HEAD(&tmp->list_entry);
- list_add(&tmp->list_entry, add_list);
-
- return 0;
-}
-
/**
* ice_add_mac_to_sync_list - creates list of mac addresses to be synced
* @netdev: the net device on which the sync is happening
@@ -318,24 +286,6 @@ static int ice_add_mac_to_unsync_list(struct net_device *netdev, const u8 *addr)
return 0;
}
-/**
- * ice_free_fltr_list - free filter lists helper
- * @dev: pointer to the device struct
- * @h: pointer to the list head to be freed
- *
- * Helper function to free filter lists previously created using
- * ice_add_mac_to_list
- */
-static void ice_free_fltr_list(struct device *dev, struct list_head *h)
-{
- struct ice_fltr_list_entry *e, *tmp;
-
- list_for_each_entry_safe(e, tmp, h, list_entry) {
- list_del(&e->list_entry);
- devm_kfree(dev, e);
- }
-}
-
/**
* ice_vsi_fltr_changed - check if filter state changed
* @vsi: VSI to be checked
@@ -3148,44 +3098,6 @@ ice_pf_vsi_setup(struct ice_pf *pf, struct ice_port_info *pi)
return ice_vsi_setup(pf, pi, ICE_VSI_PF, ICE_INVAL_VFID);
}
-/**
- * ice_vsi_add_vlan - Add vsi membership for given vlan
- * @vsi: the vsi being configured
- * @vid: vlan id to be added
- */
-static int ice_vsi_add_vlan(struct ice_vsi *vsi, u16 vid)
-{
- struct ice_fltr_list_entry *tmp;
- struct ice_pf *pf = vsi->back;
- LIST_HEAD(tmp_add_list);
- enum ice_status status;
- int err = 0;
-
- tmp = devm_kzalloc(&pf->pdev->dev, sizeof(*tmp), GFP_KERNEL);
- if (!tmp)
- return -ENOMEM;
-
- tmp->fltr_info.lkup_type = ICE_SW_LKUP_VLAN;
- tmp->fltr_info.fltr_act = ICE_FWD_TO_VSI;
- tmp->fltr_info.flag = ICE_FLTR_TX;
- tmp->fltr_info.src = vsi->vsi_num;
- tmp->fltr_info.fwd_id.vsi_id = vsi->vsi_num;
- tmp->fltr_info.l_data.vlan.vlan_id = vid;
-
- INIT_LIST_HEAD(&tmp->list_entry);
- list_add(&tmp->list_entry, &tmp_add_list);
-
- status = ice_add_vlan(&pf->hw, &tmp_add_list);
- if (status) {
- err = -ENODEV;
- dev_err(&pf->pdev->dev, "Failure Adding VLAN %d on VSI %i\n",
- vid, vsi->vsi_num);
- }
-
- ice_free_fltr_list(&pf->pdev->dev, &tmp_add_list);
- return err;
-}
-
/**
* ice_vlan_rx_add_vid - Add a vlan id filter to HW offload
* @netdev: network interface to be adjusted
@@ -3229,44 +3141,6 @@ static int ice_vlan_rx_add_vid(struct net_device *netdev,
return ret;
}
-/**
- * ice_vsi_kill_vlan - Remove VSI membership for a given VLAN
- * @vsi: the VSI being configured
- * @vid: VLAN id to be removed
- *
- * Returns 0 on success and negative on failure
- */
-static int ice_vsi_kill_vlan(struct ice_vsi *vsi, u16 vid)
-{
- struct ice_fltr_list_entry *list;
- struct ice_pf *pf = vsi->back;
- LIST_HEAD(tmp_add_list);
- int status = 0;
-
- list = devm_kzalloc(&pf->pdev->dev, sizeof(*list), GFP_KERNEL);
- if (!list)
- return -ENOMEM;
-
- list->fltr_info.lkup_type = ICE_SW_LKUP_VLAN;
- list->fltr_info.fwd_id.vsi_id = vsi->vsi_num;
- list->fltr_info.fltr_act = ICE_FWD_TO_VSI;
- list->fltr_info.l_data.vlan.vlan_id = vid;
- list->fltr_info.flag = ICE_FLTR_TX;
- list->fltr_info.src = vsi->vsi_num;
-
- INIT_LIST_HEAD(&list->list_entry);
- list_add(&list->list_entry, &tmp_add_list);
-
- if (ice_remove_vlan(&pf->hw, &tmp_add_list)) {
- dev_err(&pf->pdev->dev, "Error removing VLAN %d on vsi %i\n",
- vid, vsi->vsi_num);
- status = -EIO;
- }
-
- ice_free_fltr_list(&pf->pdev->dev, &tmp_add_list);
- return status;
-}
-
/**
* ice_vlan_rx_kill_vid - Remove a vlan id filter from HW offload
* @netdev: network interface to be adjusted
@@ -4023,78 +3897,6 @@ static int ice_fdb_del(struct ndmsg *ndm, __always_unused struct nlattr *tb[],
return err;
}
-/**
- * ice_vsi_manage_vlan_insertion - Manage VLAN insertion for the VSI for Tx
- * @vsi: the vsi being changed
- */
-static int ice_vsi_manage_vlan_insertion(struct ice_vsi *vsi)
-{
- struct device *dev = &vsi->back->pdev->dev;
- struct ice_hw *hw = &vsi->back->hw;
- struct ice_vsi_ctx ctxt = { 0 };
- enum ice_status status;
-
- /* Here we are configuring the VSI to let the driver add VLAN tags by
- * setting vlan_flags to ICE_AQ_VSI_VLAN_MODE_ALL. The actual VLAN tag
- * insertion happens in the Tx hot path, in ice_tx_map.
- */
- ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_MODE_ALL;
-
- ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
- ctxt.vsi_num = vsi->vsi_num;
-
- status = ice_aq_update_vsi(hw, &ctxt, NULL);
- if (status) {
- dev_err(dev, "update VSI for VLAN insert failed, err %d aq_err %d\n",
- status, hw->adminq.sq_last_status);
- return -EIO;
- }
-
- vsi->info.vlan_flags = ctxt.info.vlan_flags;
- return 0;
-}
-
-/**
- * ice_vsi_manage_vlan_stripping - Manage VLAN stripping for the VSI for Rx
- * @vsi: the vsi being changed
- * @ena: boolean value indicating if this is a enable or disable request
- */
-static int ice_vsi_manage_vlan_stripping(struct ice_vsi *vsi, bool ena)
-{
- struct device *dev = &vsi->back->pdev->dev;
- struct ice_hw *hw = &vsi->back->hw;
- struct ice_vsi_ctx ctxt = { 0 };
- enum ice_status status;
-
- /* Here we are configuring what the VSI should do with the VLAN tag in
- * the Rx packet. We can either leave the tag in the packet or put it in
- * the Rx descriptor.
- */
- if (ena) {
- /* Strip VLAN tag from Rx packet and put it in the desc */
- ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_EMOD_STR_BOTH;
- } else {
- /* Disable stripping. Leave tag in packet */
- ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_EMOD_NOTHING;
- }
-
- /* Allow all packets untagged/tagged */
- ctxt.info.vlan_flags |= ICE_AQ_VSI_VLAN_MODE_ALL;
-
- ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
- ctxt.vsi_num = vsi->vsi_num;
-
- status = ice_aq_update_vsi(hw, &ctxt, NULL);
- if (status) {
- dev_err(dev, "update VSI for VALN strip failed, ena = %d err %d aq_err %d\n",
- ena, status, hw->adminq.sq_last_status);
- return -EIO;
- }
-
- vsi->info.vlan_flags = ctxt.info.vlan_flags;
- return 0;
-}
-
/**
* ice_set_features - set the netdev feature flags
* @netdev: ptr to the netdev being adjusted
@@ -4727,122 +4529,6 @@ static void ice_fetch_u64_stats_per_ring(struct ice_ring *ring, u64 *pkts,
} while (u64_stats_fetch_retry_irq(&ring->syncp, start));
}
-/**
- * ice_stat_update40 - read 40 bit stat from the chip and update stat values
- * @hw: ptr to the hardware info
- * @hireg: high 32 bit HW register to read from
- * @loreg: low 32 bit HW register to read from
- * @prev_stat_loaded: bool to specify if previous stats are loaded
- * @prev_stat: ptr to previous loaded stat value
- * @cur_stat: ptr to current stat value
- */
-static void ice_stat_update40(struct ice_hw *hw, u32 hireg, u32 loreg,
- bool prev_stat_loaded, u64 *prev_stat,
- u64 *cur_stat)
-{
- u64 new_data;
-
- new_data = rd32(hw, loreg);
- new_data |= ((u64)(rd32(hw, hireg) & 0xFFFF)) << 32;
-
- /* device stats are not reset at PFR, they likely will not be zeroed
- * when the driver starts. So save the first values read and use them as
- * offsets to be subtracted from the raw values in order to report stats
- * that count from zero.
- */
- if (!prev_stat_loaded)
- *prev_stat = new_data;
- if (likely(new_data >= *prev_stat))
- *cur_stat = new_data - *prev_stat;
- else
- /* to manage the potential roll-over */
- *cur_stat = (new_data + BIT_ULL(40)) - *prev_stat;
- *cur_stat &= 0xFFFFFFFFFFULL;
-}
-
-/**
- * ice_stat_update32 - read 32 bit stat from the chip and update stat values
- * @hw: ptr to the hardware info
- * @reg: HW register to read from
- * @prev_stat_loaded: bool to specify if previous stats are loaded
- * @prev_stat: ptr to previous loaded stat value
- * @cur_stat: ptr to current stat value
- */
-static void ice_stat_update32(struct ice_hw *hw, u32 reg, bool prev_stat_loaded,
- u64 *prev_stat, u64 *cur_stat)
-{
- u32 new_data;
-
- new_data = rd32(hw, reg);
-
- /* device stats are not reset at PFR, they likely will not be zeroed
- * when the driver starts. So save the first values read and use them as
- * offsets to be subtracted from the raw values in order to report stats
- * that count from zero.
- */
- if (!prev_stat_loaded)
- *prev_stat = new_data;
- if (likely(new_data >= *prev_stat))
- *cur_stat = new_data - *prev_stat;
- else
- /* to manage the potential roll-over */
- *cur_stat = (new_data + BIT_ULL(32)) - *prev_stat;
-}
-
-/**
- * ice_update_eth_stats - Update VSI-specific ethernet statistics counters
- * @vsi: the VSI to be updated
- */
-static void ice_update_eth_stats(struct ice_vsi *vsi)
-{
- struct ice_eth_stats *prev_es, *cur_es;
- struct ice_hw *hw = &vsi->back->hw;
- u16 vsi_num = vsi->vsi_num; /* HW absolute index of a VSI */
-
- prev_es = &vsi->eth_stats_prev;
- cur_es = &vsi->eth_stats;
-
- ice_stat_update40(hw, GLV_GORCH(vsi_num), GLV_GORCL(vsi_num),
- vsi->stat_offsets_loaded, &prev_es->rx_bytes,
- &cur_es->rx_bytes);
-
- ice_stat_update40(hw, GLV_UPRCH(vsi_num), GLV_UPRCL(vsi_num),
- vsi->stat_offsets_loaded, &prev_es->rx_unicast,
- &cur_es->rx_unicast);
-
- ice_stat_update40(hw, GLV_MPRCH(vsi_num), GLV_MPRCL(vsi_num),
- vsi->stat_offsets_loaded, &prev_es->rx_multicast,
- &cur_es->rx_multicast);
-
- ice_stat_update40(hw, GLV_BPRCH(vsi_num), GLV_BPRCL(vsi_num),
- vsi->stat_offsets_loaded, &prev_es->rx_broadcast,
- &cur_es->rx_broadcast);
-
- ice_stat_update32(hw, GLV_RDPC(vsi_num), vsi->stat_offsets_loaded,
- &prev_es->rx_discards, &cur_es->rx_discards);
-
- ice_stat_update40(hw, GLV_GOTCH(vsi_num), GLV_GOTCL(vsi_num),
- vsi->stat_offsets_loaded, &prev_es->tx_bytes,
- &cur_es->tx_bytes);
-
- ice_stat_update40(hw, GLV_UPTCH(vsi_num), GLV_UPTCL(vsi_num),
- vsi->stat_offsets_loaded, &prev_es->tx_unicast,
- &cur_es->tx_unicast);
-
- ice_stat_update40(hw, GLV_MPTCH(vsi_num), GLV_MPTCL(vsi_num),
- vsi->stat_offsets_loaded, &prev_es->tx_multicast,
- &cur_es->tx_multicast);
-
- ice_stat_update40(hw, GLV_BPTCH(vsi_num), GLV_BPTCL(vsi_num),
- vsi->stat_offsets_loaded, &prev_es->tx_broadcast,
- &cur_es->tx_broadcast);
-
- ice_stat_update32(hw, GLV_TEPC(vsi_num), vsi->stat_offsets_loaded,
- &prev_es->tx_errors, &cur_es->tx_errors);
-
- vsi->stat_offsets_loaded = true;
-}
-
/**
* ice_update_vsi_ring_stats - Update VSI stats counters
* @vsi: the VSI to be updated
--
2.17.1
^ permalink raw reply related
* [net-next 6/8] ice: Move common functions out of ice_main.c part 6/7
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem
Cc: Anirudh Venkataramanan, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <20181001211431.14325-1-jeffrey.t.kirsher@intel.com>
From: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
This patch continues the code move out of ice_main.c
The following top level functions (and related dependency functions) were
moved to ice_lib.c:
ice_vsi_setup_vector_base
ice_vsi_alloc_q_vectors
ice_vsi_get_qs
The following functions were made static again:
ice_vsi_free_arrays
ice_vsi_clear_rings
Also, in this patch, the netdev and NAPI registration logic was de-coupled
from the VSI creation logic (ice_vsi_setup) as for SR-IOV, while we want to
create VF VSIs using ice_vsi_setup, we don't want to create netdevs.
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_lib.c | 463 +++++++++++++++++++-
drivers/net/ethernet/intel/ice/ice_lib.h | 16 +-
drivers/net/ethernet/intel/ice/ice_main.c | 492 +++-------------------
3 files changed, 521 insertions(+), 450 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 6ba82337d017..232ca06974ea 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -346,7 +346,7 @@ void ice_vsi_delete(struct ice_vsi *vsi)
* @vsi: pointer to VSI being cleared
* @free_qvectors: bool to specify if q_vectors should be deallocated
*/
-void ice_vsi_free_arrays(struct ice_vsi *vsi, bool free_qvectors)
+static void ice_vsi_free_arrays(struct ice_vsi *vsi, bool free_qvectors)
{
struct ice_pf *pf = vsi->back;
@@ -423,6 +423,141 @@ irqreturn_t ice_msix_clean_rings(int __always_unused irq, void *data)
return IRQ_HANDLED;
}
+/**
+ * ice_vsi_get_qs_contig - Assign a contiguous chunk of queues to VSI
+ * @vsi: the VSI getting queues
+ *
+ * Return 0 on success and a negative value on error
+ */
+static int ice_vsi_get_qs_contig(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ int offset, ret = 0;
+
+ mutex_lock(&pf->avail_q_mutex);
+ /* look for contiguous block of queues for Tx */
+ offset = bitmap_find_next_zero_area(pf->avail_txqs, ICE_MAX_TXQS,
+ 0, vsi->alloc_txq, 0);
+ if (offset < ICE_MAX_TXQS) {
+ int i;
+
+ bitmap_set(pf->avail_txqs, offset, vsi->alloc_txq);
+ for (i = 0; i < vsi->alloc_txq; i++)
+ vsi->txq_map[i] = i + offset;
+ } else {
+ ret = -ENOMEM;
+ vsi->tx_mapping_mode = ICE_VSI_MAP_SCATTER;
+ }
+
+ /* look for contiguous block of queues for Rx */
+ offset = bitmap_find_next_zero_area(pf->avail_rxqs, ICE_MAX_RXQS,
+ 0, vsi->alloc_rxq, 0);
+ if (offset < ICE_MAX_RXQS) {
+ int i;
+
+ bitmap_set(pf->avail_rxqs, offset, vsi->alloc_rxq);
+ for (i = 0; i < vsi->alloc_rxq; i++)
+ vsi->rxq_map[i] = i + offset;
+ } else {
+ ret = -ENOMEM;
+ vsi->rx_mapping_mode = ICE_VSI_MAP_SCATTER;
+ }
+ mutex_unlock(&pf->avail_q_mutex);
+
+ return ret;
+}
+
+/**
+ * ice_vsi_get_qs_scatter - Assign a scattered queues to VSI
+ * @vsi: the VSI getting queues
+ *
+ * Return 0 on success and a negative value on error
+ */
+static int ice_vsi_get_qs_scatter(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ int i, index = 0;
+
+ mutex_lock(&pf->avail_q_mutex);
+
+ if (vsi->tx_mapping_mode == ICE_VSI_MAP_SCATTER) {
+ for (i = 0; i < vsi->alloc_txq; i++) {
+ index = find_next_zero_bit(pf->avail_txqs,
+ ICE_MAX_TXQS, index);
+ if (index < ICE_MAX_TXQS) {
+ set_bit(index, pf->avail_txqs);
+ vsi->txq_map[i] = index;
+ } else {
+ goto err_scatter_tx;
+ }
+ }
+ }
+
+ if (vsi->rx_mapping_mode == ICE_VSI_MAP_SCATTER) {
+ for (i = 0; i < vsi->alloc_rxq; i++) {
+ index = find_next_zero_bit(pf->avail_rxqs,
+ ICE_MAX_RXQS, index);
+ if (index < ICE_MAX_RXQS) {
+ set_bit(index, pf->avail_rxqs);
+ vsi->rxq_map[i] = index;
+ } else {
+ goto err_scatter_rx;
+ }
+ }
+ }
+
+ mutex_unlock(&pf->avail_q_mutex);
+ return 0;
+
+err_scatter_rx:
+ /* unflag any queues we have grabbed (i is failed position) */
+ for (index = 0; index < i; index++) {
+ clear_bit(vsi->rxq_map[index], pf->avail_rxqs);
+ vsi->rxq_map[index] = 0;
+ }
+ i = vsi->alloc_txq;
+err_scatter_tx:
+ /* i is either position of failed attempt or vsi->alloc_txq */
+ for (index = 0; index < i; index++) {
+ clear_bit(vsi->txq_map[index], pf->avail_txqs);
+ vsi->txq_map[index] = 0;
+ }
+
+ mutex_unlock(&pf->avail_q_mutex);
+ return -ENOMEM;
+}
+
+/**
+ * ice_vsi_get_qs - Assign queues from PF to VSI
+ * @vsi: the VSI to assign queues to
+ *
+ * Returns 0 on success and a negative value on error
+ */
+int ice_vsi_get_qs(struct ice_vsi *vsi)
+{
+ int ret = 0;
+
+ vsi->tx_mapping_mode = ICE_VSI_MAP_CONTIG;
+ vsi->rx_mapping_mode = ICE_VSI_MAP_CONTIG;
+
+ /* NOTE: ice_vsi_get_qs_contig() will set the Rx/Tx mapping
+ * modes individually to scatter if assigning contiguous queues
+ * to Rx or Tx fails
+ */
+ ret = ice_vsi_get_qs_contig(vsi);
+ if (ret < 0) {
+ if (vsi->tx_mapping_mode == ICE_VSI_MAP_SCATTER)
+ vsi->alloc_txq = max_t(u16, vsi->alloc_txq,
+ ICE_MAX_SCATTER_TXQS);
+ if (vsi->rx_mapping_mode == ICE_VSI_MAP_SCATTER)
+ vsi->alloc_rxq = max_t(u16, vsi->alloc_rxq,
+ ICE_MAX_SCATTER_RXQS);
+ ret = ice_vsi_get_qs_scatter(vsi);
+ }
+
+ return ret;
+}
+
/**
* ice_vsi_put_qs - Release queues from VSI to PF
* @vsi: the VSI that is going to release queues
@@ -447,6 +582,22 @@ void ice_vsi_put_qs(struct ice_vsi *vsi)
mutex_unlock(&pf->avail_q_mutex);
}
+/**
+ * ice_rss_clean - Delete RSS related VSI structures that hold user inputs
+ * @vsi: the VSI being removed
+ */
+static void ice_rss_clean(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf;
+
+ pf = vsi->back;
+
+ if (vsi->rss_hkey_user)
+ devm_kfree(&pf->pdev->dev, vsi->rss_hkey_user);
+ if (vsi->rss_lut_user)
+ devm_kfree(&pf->pdev->dev, vsi->rss_lut_user);
+}
+
/**
* ice_vsi_set_rss_params - Setup RSS capabilities per VSI type
* @vsi: the VSI being configured
@@ -685,11 +836,183 @@ int ice_vsi_init(struct ice_vsi *vsi)
return ret;
}
+/**
+ * ice_free_q_vector - Free memory allocated for a specific interrupt vector
+ * @vsi: VSI having the memory freed
+ * @v_idx: index of the vector to be freed
+ */
+static void ice_free_q_vector(struct ice_vsi *vsi, int v_idx)
+{
+ struct ice_q_vector *q_vector;
+ struct ice_ring *ring;
+
+ if (!vsi->q_vectors[v_idx]) {
+ dev_dbg(&vsi->back->pdev->dev, "Queue vector at index %d not found\n",
+ v_idx);
+ return;
+ }
+ q_vector = vsi->q_vectors[v_idx];
+
+ ice_for_each_ring(ring, q_vector->tx)
+ ring->q_vector = NULL;
+ ice_for_each_ring(ring, q_vector->rx)
+ ring->q_vector = NULL;
+
+ /* only VSI with an associated netdev is set up with NAPI */
+ if (vsi->netdev)
+ netif_napi_del(&q_vector->napi);
+
+ devm_kfree(&vsi->back->pdev->dev, q_vector);
+ vsi->q_vectors[v_idx] = NULL;
+}
+
+/**
+ * ice_vsi_free_q_vectors - Free memory allocated for interrupt vectors
+ * @vsi: the VSI having memory freed
+ */
+void ice_vsi_free_q_vectors(struct ice_vsi *vsi)
+{
+ int v_idx;
+
+ for (v_idx = 0; v_idx < vsi->num_q_vectors; v_idx++)
+ ice_free_q_vector(vsi, v_idx);
+}
+
+/**
+ * ice_vsi_alloc_q_vector - Allocate memory for a single interrupt vector
+ * @vsi: the VSI being configured
+ * @v_idx: index of the vector in the VSI struct
+ *
+ * We allocate one q_vector. If allocation fails we return -ENOMEM.
+ */
+static int ice_vsi_alloc_q_vector(struct ice_vsi *vsi, int v_idx)
+{
+ struct ice_pf *pf = vsi->back;
+ struct ice_q_vector *q_vector;
+
+ /* allocate q_vector */
+ q_vector = devm_kzalloc(&pf->pdev->dev, sizeof(*q_vector), GFP_KERNEL);
+ if (!q_vector)
+ return -ENOMEM;
+
+ q_vector->vsi = vsi;
+ q_vector->v_idx = v_idx;
+ /* only set affinity_mask if the CPU is online */
+ if (cpu_online(v_idx))
+ cpumask_set_cpu(v_idx, &q_vector->affinity_mask);
+
+ /* This will not be called in the driver load path because the netdev
+ * will not be created yet. All other cases with register the NAPI
+ * handler here (i.e. resume, reset/rebuild, etc.)
+ */
+ if (vsi->netdev)
+ netif_napi_add(vsi->netdev, &q_vector->napi, ice_napi_poll,
+ NAPI_POLL_WEIGHT);
+
+ /* tie q_vector and VSI together */
+ vsi->q_vectors[v_idx] = q_vector;
+
+ return 0;
+}
+
+/**
+ * ice_vsi_alloc_q_vectors - Allocate memory for interrupt vectors
+ * @vsi: the VSI being configured
+ *
+ * We allocate one q_vector per queue interrupt. If allocation fails we
+ * return -ENOMEM.
+ */
+int ice_vsi_alloc_q_vectors(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ int v_idx = 0, num_q_vectors;
+ int err;
+
+ if (vsi->q_vectors[0]) {
+ dev_dbg(&pf->pdev->dev, "VSI %d has existing q_vectors\n",
+ vsi->vsi_num);
+ return -EEXIST;
+ }
+
+ if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags)) {
+ num_q_vectors = vsi->num_q_vectors;
+ } else {
+ err = -EINVAL;
+ goto err_out;
+ }
+
+ for (v_idx = 0; v_idx < num_q_vectors; v_idx++) {
+ err = ice_vsi_alloc_q_vector(vsi, v_idx);
+ if (err)
+ goto err_out;
+ }
+
+ return 0;
+
+err_out:
+ while (v_idx--)
+ ice_free_q_vector(vsi, v_idx);
+
+ dev_err(&pf->pdev->dev,
+ "Failed to allocate %d q_vector for VSI %d, ret=%d\n",
+ vsi->num_q_vectors, vsi->vsi_num, err);
+ vsi->num_q_vectors = 0;
+ return err;
+}
+
+/**
+ * ice_vsi_setup_vector_base - Set up the base vector for the given VSI
+ * @vsi: ptr to the VSI
+ *
+ * This should only be called after ice_vsi_alloc() which allocates the
+ * corresponding SW VSI structure and initializes num_queue_pairs for the
+ * newly allocated VSI.
+ *
+ * Returns 0 on success or negative on failure
+ */
+int ice_vsi_setup_vector_base(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ int num_q_vectors = 0;
+
+ if (vsi->base_vector) {
+ dev_dbg(&pf->pdev->dev, "VSI %d has non-zero base vector %d\n",
+ vsi->vsi_num, vsi->base_vector);
+ return -EEXIST;
+ }
+
+ if (!test_bit(ICE_FLAG_MSIX_ENA, pf->flags))
+ return -ENOENT;
+
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ num_q_vectors = vsi->num_q_vectors;
+ break;
+ default:
+ dev_warn(&vsi->back->pdev->dev, "Unknown VSI type %d\n",
+ vsi->type);
+ break;
+ }
+
+ if (num_q_vectors)
+ vsi->base_vector = ice_get_res(pf, pf->irq_tracker,
+ num_q_vectors, vsi->idx);
+
+ if (vsi->base_vector < 0) {
+ dev_err(&pf->pdev->dev,
+ "Failed to get tracking for %d vectors for VSI %d, err=%d\n",
+ num_q_vectors, vsi->vsi_num, vsi->base_vector);
+ return -ENOENT;
+ }
+
+ return 0;
+}
+
/**
* ice_vsi_clear_rings - Deallocates the Tx and Rx rings for VSI
* @vsi: the VSI having rings deallocated
*/
-void ice_vsi_clear_rings(struct ice_vsi *vsi)
+static void ice_vsi_clear_rings(struct ice_vsi *vsi)
{
int i;
@@ -1674,6 +1997,142 @@ void ice_vsi_dis_irq(struct ice_vsi *vsi)
}
}
+/**
+ * ice_vsi_release - Delete a VSI and free its resources
+ * @vsi: the VSI being removed
+ *
+ * Returns 0 on success or < 0 on error
+ */
+int ice_vsi_release(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf;
+
+ if (!vsi->back)
+ return -ENODEV;
+ pf = vsi->back;
+ /* do not unregister and free netdevs while driver is in the reset
+ * recovery pending state. Since reset/rebuild happens through PF
+ * service task workqueue, its not a good idea to unregister netdev
+ * that is associated to the PF that is running the work queue items
+ * currently. This is done to avoid check_flush_dependency() warning
+ * on this wq
+ */
+ if (vsi->netdev && !ice_is_reset_recovery_pending(pf->state)) {
+ unregister_netdev(vsi->netdev);
+ free_netdev(vsi->netdev);
+ vsi->netdev = NULL;
+ }
+
+ if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
+ ice_rss_clean(vsi);
+
+ /* Disable VSI and free resources */
+ ice_vsi_dis_irq(vsi);
+ ice_vsi_close(vsi);
+
+ /* reclaim interrupt vectors back to PF */
+ ice_free_res(vsi->back->irq_tracker, vsi->base_vector, vsi->idx);
+ pf->num_avail_msix += vsi->num_q_vectors;
+
+ ice_remove_vsi_fltr(&pf->hw, vsi->vsi_num);
+ ice_vsi_delete(vsi);
+ ice_vsi_free_q_vectors(vsi);
+ ice_vsi_clear_rings(vsi);
+
+ ice_vsi_put_qs(vsi);
+ pf->q_left_tx += vsi->alloc_txq;
+ pf->q_left_rx += vsi->alloc_rxq;
+
+ /* retain SW VSI data structure since it is needed to unregister and
+ * free VSI netdev when PF is not in reset recovery pending state,\
+ * for ex: during rmmod.
+ */
+ if (!ice_is_reset_recovery_pending(pf->state))
+ ice_vsi_clear(vsi);
+
+ return 0;
+}
+
+/**
+ * ice_vsi_rebuild - Rebuild VSI after reset
+ * @vsi: VSI to be rebuild
+ *
+ * Returns 0 on success and negative value on failure
+ */
+int ice_vsi_rebuild(struct ice_vsi *vsi)
+{
+ u16 max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 };
+ int ret, i;
+
+ if (!vsi)
+ return -EINVAL;
+
+ ice_vsi_free_q_vectors(vsi);
+ ice_free_res(vsi->back->irq_tracker, vsi->base_vector, vsi->idx);
+ vsi->base_vector = 0;
+ ice_vsi_clear_rings(vsi);
+ ice_vsi_free_arrays(vsi, false);
+ ice_vsi_set_num_qs(vsi);
+
+ /* Initialize VSI struct elements and create VSI in FW */
+ ret = ice_vsi_init(vsi);
+ if (ret < 0)
+ goto err_vsi;
+
+ ret = ice_vsi_alloc_arrays(vsi, false);
+ if (ret < 0)
+ goto err_vsi;
+
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ ret = ice_vsi_alloc_q_vectors(vsi);
+ if (ret)
+ goto err_rings;
+
+ ret = ice_vsi_setup_vector_base(vsi);
+ if (ret)
+ goto err_vectors;
+
+ ret = ice_vsi_alloc_rings(vsi);
+ if (ret)
+ goto err_vectors;
+
+ ice_vsi_map_rings_to_vectors(vsi);
+ break;
+ default:
+ break;
+ }
+
+ ice_vsi_set_tc_cfg(vsi);
+
+ /* configure VSI nodes based on number of queues and TC's */
+ for (i = 0; i < vsi->tc_cfg.numtc; i++)
+ max_txqs[i] = vsi->num_txq;
+
+ ret = ice_cfg_vsi_lan(vsi->port_info, vsi->vsi_num,
+ vsi->tc_cfg.ena_tc, max_txqs);
+ if (ret) {
+ dev_info(&vsi->back->pdev->dev,
+ "Failed VSI lan queue config\n");
+ goto err_vectors;
+ }
+ return 0;
+
+err_vectors:
+ ice_vsi_free_q_vectors(vsi);
+err_rings:
+ if (vsi->netdev) {
+ vsi->current_netdev_flags = 0;
+ unregister_netdev(vsi->netdev);
+ free_netdev(vsi->netdev);
+ vsi->netdev = NULL;
+ }
+err_vsi:
+ ice_vsi_clear(vsi);
+ set_bit(__ICE_RESET_FAILED, vsi->back->state);
+ return ret;
+}
+
/**
* ice_is_reset_recovery_pending - schedule a reset
* @state: pf state field
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
index 002bbca8e7ea..aaab3fc4b018 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -6,6 +6,12 @@
#include "ice.h"
+int ice_vsi_setup_vector_base(struct ice_vsi *vsi);
+
+int ice_vsi_alloc_q_vectors(struct ice_vsi *vsi);
+
+int ice_vsi_get_qs(struct ice_vsi *vsi);
+
void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi);
int ice_vsi_alloc_rings(struct ice_vsi *vsi);
@@ -18,10 +24,6 @@ int ice_get_free_slot(void *array, int size, int curr);
int ice_vsi_init(struct ice_vsi *vsi);
-void ice_vsi_free_arrays(struct ice_vsi *vsi, bool free_qvectors);
-
-void ice_vsi_clear_rings(struct ice_vsi *vsi);
-
int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors);
int ice_add_mac_to_list(struct ice_vsi *vsi, struct list_head *add_list,
@@ -57,6 +59,8 @@ void ice_vsi_delete(struct ice_vsi *vsi);
int ice_vsi_clear(struct ice_vsi *vsi);
+int ice_vsi_release(struct ice_vsi *vsi);
+
void ice_vsi_close(struct ice_vsi *vsi);
int ice_free_res(struct ice_res_tracker *res, u16 index, u16 id);
@@ -64,8 +68,12 @@ int ice_free_res(struct ice_res_tracker *res, u16 index, u16 id);
int
ice_get_res(struct ice_pf *pf, struct ice_res_tracker *res, u16 needed, u16 id);
+int ice_vsi_rebuild(struct ice_vsi *vsi);
+
bool ice_is_reset_recovery_pending(unsigned long *state);
+void ice_vsi_free_q_vectors(struct ice_vsi *vsi);
+
void ice_vsi_put_qs(struct ice_vsi *vsi);
void ice_vsi_dis_irq(struct ice_vsi *vsi);
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index f458ff285a34..3927b18b45a0 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -32,7 +32,6 @@ static const struct net_device_ops ice_netdev_ops;
static void ice_pf_dis_all_vsi(struct ice_pf *pf);
static void ice_rebuild(struct ice_pf *pf);
-static int ice_vsi_release(struct ice_vsi *vsi);
static void ice_vsi_release_all(struct ice_pf *pf);
static void ice_update_vsi_stats(struct ice_vsi *vsi);
@@ -1465,185 +1464,43 @@ static int ice_req_irq_msix_misc(struct ice_pf *pf)
}
/**
- * ice_vsi_get_qs_contig - Assign a contiguous chunk of queues to VSI
- * @vsi: the VSI getting queues
- *
- * Return 0 on success and a negative value on error
- */
-static int ice_vsi_get_qs_contig(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- int offset, ret = 0;
-
- mutex_lock(&pf->avail_q_mutex);
- /* look for contiguous block of queues for tx */
- offset = bitmap_find_next_zero_area(pf->avail_txqs, ICE_MAX_TXQS,
- 0, vsi->alloc_txq, 0);
- if (offset < ICE_MAX_TXQS) {
- int i;
-
- bitmap_set(pf->avail_txqs, offset, vsi->alloc_txq);
- for (i = 0; i < vsi->alloc_txq; i++)
- vsi->txq_map[i] = i + offset;
- } else {
- ret = -ENOMEM;
- vsi->tx_mapping_mode = ICE_VSI_MAP_SCATTER;
- }
-
- /* look for contiguous block of queues for rx */
- offset = bitmap_find_next_zero_area(pf->avail_rxqs, ICE_MAX_RXQS,
- 0, vsi->alloc_rxq, 0);
- if (offset < ICE_MAX_RXQS) {
- int i;
-
- bitmap_set(pf->avail_rxqs, offset, vsi->alloc_rxq);
- for (i = 0; i < vsi->alloc_rxq; i++)
- vsi->rxq_map[i] = i + offset;
- } else {
- ret = -ENOMEM;
- vsi->rx_mapping_mode = ICE_VSI_MAP_SCATTER;
- }
- mutex_unlock(&pf->avail_q_mutex);
-
- return ret;
-}
-
-/**
- * ice_vsi_get_qs_scatter - Assign a scattered queues to VSI
- * @vsi: the VSI getting queues
- *
- * Return 0 on success and a negative value on error
+ * ice_napi_del - Remove NAPI handler for the VSI
+ * @vsi: VSI for which NAPI handler is to be removed
*/
-static int ice_vsi_get_qs_scatter(struct ice_vsi *vsi)
+static void ice_napi_del(struct ice_vsi *vsi)
{
- struct ice_pf *pf = vsi->back;
- int i, index = 0;
-
- mutex_lock(&pf->avail_q_mutex);
-
- if (vsi->tx_mapping_mode == ICE_VSI_MAP_SCATTER) {
- for (i = 0; i < vsi->alloc_txq; i++) {
- index = find_next_zero_bit(pf->avail_txqs,
- ICE_MAX_TXQS, index);
- if (index < ICE_MAX_TXQS) {
- set_bit(index, pf->avail_txqs);
- vsi->txq_map[i] = index;
- } else {
- goto err_scatter_tx;
- }
- }
- }
-
- if (vsi->rx_mapping_mode == ICE_VSI_MAP_SCATTER) {
- for (i = 0; i < vsi->alloc_rxq; i++) {
- index = find_next_zero_bit(pf->avail_rxqs,
- ICE_MAX_RXQS, index);
- if (index < ICE_MAX_RXQS) {
- set_bit(index, pf->avail_rxqs);
- vsi->rxq_map[i] = index;
- } else {
- goto err_scatter_rx;
- }
- }
- }
-
- mutex_unlock(&pf->avail_q_mutex);
- return 0;
+ int v_idx;
-err_scatter_rx:
- /* unflag any queues we have grabbed (i is failed position) */
- for (index = 0; index < i; index++) {
- clear_bit(vsi->rxq_map[index], pf->avail_rxqs);
- vsi->rxq_map[index] = 0;
- }
- i = vsi->alloc_txq;
-err_scatter_tx:
- /* i is either position of failed attempt or vsi->alloc_txq */
- for (index = 0; index < i; index++) {
- clear_bit(vsi->txq_map[index], pf->avail_txqs);
- vsi->txq_map[index] = 0;
- }
+ if (!vsi->netdev)
+ return;
- mutex_unlock(&pf->avail_q_mutex);
- return -ENOMEM;
+ for (v_idx = 0; v_idx < vsi->num_q_vectors; v_idx++)
+ netif_napi_del(&vsi->q_vectors[v_idx]->napi);
}
/**
- * ice_vsi_get_qs - Assign queues from PF to VSI
- * @vsi: the VSI to assign queues to
+ * ice_napi_add - register NAPI handler for the VSI
+ * @vsi: VSI for which NAPI handler is to be registered
*
- * Returns 0 on success and a negative value on error
+ * This function is only called in the driver's load path. Registering the NAPI
+ * handler is done in ice_vsi_alloc_q_vector() for all other cases (i.e. resume,
+ * reset/rebuild, etc.)
*/
-static int ice_vsi_get_qs(struct ice_vsi *vsi)
+static void ice_napi_add(struct ice_vsi *vsi)
{
- int ret = 0;
-
- vsi->tx_mapping_mode = ICE_VSI_MAP_CONTIG;
- vsi->rx_mapping_mode = ICE_VSI_MAP_CONTIG;
-
- /* NOTE: ice_vsi_get_qs_contig() will set the rx/tx mapping
- * modes individually to scatter if assigning contiguous queues
- * to rx or tx fails
- */
- ret = ice_vsi_get_qs_contig(vsi);
- if (ret < 0) {
- if (vsi->tx_mapping_mode == ICE_VSI_MAP_SCATTER)
- vsi->alloc_txq = max_t(u16, vsi->alloc_txq,
- ICE_MAX_SCATTER_TXQS);
- if (vsi->rx_mapping_mode == ICE_VSI_MAP_SCATTER)
- vsi->alloc_rxq = max_t(u16, vsi->alloc_rxq,
- ICE_MAX_SCATTER_RXQS);
- ret = ice_vsi_get_qs_scatter(vsi);
- }
-
- return ret;
-}
-
-/**
- * ice_free_q_vector - Free memory allocated for a specific interrupt vector
- * @vsi: VSI having the memory freed
- * @v_idx: index of the vector to be freed
- */
-static void ice_free_q_vector(struct ice_vsi *vsi, int v_idx)
-{
- struct ice_q_vector *q_vector;
- struct ice_ring *ring;
+ int v_idx;
- if (!vsi->q_vectors[v_idx]) {
- dev_dbg(&vsi->back->pdev->dev, "Queue vector at index %d not found\n",
- v_idx);
+ if (!vsi->netdev)
return;
- }
- q_vector = vsi->q_vectors[v_idx];
-
- ice_for_each_ring(ring, q_vector->tx)
- ring->q_vector = NULL;
- ice_for_each_ring(ring, q_vector->rx)
- ring->q_vector = NULL;
-
- /* only VSI with an associated netdev is set up with NAPI */
- if (vsi->netdev)
- netif_napi_del(&q_vector->napi);
-
- devm_kfree(&vsi->back->pdev->dev, q_vector);
- vsi->q_vectors[v_idx] = NULL;
-}
-
-/**
- * ice_vsi_free_q_vectors - Free memory allocated for interrupt vectors
- * @vsi: the VSI having memory freed
- */
-static void ice_vsi_free_q_vectors(struct ice_vsi *vsi)
-{
- int v_idx;
for (v_idx = 0; v_idx < vsi->num_q_vectors; v_idx++)
- ice_free_q_vector(vsi, v_idx);
+ netif_napi_add(vsi->netdev, &vsi->q_vectors[v_idx]->napi,
+ ice_napi_poll, NAPI_POLL_WEIGHT);
}
/**
- * ice_cfg_netdev - Setup the netdev flags
- * @vsi: the VSI being configured
+ * ice_cfg_netdev - Allocate, configure and register a netdev
+ * @vsi: the VSI associated with the new netdev
*
* Returns 0 on success, negative value on failure
*/
@@ -1656,6 +1513,7 @@ static int ice_cfg_netdev(struct ice_vsi *vsi)
struct ice_netdev_priv *np;
struct net_device *netdev;
u8 mac_addr[ETH_ALEN];
+ int err;
netdev = alloc_etherdev_mqs(sizeof(struct ice_netdev_priv),
vsi->alloc_txq, vsi->alloc_rxq);
@@ -1713,130 +1571,14 @@ static int ice_cfg_netdev(struct ice_vsi *vsi)
netdev->min_mtu = ETH_MIN_MTU;
netdev->max_mtu = ICE_MAX_MTU;
- return 0;
-}
-
-/**
- * ice_vsi_alloc_q_vector - Allocate memory for a single interrupt vector
- * @vsi: the VSI being configured
- * @v_idx: index of the vector in the vsi struct
- *
- * We allocate one q_vector. If allocation fails we return -ENOMEM.
- */
-static int ice_vsi_alloc_q_vector(struct ice_vsi *vsi, int v_idx)
-{
- struct ice_pf *pf = vsi->back;
- struct ice_q_vector *q_vector;
-
- /* allocate q_vector */
- q_vector = devm_kzalloc(&pf->pdev->dev, sizeof(*q_vector), GFP_KERNEL);
- if (!q_vector)
- return -ENOMEM;
-
- q_vector->vsi = vsi;
- q_vector->v_idx = v_idx;
- /* only set affinity_mask if the CPU is online */
- if (cpu_online(v_idx))
- cpumask_set_cpu(v_idx, &q_vector->affinity_mask);
-
- if (vsi->netdev)
- netif_napi_add(vsi->netdev, &q_vector->napi, ice_napi_poll,
- NAPI_POLL_WEIGHT);
- /* tie q_vector and vsi together */
- vsi->q_vectors[v_idx] = q_vector;
-
- return 0;
-}
-
-/**
- * ice_vsi_alloc_q_vectors - Allocate memory for interrupt vectors
- * @vsi: the VSI being configured
- *
- * We allocate one q_vector per queue interrupt. If allocation fails we
- * return -ENOMEM.
- */
-static int ice_vsi_alloc_q_vectors(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- int v_idx = 0, num_q_vectors;
- int err;
-
- if (vsi->q_vectors[0]) {
- dev_dbg(&pf->pdev->dev, "VSI %d has existing q_vectors\n",
- vsi->vsi_num);
- return -EEXIST;
- }
-
- if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags)) {
- num_q_vectors = vsi->num_q_vectors;
- } else {
- err = -EINVAL;
- goto err_out;
- }
-
- for (v_idx = 0; v_idx < num_q_vectors; v_idx++) {
- err = ice_vsi_alloc_q_vector(vsi, v_idx);
- if (err)
- goto err_out;
- }
-
- return 0;
-
-err_out:
- while (v_idx--)
- ice_free_q_vector(vsi, v_idx);
-
- dev_err(&pf->pdev->dev,
- "Failed to allocate %d q_vector for VSI %d, ret=%d\n",
- vsi->num_q_vectors, vsi->vsi_num, err);
- vsi->num_q_vectors = 0;
- return err;
-}
-
-/**
- * ice_vsi_setup_vector_base - Set up the base vector for the given VSI
- * @vsi: ptr to the VSI
- *
- * This should only be called after ice_vsi_alloc() which allocates the
- * corresponding SW VSI structure and initializes num_queue_pairs for the
- * newly allocated VSI.
- *
- * Returns 0 on success or negative on failure
- */
-static int ice_vsi_setup_vector_base(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- int num_q_vectors = 0;
-
- if (vsi->base_vector) {
- dev_dbg(&pf->pdev->dev, "VSI %d has non-zero base vector %d\n",
- vsi->vsi_num, vsi->base_vector);
- return -EEXIST;
- }
-
- if (!test_bit(ICE_FLAG_MSIX_ENA, pf->flags))
- return -ENOENT;
-
- switch (vsi->type) {
- case ICE_VSI_PF:
- num_q_vectors = vsi->num_q_vectors;
- break;
- default:
- dev_warn(&vsi->back->pdev->dev, "Unknown VSI type %d\n",
- vsi->type);
- break;
- }
+ err = register_netdev(vsi->netdev);
+ if (err)
+ return err;
- if (num_q_vectors)
- vsi->base_vector = ice_get_res(pf, pf->irq_tracker,
- num_q_vectors, vsi->idx);
+ netif_carrier_off(vsi->netdev);
- if (vsi->base_vector < 0) {
- dev_err(&pf->pdev->dev,
- "Failed to get tracking for %d vectors for VSI %d, err=%d\n",
- num_q_vectors, vsi->vsi_num, vsi->base_vector);
- return -ENOENT;
- }
+ /* make sure transmit queues start off as stopped */
+ netif_tx_stop_all_queues(vsi->netdev);
return 0;
}
@@ -1918,87 +1660,6 @@ static int ice_vsi_cfg_rss(struct ice_vsi *vsi)
return err;
}
-/**
- * ice_vsi_rebuild - Rebuild VSI after reset
- * @vsi: vsi to be rebuild
- *
- * Returns 0 on success and negative value on failure
- */
-static int ice_vsi_rebuild(struct ice_vsi *vsi)
-{
- u16 max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 };
- int ret, i;
-
- if (!vsi)
- return -EINVAL;
-
- ice_vsi_free_q_vectors(vsi);
- ice_free_res(vsi->back->irq_tracker, vsi->base_vector, vsi->idx);
- vsi->base_vector = 0;
- ice_vsi_clear_rings(vsi);
- ice_vsi_free_arrays(vsi, false);
- ice_vsi_set_num_qs(vsi);
-
- /* Initialize VSI struct elements and create VSI in FW */
- ret = ice_vsi_init(vsi);
- if (ret < 0)
- goto err_vsi;
-
- ret = ice_vsi_alloc_arrays(vsi, false);
- if (ret < 0)
- goto err_vsi;
-
- switch (vsi->type) {
- case ICE_VSI_PF:
- /* fall through */
- ret = ice_vsi_alloc_q_vectors(vsi);
- if (ret)
- goto err_rings;
-
- ret = ice_vsi_setup_vector_base(vsi);
- if (ret)
- goto err_vectors;
-
- ret = ice_vsi_alloc_rings(vsi);
- if (ret)
- goto err_vectors;
-
- ice_vsi_map_rings_to_vectors(vsi);
- break;
- default:
- break;
- }
-
- ice_vsi_set_tc_cfg(vsi);
-
- /* configure VSI nodes based on number of queues and TC's */
- for (i = 0; i < vsi->tc_cfg.numtc; i++)
- max_txqs[i] = vsi->num_txq;
-
- ret = ice_cfg_vsi_lan(vsi->port_info, vsi->vsi_num,
- vsi->tc_cfg.ena_tc, max_txqs);
- if (ret) {
- dev_info(&vsi->back->pdev->dev,
- "Failed VSI lan queue config\n");
- goto err_vectors;
- }
- return 0;
-
-err_vectors:
- ice_vsi_free_q_vectors(vsi);
-err_rings:
- if (vsi->netdev) {
- vsi->current_netdev_flags = 0;
- unregister_netdev(vsi->netdev);
- free_netdev(vsi->netdev);
- vsi->netdev = NULL;
- }
-err_vsi:
- ice_vsi_clear(vsi);
- set_bit(__ICE_RESET_FAILED, vsi->back->state);
- return ret;
-}
-
/**
* ice_vsi_setup - Set up a VSI by a given type
* @pf: board private structure
@@ -2237,6 +1898,18 @@ static int ice_setup_pf_sw(struct ice_pf *pf)
goto unroll_vsi_setup;
}
+ status = ice_cfg_netdev(vsi);
+ if (status) {
+ status = -ENODEV;
+ goto unroll_vsi_setup;
+ }
+
+ /* registering the NAPI handler requires both the queues and
+ * netdev to be created, which are done in ice_pf_vsi_setup()
+ * and ice_cfg_netdev() respectively
+ */
+ ice_napi_add(vsi);
+
/* To add a MAC filter, first add the MAC to a list and then
* pass the list to ice_add_mac.
*/
@@ -2245,7 +1918,7 @@ static int ice_setup_pf_sw(struct ice_pf *pf)
status = ice_add_mac_to_list(vsi, &tmp_add_list,
vsi->port_info->mac.perm_addr);
if (status)
- goto unroll_vsi_setup;
+ goto unroll_napi_add;
/* VSI needs to receive broadcast traffic, so add the broadcast
* MAC address to the list as well.
@@ -2269,16 +1942,20 @@ static int ice_setup_pf_sw(struct ice_pf *pf)
free_mac_list:
ice_free_fltr_list(&pf->pdev->dev, &tmp_add_list);
-unroll_vsi_setup:
+unroll_napi_add:
if (vsi) {
- ice_vsi_free_q_vectors(vsi);
- if (vsi->netdev && vsi->netdev->reg_state == NETREG_REGISTERED)
- unregister_netdev(vsi->netdev);
+ ice_napi_del(vsi);
if (vsi->netdev) {
+ if (vsi->netdev->reg_state == NETREG_REGISTERED)
+ unregister_netdev(vsi->netdev);
free_netdev(vsi->netdev);
vsi->netdev = NULL;
}
+ }
+unroll_vsi_setup:
+ if (vsi) {
+ ice_vsi_free_q_vectors(vsi);
ice_vsi_delete(vsi);
ice_vsi_put_qs(vsi);
pf->q_left_tx += vsi->alloc_txq;
@@ -3590,79 +3267,6 @@ static int ice_vsi_open(struct ice_vsi *vsi)
return err;
}
-
-/**
- * ice_rss_clean - Delete RSS related VSI structures that hold user inputs
- * @vsi: the VSI being removed
- */
-static void ice_rss_clean(struct ice_vsi *vsi)
-{
- struct ice_pf *pf;
-
- pf = vsi->back;
-
- if (vsi->rss_hkey_user)
- devm_kfree(&pf->pdev->dev, vsi->rss_hkey_user);
- if (vsi->rss_lut_user)
- devm_kfree(&pf->pdev->dev, vsi->rss_lut_user);
-}
-
-/**
- * ice_vsi_release - Delete a VSI and free its resources
- * @vsi: the VSI being removed
- *
- * Returns 0 on success or < 0 on error
- */
-static int ice_vsi_release(struct ice_vsi *vsi)
-{
- struct ice_pf *pf;
-
- if (!vsi->back)
- return -ENODEV;
- pf = vsi->back;
- /* do not unregister and free netdevs while driver is in the reset
- * recovery pending state. Since reset/rebuild happens through PF
- * service task workqueue, its not a good idea to unregister netdev
- * that is associated to the PF that is running the work queue items
- * currently. This is done to avoid check_flush_dependency() warning
- * on this wq
- */
- if (vsi->netdev && !ice_is_reset_recovery_pending(pf->state)) {
- unregister_netdev(vsi->netdev);
- free_netdev(vsi->netdev);
- vsi->netdev = NULL;
- }
-
- if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
- ice_rss_clean(vsi);
-
- /* Disable VSI and free resources */
- ice_vsi_dis_irq(vsi);
- ice_vsi_close(vsi);
-
- /* reclaim interrupt vectors back to PF */
- ice_free_res(vsi->back->irq_tracker, vsi->base_vector, vsi->idx);
- pf->num_avail_msix += vsi->num_q_vectors;
-
- ice_remove_vsi_fltr(&pf->hw, vsi->vsi_num);
- ice_vsi_delete(vsi);
- ice_vsi_free_q_vectors(vsi);
- ice_vsi_clear_rings(vsi);
-
- ice_vsi_put_qs(vsi);
- pf->q_left_tx += vsi->alloc_txq;
- pf->q_left_rx += vsi->alloc_rxq;
-
- /* retain SW VSI data structure since it is needed to unregister and
- * free VSI netdev when PF is not in reset recovery pending state,\
- * for ex: during rmmod.
- */
- if (!ice_is_reset_recovery_pending(pf->state))
- ice_vsi_clear(vsi);
-
- return 0;
-}
-
/**
* ice_vsi_release_all - Delete all VSIs
* @pf: PF from which all VSIs are being removed
--
2.17.1
^ permalink raw reply related
* [net-next 4/8] ice: Move common functions out of ice_main.c part 4/7
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem
Cc: Anirudh Venkataramanan, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <20181001211431.14325-1-jeffrey.t.kirsher@intel.com>
From: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
This patch continues the code move out of ice_main.c
The following top level functions (and related dependency functions) were
moved to ice_lib.c:
ice_vsi_alloc_rings
ice_vsi_set_rss_params
ice_vsi_set_num_qs
ice_get_free_slot
ice_vsi_init
ice_vsi_clear_rings
ice_vsi_alloc_arrays
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_lib.c | 414 +++++++++++++++++++++
drivers/net/ethernet/intel/ice/ice_lib.h | 14 +
drivers/net/ethernet/intel/ice/ice_main.c | 418 +---------------------
3 files changed, 429 insertions(+), 417 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 474ce5828bd4..df20d68c92ab 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -225,6 +225,102 @@ static int ice_vsi_ctrl_rx_rings(struct ice_vsi *vsi, bool ena)
return ret;
}
+/**
+ * ice_vsi_alloc_arrays - Allocate queue and vector pointer arrays for the VSI
+ * @vsi: VSI pointer
+ * @alloc_qvectors: a bool to specify if q_vectors need to be allocated.
+ *
+ * On error: returns error code (negative)
+ * On success: returns 0
+ */
+int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
+{
+ struct ice_pf *pf = vsi->back;
+
+ /* allocate memory for both Tx and Rx ring pointers */
+ vsi->tx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_txq,
+ sizeof(struct ice_ring *), GFP_KERNEL);
+ if (!vsi->tx_rings)
+ goto err_txrings;
+
+ vsi->rx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_rxq,
+ sizeof(struct ice_ring *), GFP_KERNEL);
+ if (!vsi->rx_rings)
+ goto err_rxrings;
+
+ if (alloc_qvectors) {
+ /* allocate memory for q_vector pointers */
+ vsi->q_vectors = devm_kcalloc(&pf->pdev->dev,
+ vsi->num_q_vectors,
+ sizeof(struct ice_q_vector *),
+ GFP_KERNEL);
+ if (!vsi->q_vectors)
+ goto err_vectors;
+ }
+
+ return 0;
+
+err_vectors:
+ devm_kfree(&pf->pdev->dev, vsi->rx_rings);
+err_rxrings:
+ devm_kfree(&pf->pdev->dev, vsi->tx_rings);
+err_txrings:
+ return -ENOMEM;
+}
+
+/**
+ * ice_vsi_set_num_qs - Set num queues, descriptors and vectors for a VSI
+ * @vsi: the VSI being configured
+ *
+ * Return 0 on success and a negative value on error
+ */
+void ice_vsi_set_num_qs(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ vsi->alloc_txq = pf->num_lan_tx;
+ vsi->alloc_rxq = pf->num_lan_rx;
+ vsi->num_desc = ALIGN(ICE_DFLT_NUM_DESC, ICE_REQ_DESC_MULTIPLE);
+ vsi->num_q_vectors = max_t(int, pf->num_lan_rx, pf->num_lan_tx);
+ break;
+ default:
+ dev_warn(&vsi->back->pdev->dev, "Unknown VSI type %d\n",
+ vsi->type);
+ break;
+ }
+}
+
+/**
+ * ice_get_free_slot - get the next non-NULL location index in array
+ * @array: array to search
+ * @size: size of the array
+ * @curr: last known occupied index to be used as a search hint
+ *
+ * void * is being used to keep the functionality generic. This lets us use this
+ * function on any array of pointers.
+ */
+int ice_get_free_slot(void *array, int size, int curr)
+{
+ int **tmp_array = (int **)array;
+ int next;
+
+ if (curr < (size - 1) && !tmp_array[curr + 1]) {
+ next = curr + 1;
+ } else {
+ int i = 0;
+
+ while ((i < size) && (tmp_array[i]))
+ i++;
+ if (i == size)
+ next = ICE_NO_VSI;
+ else
+ next = i;
+ }
+ return next;
+}
+
/**
* ice_vsi_delete - delete a VSI from the switch
* @vsi: pointer to VSI being removed
@@ -286,6 +382,324 @@ void ice_vsi_put_qs(struct ice_vsi *vsi)
mutex_unlock(&pf->avail_q_mutex);
}
+/**
+ * ice_vsi_set_rss_params - Setup RSS capabilities per VSI type
+ * @vsi: the VSI being configured
+ */
+void ice_vsi_set_rss_params(struct ice_vsi *vsi)
+{
+ struct ice_hw_common_caps *cap;
+ struct ice_pf *pf = vsi->back;
+
+ if (!test_bit(ICE_FLAG_RSS_ENA, pf->flags)) {
+ vsi->rss_size = 1;
+ return;
+ }
+
+ cap = &pf->hw.func_caps.common_cap;
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ /* PF VSI will inherit RSS instance of PF */
+ vsi->rss_table_size = cap->rss_table_size;
+ vsi->rss_size = min_t(int, num_online_cpus(),
+ BIT(cap->rss_table_entry_width));
+ vsi->rss_lut_type = ICE_AQC_GSET_RSS_LUT_TABLE_TYPE_PF;
+ break;
+ default:
+ dev_warn(&pf->pdev->dev, "Unknown VSI type %d\n",
+ vsi->type);
+ break;
+ }
+}
+
+/**
+ * ice_set_dflt_vsi_ctx - Set default VSI context before adding a VSI
+ * @ctxt: the VSI context being set
+ *
+ * This initializes a default VSI context for all sections except the Queues.
+ */
+static void ice_set_dflt_vsi_ctx(struct ice_vsi_ctx *ctxt)
+{
+ u32 table = 0;
+
+ memset(&ctxt->info, 0, sizeof(ctxt->info));
+ /* VSI's should be allocated from shared pool */
+ ctxt->alloc_from_pool = true;
+ /* Src pruning enabled by default */
+ ctxt->info.sw_flags = ICE_AQ_VSI_SW_FLAG_SRC_PRUNE;
+ /* Traffic from VSI can be sent to LAN */
+ ctxt->info.sw_flags2 = ICE_AQ_VSI_SW_FLAG_LAN_ENA;
+ /* By default bits 3 and 4 in vlan_flags are 0's which results in legacy
+ * behavior (show VLAN, DEI, and UP) in descriptor. Also, allow all
+ * packets untagged/tagged.
+ */
+ ctxt->info.vlan_flags = ((ICE_AQ_VSI_VLAN_MODE_ALL &
+ ICE_AQ_VSI_VLAN_MODE_M) >>
+ ICE_AQ_VSI_VLAN_MODE_S);
+ /* Have 1:1 UP mapping for both ingress/egress tables */
+ table |= ICE_UP_TABLE_TRANSLATE(0, 0);
+ table |= ICE_UP_TABLE_TRANSLATE(1, 1);
+ table |= ICE_UP_TABLE_TRANSLATE(2, 2);
+ table |= ICE_UP_TABLE_TRANSLATE(3, 3);
+ table |= ICE_UP_TABLE_TRANSLATE(4, 4);
+ table |= ICE_UP_TABLE_TRANSLATE(5, 5);
+ table |= ICE_UP_TABLE_TRANSLATE(6, 6);
+ table |= ICE_UP_TABLE_TRANSLATE(7, 7);
+ ctxt->info.ingress_table = cpu_to_le32(table);
+ ctxt->info.egress_table = cpu_to_le32(table);
+ /* Have 1:1 UP mapping for outer to inner UP table */
+ ctxt->info.outer_up_table = cpu_to_le32(table);
+ /* No Outer tag support outer_tag_flags remains to zero */
+}
+
+/**
+ * ice_vsi_setup_q_map - Setup a VSI queue map
+ * @vsi: the VSI being configured
+ * @ctxt: VSI context structure
+ */
+static void ice_vsi_setup_q_map(struct ice_vsi *vsi, struct ice_vsi_ctx *ctxt)
+{
+ u16 offset = 0, qmap = 0, numq_tc;
+ u16 pow = 0, max_rss = 0, qcount;
+ u16 qcount_tx = vsi->alloc_txq;
+ u16 qcount_rx = vsi->alloc_rxq;
+ bool ena_tc0 = false;
+ int i;
+
+ /* at least TC0 should be enabled by default */
+ if (vsi->tc_cfg.numtc) {
+ if (!(vsi->tc_cfg.ena_tc & BIT(0)))
+ ena_tc0 = true;
+ } else {
+ ena_tc0 = true;
+ }
+
+ if (ena_tc0) {
+ vsi->tc_cfg.numtc++;
+ vsi->tc_cfg.ena_tc |= 1;
+ }
+
+ numq_tc = qcount_rx / vsi->tc_cfg.numtc;
+
+ /* TC mapping is a function of the number of Rx queues assigned to the
+ * VSI for each traffic class and the offset of these queues.
+ * The first 10 bits are for queue offset for TC0, next 4 bits for no:of
+ * queues allocated to TC0. No:of queues is a power-of-2.
+ *
+ * If TC is not enabled, the queue offset is set to 0, and allocate one
+ * queue, this way, traffic for the given TC will be sent to the default
+ * queue.
+ *
+ * Setup number and offset of Rx queues for all TCs for the VSI
+ */
+
+ /* qcount will change if RSS is enabled */
+ if (test_bit(ICE_FLAG_RSS_ENA, vsi->back->flags)) {
+ if (vsi->type == ICE_VSI_PF)
+ max_rss = ICE_MAX_LG_RSS_QS;
+ else
+ max_rss = ICE_MAX_SMALL_RSS_QS;
+
+ qcount = min_t(int, numq_tc, max_rss);
+ qcount = min_t(int, qcount, vsi->rss_size);
+ } else {
+ qcount = numq_tc;
+ }
+
+ /* find the (rounded up) power-of-2 of qcount */
+ pow = order_base_2(qcount);
+
+ for (i = 0; i < ICE_MAX_TRAFFIC_CLASS; i++) {
+ if (!(vsi->tc_cfg.ena_tc & BIT(i))) {
+ /* TC is not enabled */
+ vsi->tc_cfg.tc_info[i].qoffset = 0;
+ vsi->tc_cfg.tc_info[i].qcount = 1;
+ ctxt->info.tc_mapping[i] = 0;
+ continue;
+ }
+
+ /* TC is enabled */
+ vsi->tc_cfg.tc_info[i].qoffset = offset;
+ vsi->tc_cfg.tc_info[i].qcount = qcount;
+
+ qmap = ((offset << ICE_AQ_VSI_TC_Q_OFFSET_S) &
+ ICE_AQ_VSI_TC_Q_OFFSET_M) |
+ ((pow << ICE_AQ_VSI_TC_Q_NUM_S) &
+ ICE_AQ_VSI_TC_Q_NUM_M);
+ offset += qcount;
+ ctxt->info.tc_mapping[i] = cpu_to_le16(qmap);
+ }
+
+ vsi->num_txq = qcount_tx;
+ vsi->num_rxq = offset;
+
+ /* Rx queue mapping */
+ ctxt->info.mapping_flags |= cpu_to_le16(ICE_AQ_VSI_Q_MAP_CONTIG);
+ /* q_mapping buffer holds the info for the first queue allocated for
+ * this VSI in the PF space and also the number of queues associated
+ * with this VSI.
+ */
+ ctxt->info.q_mapping[0] = cpu_to_le16(vsi->rxq_map[0]);
+ ctxt->info.q_mapping[1] = cpu_to_le16(vsi->num_rxq);
+}
+
+/**
+ * ice_set_rss_vsi_ctx - Set RSS VSI context before adding a VSI
+ * @ctxt: the VSI context being set
+ * @vsi: the VSI being configured
+ */
+static void ice_set_rss_vsi_ctx(struct ice_vsi_ctx *ctxt, struct ice_vsi *vsi)
+{
+ u8 lut_type, hash_type;
+
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ /* PF VSI will inherit RSS instance of PF */
+ lut_type = ICE_AQ_VSI_Q_OPT_RSS_LUT_PF;
+ hash_type = ICE_AQ_VSI_Q_OPT_RSS_TPLZ;
+ break;
+ default:
+ dev_warn(&vsi->back->pdev->dev, "Unknown VSI type %d\n",
+ vsi->type);
+ return;
+ }
+
+ ctxt->info.q_opt_rss = ((lut_type << ICE_AQ_VSI_Q_OPT_RSS_LUT_S) &
+ ICE_AQ_VSI_Q_OPT_RSS_LUT_M) |
+ ((hash_type << ICE_AQ_VSI_Q_OPT_RSS_HASH_S) &
+ ICE_AQ_VSI_Q_OPT_RSS_HASH_M);
+}
+
+/**
+ * ice_vsi_init - Create and initialize a VSI
+ * @vsi: the VSI being configured
+ *
+ * This initializes a VSI context depending on the VSI type to be added and
+ * passes it down to the add_vsi aq command to create a new VSI.
+ */
+int ice_vsi_init(struct ice_vsi *vsi)
+{
+ struct ice_vsi_ctx ctxt = { 0 };
+ struct ice_pf *pf = vsi->back;
+ struct ice_hw *hw = &pf->hw;
+ int ret = 0;
+
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ ctxt.flags = ICE_AQ_VSI_TYPE_PF;
+ break;
+ default:
+ return -ENODEV;
+ }
+
+ ice_set_dflt_vsi_ctx(&ctxt);
+ /* if the switch is in VEB mode, allow VSI loopback */
+ if (vsi->vsw->bridge_mode == BRIDGE_MODE_VEB)
+ ctxt.info.sw_flags |= ICE_AQ_VSI_SW_FLAG_ALLOW_LB;
+
+ /* Set LUT type and HASH type if RSS is enabled */
+ if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
+ ice_set_rss_vsi_ctx(&ctxt, vsi);
+
+ ctxt.info.sw_id = vsi->port_info->sw_id;
+ ice_vsi_setup_q_map(vsi, &ctxt);
+
+ ret = ice_add_vsi(hw, vsi->idx, &ctxt, NULL);
+ if (ret) {
+ dev_err(&pf->pdev->dev,
+ "Add VSI failed, err %d\n", ret);
+ return -EIO;
+ }
+
+ /* keep context for update VSI operations */
+ vsi->info = ctxt.info;
+
+ /* record VSI number returned */
+ vsi->vsi_num = ctxt.vsi_num;
+
+ return ret;
+}
+
+/**
+ * ice_vsi_clear_rings - Deallocates the Tx and Rx rings for VSI
+ * @vsi: the VSI having rings deallocated
+ */
+void ice_vsi_clear_rings(struct ice_vsi *vsi)
+{
+ int i;
+
+ if (vsi->tx_rings) {
+ for (i = 0; i < vsi->alloc_txq; i++) {
+ if (vsi->tx_rings[i]) {
+ kfree_rcu(vsi->tx_rings[i], rcu);
+ vsi->tx_rings[i] = NULL;
+ }
+ }
+ }
+ if (vsi->rx_rings) {
+ for (i = 0; i < vsi->alloc_rxq; i++) {
+ if (vsi->rx_rings[i]) {
+ kfree_rcu(vsi->rx_rings[i], rcu);
+ vsi->rx_rings[i] = NULL;
+ }
+ }
+ }
+}
+
+/**
+ * ice_vsi_alloc_rings - Allocates Tx and Rx rings for the VSI
+ * @vsi: VSI which is having rings allocated
+ */
+int ice_vsi_alloc_rings(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ int i;
+
+ /* Allocate tx_rings */
+ for (i = 0; i < vsi->alloc_txq; i++) {
+ struct ice_ring *ring;
+
+ /* allocate with kzalloc(), free with kfree_rcu() */
+ ring = kzalloc(sizeof(*ring), GFP_KERNEL);
+
+ if (!ring)
+ goto err_out;
+
+ ring->q_index = i;
+ ring->reg_idx = vsi->txq_map[i];
+ ring->ring_active = false;
+ ring->vsi = vsi;
+ ring->dev = &pf->pdev->dev;
+ ring->count = vsi->num_desc;
+ vsi->tx_rings[i] = ring;
+ }
+
+ /* Allocate rx_rings */
+ for (i = 0; i < vsi->alloc_rxq; i++) {
+ struct ice_ring *ring;
+
+ /* allocate with kzalloc(), free with kfree_rcu() */
+ ring = kzalloc(sizeof(*ring), GFP_KERNEL);
+ if (!ring)
+ goto err_out;
+
+ ring->q_index = i;
+ ring->reg_idx = vsi->rxq_map[i];
+ ring->ring_active = false;
+ ring->vsi = vsi;
+ ring->netdev = vsi->netdev;
+ ring->dev = &pf->pdev->dev;
+ ring->count = vsi->num_desc;
+ vsi->rx_rings[i] = ring;
+ }
+
+ return 0;
+
+err_out:
+ ice_vsi_clear_rings(vsi);
+ return -ENOMEM;
+}
+
/**
* ice_add_mac_to_list - Add a mac address filter entry to the list
* @vsi: the VSI to be forwarded to
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
index cbde99f38798..ab5a4e1edcf8 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -6,6 +6,20 @@
#include "ice.h"
+int ice_vsi_alloc_rings(struct ice_vsi *vsi);
+
+void ice_vsi_set_rss_params(struct ice_vsi *vsi);
+
+void ice_vsi_set_num_qs(struct ice_vsi *vsi);
+
+int ice_get_free_slot(void *array, int size, int curr);
+
+int ice_vsi_init(struct ice_vsi *vsi);
+
+void ice_vsi_clear_rings(struct ice_vsi *vsi);
+
+int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors);
+
int ice_add_mac_to_list(struct ice_vsi *vsi, struct list_head *add_list,
const u8 *macaddr);
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 839c11198e14..73a6574c0d7d 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -33,6 +33,7 @@ static const struct net_device_ops ice_netdev_ops;
static void ice_pf_dis_all_vsi(struct ice_pf *pf);
static void ice_rebuild(struct ice_pf *pf);
static int ice_vsi_release(struct ice_vsi *vsi);
+
static void ice_vsi_release_all(struct ice_pf *pf);
static void ice_update_vsi_stats(struct ice_vsi *vsi);
static void ice_update_pf_stats(struct ice_pf *pf);
@@ -112,35 +113,6 @@ static void ice_check_for_hang_subtask(struct ice_pf *pf)
}
}
-/**
- * ice_get_free_slot - get the next non-NULL location index in array
- * @array: array to search
- * @size: size of the array
- * @curr: last known occupied index to be used as a search hint
- *
- * void * is being used to keep the functionality generic. This lets us use this
- * function on any array of pointers.
- */
-static int ice_get_free_slot(void *array, int size, int curr)
-{
- int **tmp_array = (int **)array;
- int next;
-
- if (curr < (size - 1) && !tmp_array[curr + 1]) {
- next = curr + 1;
- } else {
- int i = 0;
-
- while ((i < size) && (tmp_array[i]))
- i++;
- if (i == size)
- next = ICE_NO_VSI;
- else
- next = i;
- }
- return next;
-}
-
/**
* ice_add_mac_to_sync_list - creates list of mac addresses to be synced
* @netdev: the net device on which the sync is happening
@@ -1210,327 +1182,6 @@ static int ice_vsi_req_irq_msix(struct ice_vsi *vsi, char *basename)
return err;
}
-/**
- * ice_vsi_set_rss_params - Setup RSS capabilities per VSI type
- * @vsi: the VSI being configured
- */
-static void ice_vsi_set_rss_params(struct ice_vsi *vsi)
-{
- struct ice_hw_common_caps *cap;
- struct ice_pf *pf = vsi->back;
-
- if (!test_bit(ICE_FLAG_RSS_ENA, pf->flags)) {
- vsi->rss_size = 1;
- return;
- }
-
- cap = &pf->hw.func_caps.common_cap;
- switch (vsi->type) {
- case ICE_VSI_PF:
- /* PF VSI will inherit RSS instance of PF */
- vsi->rss_table_size = cap->rss_table_size;
- vsi->rss_size = min_t(int, num_online_cpus(),
- BIT(cap->rss_table_entry_width));
- vsi->rss_lut_type = ICE_AQC_GSET_RSS_LUT_TABLE_TYPE_PF;
- break;
- default:
- dev_warn(&pf->pdev->dev, "Unknown VSI type %d\n", vsi->type);
- break;
- }
-}
-
-/**
- * ice_vsi_setup_q_map - Setup a VSI queue map
- * @vsi: the VSI being configured
- * @ctxt: VSI context structure
- */
-static void ice_vsi_setup_q_map(struct ice_vsi *vsi, struct ice_vsi_ctx *ctxt)
-{
- u16 offset = 0, qmap = 0, numq_tc;
- u16 pow = 0, max_rss = 0, qcount;
- u16 qcount_tx = vsi->alloc_txq;
- u16 qcount_rx = vsi->alloc_rxq;
- bool ena_tc0 = false;
- int i;
-
- /* at least TC0 should be enabled by default */
- if (vsi->tc_cfg.numtc) {
- if (!(vsi->tc_cfg.ena_tc & BIT(0)))
- ena_tc0 = true;
- } else {
- ena_tc0 = true;
- }
-
- if (ena_tc0) {
- vsi->tc_cfg.numtc++;
- vsi->tc_cfg.ena_tc |= 1;
- }
-
- numq_tc = qcount_rx / vsi->tc_cfg.numtc;
-
- /* TC mapping is a function of the number of Rx queues assigned to the
- * VSI for each traffic class and the offset of these queues.
- * The first 10 bits are for queue offset for TC0, next 4 bits for no:of
- * queues allocated to TC0. No:of queues is a power-of-2.
- *
- * If TC is not enabled, the queue offset is set to 0, and allocate one
- * queue, this way, traffic for the given TC will be sent to the default
- * queue.
- *
- * Setup number and offset of Rx queues for all TCs for the VSI
- */
-
- /* qcount will change if RSS is enabled */
- if (test_bit(ICE_FLAG_RSS_ENA, vsi->back->flags)) {
- if (vsi->type == ICE_VSI_PF)
- max_rss = ICE_MAX_LG_RSS_QS;
- else
- max_rss = ICE_MAX_SMALL_RSS_QS;
-
- qcount = min_t(int, numq_tc, max_rss);
- qcount = min_t(int, qcount, vsi->rss_size);
- } else {
- qcount = numq_tc;
- }
-
- /* find the (rounded up) power-of-2 of qcount */
- pow = order_base_2(qcount);
-
- for (i = 0; i < ICE_MAX_TRAFFIC_CLASS; i++) {
- if (!(vsi->tc_cfg.ena_tc & BIT(i))) {
- /* TC is not enabled */
- vsi->tc_cfg.tc_info[i].qoffset = 0;
- vsi->tc_cfg.tc_info[i].qcount = 1;
- ctxt->info.tc_mapping[i] = 0;
- continue;
- }
-
- /* TC is enabled */
- vsi->tc_cfg.tc_info[i].qoffset = offset;
- vsi->tc_cfg.tc_info[i].qcount = qcount;
-
- qmap = ((offset << ICE_AQ_VSI_TC_Q_OFFSET_S) &
- ICE_AQ_VSI_TC_Q_OFFSET_M) |
- ((pow << ICE_AQ_VSI_TC_Q_NUM_S) &
- ICE_AQ_VSI_TC_Q_NUM_M);
- offset += qcount;
- ctxt->info.tc_mapping[i] = cpu_to_le16(qmap);
- }
-
- vsi->num_txq = qcount_tx;
- vsi->num_rxq = offset;
-
- /* Rx queue mapping */
- ctxt->info.mapping_flags |= cpu_to_le16(ICE_AQ_VSI_Q_MAP_CONTIG);
- /* q_mapping buffer holds the info for the first queue allocated for
- * this VSI in the PF space and also the number of queues associated
- * with this VSI.
- */
- ctxt->info.q_mapping[0] = cpu_to_le16(vsi->rxq_map[0]);
- ctxt->info.q_mapping[1] = cpu_to_le16(vsi->num_rxq);
-}
-
-/**
- * ice_set_dflt_vsi_ctx - Set default VSI context before adding a VSI
- * @ctxt: the VSI context being set
- *
- * This initializes a default VSI context for all sections except the Queues.
- */
-static void ice_set_dflt_vsi_ctx(struct ice_vsi_ctx *ctxt)
-{
- u32 table = 0;
-
- memset(&ctxt->info, 0, sizeof(ctxt->info));
- /* VSI's should be allocated from shared pool */
- ctxt->alloc_from_pool = true;
- /* Src pruning enabled by default */
- ctxt->info.sw_flags = ICE_AQ_VSI_SW_FLAG_SRC_PRUNE;
- /* Traffic from VSI can be sent to LAN */
- ctxt->info.sw_flags2 = ICE_AQ_VSI_SW_FLAG_LAN_ENA;
-
- /* By default bits 3 and 4 in vlan_flags are 0's which results in legacy
- * behavior (show VLAN, DEI, and UP) in descriptor. Also, allow all
- * packets untagged/tagged.
- */
- ctxt->info.vlan_flags = ((ICE_AQ_VSI_VLAN_MODE_ALL &
- ICE_AQ_VSI_VLAN_MODE_M) >>
- ICE_AQ_VSI_VLAN_MODE_S);
-
- /* Have 1:1 UP mapping for both ingress/egress tables */
- table |= ICE_UP_TABLE_TRANSLATE(0, 0);
- table |= ICE_UP_TABLE_TRANSLATE(1, 1);
- table |= ICE_UP_TABLE_TRANSLATE(2, 2);
- table |= ICE_UP_TABLE_TRANSLATE(3, 3);
- table |= ICE_UP_TABLE_TRANSLATE(4, 4);
- table |= ICE_UP_TABLE_TRANSLATE(5, 5);
- table |= ICE_UP_TABLE_TRANSLATE(6, 6);
- table |= ICE_UP_TABLE_TRANSLATE(7, 7);
- ctxt->info.ingress_table = cpu_to_le32(table);
- ctxt->info.egress_table = cpu_to_le32(table);
- /* Have 1:1 UP mapping for outer to inner UP table */
- ctxt->info.outer_up_table = cpu_to_le32(table);
- /* No Outer tag support outer_tag_flags remains to zero */
-}
-
-/**
- * ice_set_rss_vsi_ctx - Set RSS VSI context before adding a VSI
- * @ctxt: the VSI context being set
- * @vsi: the VSI being configured
- */
-static void ice_set_rss_vsi_ctx(struct ice_vsi_ctx *ctxt, struct ice_vsi *vsi)
-{
- u8 lut_type, hash_type;
-
- switch (vsi->type) {
- case ICE_VSI_PF:
- /* PF VSI will inherit RSS instance of PF */
- lut_type = ICE_AQ_VSI_Q_OPT_RSS_LUT_PF;
- hash_type = ICE_AQ_VSI_Q_OPT_RSS_TPLZ;
- break;
- default:
- dev_warn(&vsi->back->pdev->dev, "Unknown VSI type %d\n",
- vsi->type);
- return;
- }
-
- ctxt->info.q_opt_rss = ((lut_type << ICE_AQ_VSI_Q_OPT_RSS_LUT_S) &
- ICE_AQ_VSI_Q_OPT_RSS_LUT_M) |
- ((hash_type << ICE_AQ_VSI_Q_OPT_RSS_HASH_S) &
- ICE_AQ_VSI_Q_OPT_RSS_HASH_M);
-}
-
-/**
- * ice_vsi_init - Create and initialize a VSI
- * @vsi: the VSI being configured
- *
- * This initializes a VSI context depending on the VSI type to be added and
- * passes it down to the add_vsi aq command to create a new VSI.
- */
-static int ice_vsi_init(struct ice_vsi *vsi)
-{
- struct ice_vsi_ctx ctxt = { 0 };
- struct ice_pf *pf = vsi->back;
- struct ice_hw *hw = &pf->hw;
- int ret = 0;
-
- switch (vsi->type) {
- case ICE_VSI_PF:
- ctxt.flags = ICE_AQ_VSI_TYPE_PF;
- break;
- default:
- return -ENODEV;
- }
-
- ice_set_dflt_vsi_ctx(&ctxt);
- /* if the switch is in VEB mode, allow VSI loopback */
- if (vsi->vsw->bridge_mode == BRIDGE_MODE_VEB)
- ctxt.info.sw_flags |= ICE_AQ_VSI_SW_FLAG_ALLOW_LB;
-
- /* Set LUT type and HASH type if RSS is enabled */
- if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
- ice_set_rss_vsi_ctx(&ctxt, vsi);
-
- ctxt.info.sw_id = vsi->port_info->sw_id;
- ice_vsi_setup_q_map(vsi, &ctxt);
-
- ret = ice_add_vsi(hw, vsi->idx, &ctxt, NULL);
- if (ret) {
- dev_err(&pf->pdev->dev,
- "Add VSI failed, err %d\n", ret);
- return -EIO;
- }
-
- /* keep context for update VSI operations */
- vsi->info = ctxt.info;
-
- /* record VSI number returned */
- vsi->vsi_num = ctxt.vsi_num;
-
- return ret;
-}
-
-/**
- * ice_vsi_clear_rings - Deallocates the Tx and Rx rings for VSI
- * @vsi: the VSI having rings deallocated
- */
-static void ice_vsi_clear_rings(struct ice_vsi *vsi)
-{
- int i;
-
- if (vsi->tx_rings) {
- for (i = 0; i < vsi->alloc_txq; i++) {
- if (vsi->tx_rings[i]) {
- kfree_rcu(vsi->tx_rings[i], rcu);
- vsi->tx_rings[i] = NULL;
- }
- }
- }
- if (vsi->rx_rings) {
- for (i = 0; i < vsi->alloc_rxq; i++) {
- if (vsi->rx_rings[i]) {
- kfree_rcu(vsi->rx_rings[i], rcu);
- vsi->rx_rings[i] = NULL;
- }
- }
- }
-}
-
-/**
- * ice_vsi_alloc_rings - Allocates Tx and Rx rings for the VSI
- * @vsi: VSI which is having rings allocated
- */
-static int ice_vsi_alloc_rings(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- int i;
-
- /* Allocate tx_rings */
- for (i = 0; i < vsi->alloc_txq; i++) {
- struct ice_ring *ring;
-
- /* allocate with kzalloc(), free with kfree_rcu() */
- ring = kzalloc(sizeof(*ring), GFP_KERNEL);
-
- if (!ring)
- goto err_out;
-
- ring->q_index = i;
- ring->reg_idx = vsi->txq_map[i];
- ring->ring_active = false;
- ring->vsi = vsi;
- ring->netdev = vsi->netdev;
- ring->dev = &pf->pdev->dev;
- ring->count = vsi->num_desc;
-
- vsi->tx_rings[i] = ring;
- }
-
- /* Allocate rx_rings */
- for (i = 0; i < vsi->alloc_rxq; i++) {
- struct ice_ring *ring;
-
- /* allocate with kzalloc(), free with kfree_rcu() */
- ring = kzalloc(sizeof(*ring), GFP_KERNEL);
- if (!ring)
- goto err_out;
-
- ring->q_index = i;
- ring->reg_idx = vsi->rxq_map[i];
- ring->ring_active = false;
- ring->vsi = vsi;
- ring->netdev = vsi->netdev;
- ring->dev = &pf->pdev->dev;
- ring->count = vsi->num_desc;
- vsi->rx_rings[i] = ring;
- }
-
- return 0;
-
-err_out:
- ice_vsi_clear_rings(vsi);
- return -ENOMEM;
-}
-
/**
* ice_ena_misc_vector - enable the non-queue interrupts
* @pf: board private structure
@@ -1716,73 +1367,6 @@ static void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
}
}
-/**
- * ice_vsi_set_num_qs - Set num queues, descriptors and vectors for a VSI
- * @vsi: the VSI being configured
- *
- * Return 0 on success and a negative value on error
- */
-static void ice_vsi_set_num_qs(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
-
- switch (vsi->type) {
- case ICE_VSI_PF:
- vsi->alloc_txq = pf->num_lan_tx;
- vsi->alloc_rxq = pf->num_lan_rx;
- vsi->num_desc = ALIGN(ICE_DFLT_NUM_DESC, ICE_REQ_DESC_MULTIPLE);
- vsi->num_q_vectors = max_t(int, pf->num_lan_rx, pf->num_lan_tx);
- break;
- default:
- dev_warn(&vsi->back->pdev->dev, "Unknown VSI type %d\n",
- vsi->type);
- break;
- }
-}
-
-/**
- * ice_vsi_alloc_arrays - Allocate queue and vector pointer arrays for the vsi
- * @vsi: VSI pointer
- * @alloc_qvectors: a bool to specify if q_vectors need to be allocated.
- *
- * On error: returns error code (negative)
- * On success: returns 0
- */
-static int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
-{
- struct ice_pf *pf = vsi->back;
-
- /* allocate memory for both Tx and Rx ring pointers */
- vsi->tx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_txq,
- sizeof(struct ice_ring *), GFP_KERNEL);
- if (!vsi->tx_rings)
- goto err_txrings;
-
- vsi->rx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_rxq,
- sizeof(struct ice_ring *), GFP_KERNEL);
- if (!vsi->rx_rings)
- goto err_rxrings;
-
- if (alloc_qvectors) {
- /* allocate memory for q_vector pointers */
- vsi->q_vectors = devm_kcalloc(&pf->pdev->dev,
- vsi->num_q_vectors,
- sizeof(struct ice_q_vector *),
- GFP_KERNEL);
- if (!vsi->q_vectors)
- goto err_vectors;
- }
-
- return 0;
-
-err_vectors:
- devm_kfree(&pf->pdev->dev, vsi->rx_rings);
-err_rxrings:
- devm_kfree(&pf->pdev->dev, vsi->tx_rings);
-err_txrings:
- return -ENOMEM;
-}
-
/**
* ice_vsi_alloc - Allocates the next available struct vsi in the PF
* @pf: board private structure
--
2.17.1
^ permalink raw reply related
* [net-next 3/8] ice: Move common functions out of ice_main.c part 3/7
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem
Cc: Anirudh Venkataramanan, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <20181001211431.14325-1-jeffrey.t.kirsher@intel.com>
From: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
This patch continues the code move out of ice_main.c
The following top level functions (and related dependency functions) were
moved to ice_lib.c:
ice_vsi_delete
ice_free_res
ice_get_res
ice_is_reset_recovery_pending
ice_vsi_put_qs
ice_vsi_dis_irq
ice_vsi_free_irq
ice_vsi_free_rx_rings
ice_vsi_free_tx_rings
ice_msix_clean_rings
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_lib.c | 388 ++++++++++++++++++++++
drivers/net/ethernet/intel/ice/ice_lib.h | 22 ++
drivers/net/ethernet/intel/ice/ice_main.c | 386 ---------------------
3 files changed, 410 insertions(+), 386 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 06a54d79fba8..474ce5828bd4 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -225,6 +225,67 @@ static int ice_vsi_ctrl_rx_rings(struct ice_vsi *vsi, bool ena)
return ret;
}
+/**
+ * ice_vsi_delete - delete a VSI from the switch
+ * @vsi: pointer to VSI being removed
+ */
+void ice_vsi_delete(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ struct ice_vsi_ctx ctxt;
+ enum ice_status status;
+
+ ctxt.vsi_num = vsi->vsi_num;
+
+ memcpy(&ctxt.info, &vsi->info, sizeof(struct ice_aqc_vsi_props));
+
+ status = ice_free_vsi(&pf->hw, vsi->idx, &ctxt, false, NULL);
+ if (status)
+ dev_err(&pf->pdev->dev, "Failed to delete VSI %i in FW\n",
+ vsi->vsi_num);
+}
+
+/**
+ * ice_msix_clean_rings - MSIX mode Interrupt Handler
+ * @irq: interrupt number
+ * @data: pointer to a q_vector
+ */
+irqreturn_t ice_msix_clean_rings(int __always_unused irq, void *data)
+{
+ struct ice_q_vector *q_vector = (struct ice_q_vector *)data;
+
+ if (!q_vector->tx.ring && !q_vector->rx.ring)
+ return IRQ_HANDLED;
+
+ napi_schedule(&q_vector->napi);
+
+ return IRQ_HANDLED;
+}
+
+/**
+ * ice_vsi_put_qs - Release queues from VSI to PF
+ * @vsi: the VSI that is going to release queues
+ */
+void ice_vsi_put_qs(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ int i;
+
+ mutex_lock(&pf->avail_q_mutex);
+
+ for (i = 0; i < vsi->alloc_txq; i++) {
+ clear_bit(vsi->txq_map[i], pf->avail_txqs);
+ vsi->txq_map[i] = ICE_INVAL_Q_INDEX;
+ }
+
+ for (i = 0; i < vsi->alloc_rxq; i++) {
+ clear_bit(vsi->rxq_map[i], pf->avail_rxqs);
+ vsi->rxq_map[i] = ICE_INVAL_Q_INDEX;
+ }
+
+ mutex_unlock(&pf->avail_q_mutex);
+}
+
/**
* ice_add_mac_to_list - Add a mac address filter entry to the list
* @vsi: the VSI to be forwarded to
@@ -747,3 +808,330 @@ int ice_vsi_stop_tx_rings(struct ice_vsi *vsi)
return err;
}
+
+/**
+ * ice_cfg_vlan_pruning - enable or disable VLAN pruning on the VSI
+ * @vsi: VSI to enable or disable VLAN pruning on
+ * @ena: set to true to enable VLAN pruning and false to disable it
+ *
+ * returns 0 if VSI is updated, negative otherwise
+ */
+int ice_cfg_vlan_pruning(struct ice_vsi *vsi, bool ena)
+{
+ struct ice_vsi_ctx *ctxt;
+ struct device *dev;
+ int status;
+
+ if (!vsi)
+ return -EINVAL;
+
+ dev = &vsi->back->pdev->dev;
+ ctxt = devm_kzalloc(dev, sizeof(*ctxt), GFP_KERNEL);
+ if (!ctxt)
+ return -ENOMEM;
+
+ ctxt->info = vsi->info;
+
+ if (ena) {
+ ctxt->info.sec_flags |=
+ ICE_AQ_VSI_SEC_TX_VLAN_PRUNE_ENA <<
+ ICE_AQ_VSI_SEC_TX_PRUNE_ENA_S;
+ ctxt->info.sw_flags2 |= ICE_AQ_VSI_SW_FLAG_RX_VLAN_PRUNE_ENA;
+ } else {
+ ctxt->info.sec_flags &=
+ ~(ICE_AQ_VSI_SEC_TX_VLAN_PRUNE_ENA <<
+ ICE_AQ_VSI_SEC_TX_PRUNE_ENA_S);
+ ctxt->info.sw_flags2 &= ~ICE_AQ_VSI_SW_FLAG_RX_VLAN_PRUNE_ENA;
+ }
+
+ ctxt->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_SECURITY_VALID |
+ ICE_AQ_VSI_PROP_SW_VALID);
+ ctxt->vsi_num = vsi->vsi_num;
+ status = ice_aq_update_vsi(&vsi->back->hw, ctxt, NULL);
+ if (status) {
+ netdev_err(vsi->netdev, "%sabling VLAN pruning on VSI %d failed, err = %d, aq_err = %d\n",
+ ena ? "Ena" : "Dis", vsi->vsi_num, status,
+ vsi->back->hw.adminq.sq_last_status);
+ goto err_out;
+ }
+
+ vsi->info.sec_flags = ctxt->info.sec_flags;
+ vsi->info.sw_flags2 = ctxt->info.sw_flags2;
+
+ devm_kfree(dev, ctxt);
+ return 0;
+
+err_out:
+ devm_kfree(dev, ctxt);
+ return -EIO;
+}
+
+/**
+ * ice_vsi_release_msix - Clear the queue to Interrupt mapping in HW
+ * @vsi: the VSI being cleaned up
+ */
+static void ice_vsi_release_msix(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ u16 vector = vsi->base_vector;
+ struct ice_hw *hw = &pf->hw;
+ u32 txq = 0;
+ u32 rxq = 0;
+ int i, q;
+
+ for (i = 0; i < vsi->num_q_vectors; i++, vector++) {
+ struct ice_q_vector *q_vector = vsi->q_vectors[i];
+
+ wr32(hw, GLINT_ITR(ICE_RX_ITR, vector), 0);
+ wr32(hw, GLINT_ITR(ICE_TX_ITR, vector), 0);
+ for (q = 0; q < q_vector->num_ring_tx; q++) {
+ wr32(hw, QINT_TQCTL(vsi->txq_map[txq]), 0);
+ txq++;
+ }
+
+ for (q = 0; q < q_vector->num_ring_rx; q++) {
+ wr32(hw, QINT_RQCTL(vsi->rxq_map[rxq]), 0);
+ rxq++;
+ }
+ }
+
+ ice_flush(hw);
+}
+
+/**
+ * ice_vsi_free_irq - Free the IRQ association with the OS
+ * @vsi: the VSI being configured
+ */
+void ice_vsi_free_irq(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ int base = vsi->base_vector;
+
+ if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags)) {
+ int i;
+
+ if (!vsi->q_vectors || !vsi->irqs_ready)
+ return;
+
+ vsi->irqs_ready = false;
+ for (i = 0; i < vsi->num_q_vectors; i++) {
+ u16 vector = i + base;
+ int irq_num;
+
+ irq_num = pf->msix_entries[vector].vector;
+
+ /* free only the irqs that were actually requested */
+ if (!vsi->q_vectors[i] ||
+ !(vsi->q_vectors[i]->num_ring_tx ||
+ vsi->q_vectors[i]->num_ring_rx))
+ continue;
+
+ /* clear the affinity notifier in the IRQ descriptor */
+ irq_set_affinity_notifier(irq_num, NULL);
+
+ /* clear the affinity_mask in the IRQ descriptor */
+ irq_set_affinity_hint(irq_num, NULL);
+ synchronize_irq(irq_num);
+ devm_free_irq(&pf->pdev->dev, irq_num,
+ vsi->q_vectors[i]);
+ }
+ ice_vsi_release_msix(vsi);
+ }
+}
+
+/**
+ * ice_vsi_free_tx_rings - Free Tx resources for VSI queues
+ * @vsi: the VSI having resources freed
+ */
+void ice_vsi_free_tx_rings(struct ice_vsi *vsi)
+{
+ int i;
+
+ if (!vsi->tx_rings)
+ return;
+
+ ice_for_each_txq(vsi, i)
+ if (vsi->tx_rings[i] && vsi->tx_rings[i]->desc)
+ ice_free_tx_ring(vsi->tx_rings[i]);
+}
+
+/**
+ * ice_vsi_free_rx_rings - Free Rx resources for VSI queues
+ * @vsi: the VSI having resources freed
+ */
+void ice_vsi_free_rx_rings(struct ice_vsi *vsi)
+{
+ int i;
+
+ if (!vsi->rx_rings)
+ return;
+
+ ice_for_each_rxq(vsi, i)
+ if (vsi->rx_rings[i] && vsi->rx_rings[i]->desc)
+ ice_free_rx_ring(vsi->rx_rings[i]);
+}
+
+/**
+ * ice_free_res - free a block of resources
+ * @res: pointer to the resource
+ * @index: starting index previously returned by ice_get_res
+ * @id: identifier to track owner
+ *
+ * Returns number of resources freed
+ */
+int ice_free_res(struct ice_res_tracker *res, u16 index, u16 id)
+{
+ int count = 0;
+ int i;
+
+ if (!res || index >= res->num_entries)
+ return -EINVAL;
+
+ id |= ICE_RES_VALID_BIT;
+ for (i = index; i < res->num_entries && res->list[i] == id; i++) {
+ res->list[i] = 0;
+ count++;
+ }
+
+ return count;
+}
+
+/**
+ * ice_search_res - Search the tracker for a block of resources
+ * @res: pointer to the resource
+ * @needed: size of the block needed
+ * @id: identifier to track owner
+ *
+ * Returns the base item index of the block, or -ENOMEM for error
+ */
+static int ice_search_res(struct ice_res_tracker *res, u16 needed, u16 id)
+{
+ int start = res->search_hint;
+ int end = start;
+
+ id |= ICE_RES_VALID_BIT;
+
+ do {
+ /* skip already allocated entries */
+ if (res->list[end++] & ICE_RES_VALID_BIT) {
+ start = end;
+ if ((start + needed) > res->num_entries)
+ break;
+ }
+
+ if (end == (start + needed)) {
+ int i = start;
+
+ /* there was enough, so assign it to the requestor */
+ while (i != end)
+ res->list[i++] = id;
+
+ if (end == res->num_entries)
+ end = 0;
+
+ res->search_hint = end;
+ return start;
+ }
+ } while (1);
+
+ return -ENOMEM;
+}
+
+/**
+ * ice_get_res - get a block of resources
+ * @pf: board private structure
+ * @res: pointer to the resource
+ * @needed: size of the block needed
+ * @id: identifier to track owner
+ *
+ * Returns the base item index of the block, or -ENOMEM for error
+ * The search_hint trick and lack of advanced fit-finding only works
+ * because we're highly likely to have all the same sized requests.
+ * Linear search time and any fragmentation should be minimal.
+ */
+int
+ice_get_res(struct ice_pf *pf, struct ice_res_tracker *res, u16 needed, u16 id)
+{
+ int ret;
+
+ if (!res || !pf)
+ return -EINVAL;
+
+ if (!needed || needed > res->num_entries || id >= ICE_RES_VALID_BIT) {
+ dev_err(&pf->pdev->dev,
+ "param err: needed=%d, num_entries = %d id=0x%04x\n",
+ needed, res->num_entries, id);
+ return -EINVAL;
+ }
+
+ /* search based on search_hint */
+ ret = ice_search_res(res, needed, id);
+
+ if (ret < 0) {
+ /* previous search failed. Reset search hint and try again */
+ res->search_hint = 0;
+ ret = ice_search_res(res, needed, id);
+ }
+
+ return ret;
+}
+
+/**
+ * ice_vsi_dis_irq - Mask off queue interrupt generation on the VSI
+ * @vsi: the VSI being un-configured
+ */
+void ice_vsi_dis_irq(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = vsi->back;
+ struct ice_hw *hw = &pf->hw;
+ int base = vsi->base_vector;
+ u32 val;
+ int i;
+
+ /* disable interrupt causation from each queue */
+ if (vsi->tx_rings) {
+ ice_for_each_txq(vsi, i) {
+ if (vsi->tx_rings[i]) {
+ u16 reg;
+
+ reg = vsi->tx_rings[i]->reg_idx;
+ val = rd32(hw, QINT_TQCTL(reg));
+ val &= ~QINT_TQCTL_CAUSE_ENA_M;
+ wr32(hw, QINT_TQCTL(reg), val);
+ }
+ }
+ }
+
+ if (vsi->rx_rings) {
+ ice_for_each_rxq(vsi, i) {
+ if (vsi->rx_rings[i]) {
+ u16 reg;
+
+ reg = vsi->rx_rings[i]->reg_idx;
+ val = rd32(hw, QINT_RQCTL(reg));
+ val &= ~QINT_RQCTL_CAUSE_ENA_M;
+ wr32(hw, QINT_RQCTL(reg), val);
+ }
+ }
+ }
+
+ /* disable each interrupt */
+ if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags)) {
+ for (i = vsi->base_vector;
+ i < (vsi->num_q_vectors + vsi->base_vector); i++)
+ wr32(hw, GLINT_DYN_CTL(i), 0);
+
+ ice_flush(hw);
+ for (i = 0; i < vsi->num_q_vectors; i++)
+ synchronize_irq(pf->msix_entries[i + base].vector);
+ }
+}
+
+/**
+ * ice_is_reset_recovery_pending - schedule a reset
+ * @state: pf state field
+ */
+bool ice_is_reset_recovery_pending(unsigned long *state)
+{
+ return test_bit(__ICE_RESET_RECOVERY_PENDING, state);
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
index ad4257929b9b..cbde99f38798 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -33,4 +33,26 @@ int ice_vsi_stop_rx_rings(struct ice_vsi *vsi);
int ice_vsi_stop_tx_rings(struct ice_vsi *vsi);
+int ice_cfg_vlan_pruning(struct ice_vsi *vsi, bool ena);
+
+void ice_vsi_delete(struct ice_vsi *vsi);
+
+int ice_free_res(struct ice_res_tracker *res, u16 index, u16 id);
+
+int
+ice_get_res(struct ice_pf *pf, struct ice_res_tracker *res, u16 needed, u16 id);
+
+bool ice_is_reset_recovery_pending(unsigned long *state);
+
+void ice_vsi_put_qs(struct ice_vsi *vsi);
+
+void ice_vsi_dis_irq(struct ice_vsi *vsi);
+
+void ice_vsi_free_irq(struct ice_vsi *vsi);
+
+void ice_vsi_free_rx_rings(struct ice_vsi *vsi);
+
+void ice_vsi_free_tx_rings(struct ice_vsi *vsi);
+
+irqreturn_t ice_msix_clean_rings(int __always_unused irq, void *data);
#endif /* !_ICE_LIB_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index ececf3dabf7e..839c11198e14 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -141,109 +141,6 @@ static int ice_get_free_slot(void *array, int size, int curr)
return next;
}
-/**
- * ice_search_res - Search the tracker for a block of resources
- * @res: pointer to the resource
- * @needed: size of the block needed
- * @id: identifier to track owner
- * Returns the base item index of the block, or -ENOMEM for error
- */
-static int ice_search_res(struct ice_res_tracker *res, u16 needed, u16 id)
-{
- int start = res->search_hint;
- int end = start;
-
- id |= ICE_RES_VALID_BIT;
-
- do {
- /* skip already allocated entries */
- if (res->list[end++] & ICE_RES_VALID_BIT) {
- start = end;
- if ((start + needed) > res->num_entries)
- break;
- }
-
- if (end == (start + needed)) {
- int i = start;
-
- /* there was enough, so assign it to the requestor */
- while (i != end)
- res->list[i++] = id;
-
- if (end == res->num_entries)
- end = 0;
-
- res->search_hint = end;
- return start;
- }
- } while (1);
-
- return -ENOMEM;
-}
-
-/**
- * ice_get_res - get a block of resources
- * @pf: board private structure
- * @res: pointer to the resource
- * @needed: size of the block needed
- * @id: identifier to track owner
- *
- * Returns the base item index of the block, or -ENOMEM for error
- * The search_hint trick and lack of advanced fit-finding only works
- * because we're highly likely to have all the same sized requests.
- * Linear search time and any fragmentation should be minimal.
- */
-static int
-ice_get_res(struct ice_pf *pf, struct ice_res_tracker *res, u16 needed, u16 id)
-{
- int ret;
-
- if (!res || !pf)
- return -EINVAL;
-
- if (!needed || needed > res->num_entries || id >= ICE_RES_VALID_BIT) {
- dev_err(&pf->pdev->dev,
- "param err: needed=%d, num_entries = %d id=0x%04x\n",
- needed, res->num_entries, id);
- return -EINVAL;
- }
-
- /* search based on search_hint */
- ret = ice_search_res(res, needed, id);
-
- if (ret < 0) {
- /* previous search failed. Reset search hint and try again */
- res->search_hint = 0;
- ret = ice_search_res(res, needed, id);
- }
-
- return ret;
-}
-
-/**
- * ice_free_res - free a block of resources
- * @res: pointer to the resource
- * @index: starting index previously returned by ice_get_res
- * @id: identifier to track owner
- * Returns number of resources freed
- */
-static int ice_free_res(struct ice_res_tracker *res, u16 index, u16 id)
-{
- int count = 0;
- int i;
-
- if (!res || index >= res->num_entries)
- return -EINVAL;
-
- id |= ICE_RES_VALID_BIT;
- for (i = index; i < res->num_entries && res->list[i] == id; i++) {
- res->list[i] = 0;
- count++;
- }
-
- return count;
-}
-
/**
* ice_add_mac_to_sync_list - creates list of mac addresses to be synced
* @netdev: the net device on which the sync is happening
@@ -299,63 +196,6 @@ static bool ice_vsi_fltr_changed(struct ice_vsi *vsi)
test_bit(ICE_VSI_FLAG_VLAN_FLTR_CHANGED, vsi->flags);
}
-/**
- * ice_cfg_vlan_pruning - enable or disable VLAN pruning on the VSI
- * @vsi: VSI to enable or disable VLAN pruning on
- * @ena: set to true to enable VLAN pruning and false to disable it
- *
- * returns 0 if VSI is updated, negative otherwise
- */
-static int ice_cfg_vlan_pruning(struct ice_vsi *vsi, bool ena)
-{
- struct ice_vsi_ctx *ctxt;
- struct device *dev;
- int status;
-
- if (!vsi)
- return -EINVAL;
-
- dev = &vsi->back->pdev->dev;
- ctxt = devm_kzalloc(dev, sizeof(*ctxt), GFP_KERNEL);
- if (!ctxt)
- return -ENOMEM;
-
- ctxt->info = vsi->info;
-
- if (ena) {
- ctxt->info.sec_flags |=
- ICE_AQ_VSI_SEC_TX_VLAN_PRUNE_ENA <<
- ICE_AQ_VSI_SEC_TX_PRUNE_ENA_S;
- ctxt->info.sw_flags2 |= ICE_AQ_VSI_SW_FLAG_RX_VLAN_PRUNE_ENA;
- } else {
- ctxt->info.sec_flags &=
- ~(ICE_AQ_VSI_SEC_TX_VLAN_PRUNE_ENA <<
- ICE_AQ_VSI_SEC_TX_PRUNE_ENA_S);
- ctxt->info.sw_flags2 &= ~ICE_AQ_VSI_SW_FLAG_RX_VLAN_PRUNE_ENA;
- }
-
- ctxt->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_SECURITY_VALID |
- ICE_AQ_VSI_PROP_SW_VALID);
- ctxt->vsi_num = vsi->vsi_num;
- status = ice_aq_update_vsi(&vsi->back->hw, ctxt, NULL);
- if (status) {
- netdev_err(vsi->netdev, "%sabling VLAN pruning on VSI %d failed, err = %d, aq_err = %d\n",
- ena ? "Ena" : "Dis", vsi->vsi_num, status,
- vsi->back->hw.adminq.sq_last_status);
- goto err_out;
- }
-
- vsi->info.sec_flags = ctxt->info.sec_flags;
- vsi->info.sw_flags2 = ctxt->info.sw_flags2;
-
- devm_kfree(dev, ctxt);
- return 0;
-
-err_out:
- devm_kfree(dev, ctxt);
- return -EIO;
-}
-
/**
* ice_vsi_sync_fltr - Update the VSI filter list to the HW
* @vsi: ptr to the VSI
@@ -520,15 +360,6 @@ static void ice_sync_fltr_subtask(struct ice_pf *pf)
}
}
-/**
- * ice_is_reset_recovery_pending - schedule a reset
- * @state: pf state field
- */
-static bool ice_is_reset_recovery_pending(unsigned long int *state)
-{
- return test_bit(__ICE_RESET_RECOVERY_PENDING, state);
-}
-
/**
* ice_prepare_for_reset - prep for the core to reset
* @pf: board private structure
@@ -1292,57 +1123,6 @@ static void ice_irq_affinity_notify(struct irq_affinity_notify *notify,
*/
static void ice_irq_affinity_release(struct kref __always_unused *ref) {}
-/**
- * ice_vsi_dis_irq - Mask off queue interrupt generation on the VSI
- * @vsi: the VSI being un-configured
- */
-static void ice_vsi_dis_irq(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- struct ice_hw *hw = &pf->hw;
- int base = vsi->base_vector;
- u32 val;
- int i;
-
- /* disable interrupt causation from each queue */
- if (vsi->tx_rings) {
- ice_for_each_txq(vsi, i) {
- if (vsi->tx_rings[i]) {
- u16 reg;
-
- reg = vsi->tx_rings[i]->reg_idx;
- val = rd32(hw, QINT_TQCTL(reg));
- val &= ~QINT_TQCTL_CAUSE_ENA_M;
- wr32(hw, QINT_TQCTL(reg), val);
- }
- }
- }
-
- if (vsi->rx_rings) {
- ice_for_each_rxq(vsi, i) {
- if (vsi->rx_rings[i]) {
- u16 reg;
-
- reg = vsi->rx_rings[i]->reg_idx;
- val = rd32(hw, QINT_RQCTL(reg));
- val &= ~QINT_RQCTL_CAUSE_ENA_M;
- wr32(hw, QINT_RQCTL(reg), val);
- }
- }
- }
-
- /* disable each interrupt */
- if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags)) {
- for (i = vsi->base_vector;
- i < (vsi->num_q_vectors + vsi->base_vector); i++)
- wr32(hw, GLINT_DYN_CTL(i), 0);
-
- ice_flush(hw);
- for (i = 0; i < vsi->num_q_vectors; i++)
- synchronize_irq(pf->msix_entries[i + base].vector);
- }
-}
-
/**
* ice_vsi_ena_irq - Enable IRQ for the given VSI
* @vsi: the VSI being configured
@@ -1363,26 +1143,6 @@ static int ice_vsi_ena_irq(struct ice_vsi *vsi)
return 0;
}
-/**
- * ice_vsi_delete - delete a VSI from the switch
- * @vsi: pointer to VSI being removed
- */
-static void ice_vsi_delete(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- struct ice_vsi_ctx ctxt;
- enum ice_status status;
-
- ctxt.vsi_num = vsi->vsi_num;
-
- memcpy(&ctxt.info, &vsi->info, sizeof(struct ice_aqc_vsi_props));
-
- status = ice_free_vsi(&pf->hw, vsi->idx, &ctxt, false, NULL);
- if (status)
- dev_err(&pf->pdev->dev, "Failed to delete VSI %i in FW\n",
- vsi->vsi_num);
-}
-
/**
* ice_vsi_req_irq_msix - get MSI-X vectors from the OS for the VSI
* @vsi: the VSI being configured
@@ -1689,38 +1449,6 @@ static int ice_vsi_init(struct ice_vsi *vsi)
return ret;
}
-/**
- * ice_vsi_release_msix - Clear the queue to Interrupt mapping in HW
- * @vsi: the VSI being cleaned up
- */
-static void ice_vsi_release_msix(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- u16 vector = vsi->base_vector;
- struct ice_hw *hw = &pf->hw;
- u32 txq = 0;
- u32 rxq = 0;
- int i, q;
-
- for (i = 0; i < vsi->num_q_vectors; i++, vector++) {
- struct ice_q_vector *q_vector = vsi->q_vectors[i];
-
- wr32(hw, GLINT_ITR(ICE_RX_ITR, vector), 0);
- wr32(hw, GLINT_ITR(ICE_TX_ITR, vector), 0);
- for (q = 0; q < q_vector->num_ring_tx; q++) {
- wr32(hw, QINT_TQCTL(vsi->txq_map[txq]), 0);
- txq++;
- }
-
- for (q = 0; q < q_vector->num_ring_rx; q++) {
- wr32(hw, QINT_RQCTL(vsi->rxq_map[rxq]), 0);
- rxq++;
- }
- }
-
- ice_flush(hw);
-}
-
/**
* ice_vsi_clear_rings - Deallocates the Tx and Rx rings for VSI
* @vsi: the VSI having rings deallocated
@@ -1803,47 +1531,6 @@ static int ice_vsi_alloc_rings(struct ice_vsi *vsi)
return -ENOMEM;
}
-/**
- * ice_vsi_free_irq - Free the irq association with the OS
- * @vsi: the VSI being configured
- */
-static void ice_vsi_free_irq(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- int base = vsi->base_vector;
-
- if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags)) {
- int i;
-
- if (!vsi->q_vectors || !vsi->irqs_ready)
- return;
-
- vsi->irqs_ready = false;
- for (i = 0; i < vsi->num_q_vectors; i++) {
- u16 vector = i + base;
- int irq_num;
-
- irq_num = pf->msix_entries[vector].vector;
-
- /* free only the irqs that were actually requested */
- if (!vsi->q_vectors[i] ||
- !(vsi->q_vectors[i]->num_ring_tx ||
- vsi->q_vectors[i]->num_ring_rx))
- continue;
-
- /* clear the affinity notifier in the IRQ descriptor */
- irq_set_affinity_notifier(irq_num, NULL);
-
- /* clear the affinity_mask in the IRQ descriptor */
- irq_set_affinity_hint(irq_num, NULL);
- synchronize_irq(irq_num);
- devm_free_irq(&pf->pdev->dev, irq_num,
- vsi->q_vectors[i]);
- }
- ice_vsi_release_msix(vsi);
- }
-}
-
/**
* ice_ena_misc_vector - enable the non-queue interrupts
* @pf: board private structure
@@ -2096,23 +1783,6 @@ static int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
return -ENOMEM;
}
-/**
- * ice_msix_clean_rings - MSIX mode Interrupt Handler
- * @irq: interrupt number
- * @data: pointer to a q_vector
- */
-static irqreturn_t ice_msix_clean_rings(int __always_unused irq, void *data)
-{
- struct ice_q_vector *q_vector = (struct ice_q_vector *)data;
-
- if (!q_vector->tx.ring && !q_vector->rx.ring)
- return IRQ_HANDLED;
-
- napi_schedule(&q_vector->napi);
-
- return IRQ_HANDLED;
-}
-
/**
* ice_vsi_alloc - Allocates the next available struct vsi in the PF
* @pf: board private structure
@@ -2399,30 +2069,6 @@ static int ice_vsi_get_qs(struct ice_vsi *vsi)
return ret;
}
-/**
- * ice_vsi_put_qs - Release queues from VSI to PF
- * @vsi: the VSI thats going to release queues
- */
-static void ice_vsi_put_qs(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = vsi->back;
- int i;
-
- mutex_lock(&pf->avail_q_mutex);
-
- for (i = 0; i < vsi->alloc_txq; i++) {
- clear_bit(vsi->txq_map[i], pf->avail_txqs);
- vsi->txq_map[i] = ICE_INVAL_Q_INDEX;
- }
-
- for (i = 0; i < vsi->alloc_rxq; i++) {
- clear_bit(vsi->rxq_map[i], pf->avail_rxqs);
- vsi->rxq_map[i] = ICE_INVAL_Q_INDEX;
- }
-
- mutex_unlock(&pf->avail_q_mutex);
-}
-
/**
* ice_free_q_vector - Free memory allocated for a specific interrupt vector
* @vsi: VSI having the memory freed
@@ -4419,38 +4065,6 @@ static int ice_vsi_req_irq(struct ice_vsi *vsi, char *basename)
return err;
}
-/**
- * ice_vsi_free_tx_rings - Free Tx resources for VSI queues
- * @vsi: the VSI having resources freed
- */
-static void ice_vsi_free_tx_rings(struct ice_vsi *vsi)
-{
- int i;
-
- if (!vsi->tx_rings)
- return;
-
- ice_for_each_txq(vsi, i)
- if (vsi->tx_rings[i] && vsi->tx_rings[i]->desc)
- ice_free_tx_ring(vsi->tx_rings[i]);
-}
-
-/**
- * ice_vsi_free_rx_rings - Free Rx resources for VSI queues
- * @vsi: the VSI having resources freed
- */
-static void ice_vsi_free_rx_rings(struct ice_vsi *vsi)
-{
- int i;
-
- if (!vsi->rx_rings)
- return;
-
- ice_for_each_rxq(vsi, i)
- if (vsi->rx_rings[i] && vsi->rx_rings[i]->desc)
- ice_free_rx_ring(vsi->rx_rings[i]);
-}
-
/**
* ice_vsi_open - Called when a network interface is made active
* @vsi: the VSI to open
--
2.17.1
^ permalink raw reply related
* [net-next 7/8] ice: Move common functions out of ice_main.c part 7/7
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem
Cc: Anirudh Venkataramanan, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <20181001211431.14325-1-jeffrey.t.kirsher@intel.com>
From: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
This patch completes the code move out of ice_main.c
The following top level functions and related dependency functions) were
moved to ice_lib.c:
ice_vsi_setup
ice_vsi_cfg_tc
The following functions were made static again:
ice_vsi_setup_vector_base
ice_vsi_alloc_q_vectors
ice_vsi_get_qs
void ice_vsi_map_rings_to_vectors
ice_vsi_alloc_rings
ice_vsi_set_rss_params
ice_vsi_set_num_qs
ice_get_free_slot
ice_vsi_init
ice_vsi_alloc_arrays
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_lib.c | 253 +++++++++++++++++++++-
drivers/net/ethernet/intel/ice/ice_lib.h | 26 +--
drivers/net/ethernet/intel/ice/ice_main.c | 252 ---------------------
3 files changed, 249 insertions(+), 282 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 232ca06974ea..21e3a3e70329 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -233,7 +233,7 @@ static int ice_vsi_ctrl_rx_rings(struct ice_vsi *vsi, bool ena)
* On error: returns error code (negative)
* On success: returns 0
*/
-int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
+static int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
{
struct ice_pf *pf = vsi->back;
@@ -274,7 +274,7 @@ int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
*
* Return 0 on success and a negative value on error
*/
-void ice_vsi_set_num_qs(struct ice_vsi *vsi)
+static void ice_vsi_set_num_qs(struct ice_vsi *vsi)
{
struct ice_pf *pf = vsi->back;
@@ -301,7 +301,7 @@ void ice_vsi_set_num_qs(struct ice_vsi *vsi)
* void * is being used to keep the functionality generic. This lets us use this
* function on any array of pointers.
*/
-int ice_get_free_slot(void *array, int size, int curr)
+static int ice_get_free_slot(void *array, int size, int curr)
{
int **tmp_array = (int **)array;
int next;
@@ -423,6 +423,70 @@ irqreturn_t ice_msix_clean_rings(int __always_unused irq, void *data)
return IRQ_HANDLED;
}
+/**
+ * ice_vsi_alloc - Allocates the next available struct VSI in the PF
+ * @pf: board private structure
+ * @type: type of VSI
+ *
+ * returns a pointer to a VSI on success, NULL on failure.
+ */
+static struct ice_vsi *ice_vsi_alloc(struct ice_pf *pf, enum ice_vsi_type type)
+{
+ struct ice_vsi *vsi = NULL;
+
+ /* Need to protect the allocation of the VSIs at the PF level */
+ mutex_lock(&pf->sw_mutex);
+
+ /* If we have already allocated our maximum number of VSIs,
+ * pf->next_vsi will be ICE_NO_VSI. If not, pf->next_vsi index
+ * is available to be populated
+ */
+ if (pf->next_vsi == ICE_NO_VSI) {
+ dev_dbg(&pf->pdev->dev, "out of VSI slots!\n");
+ goto unlock_pf;
+ }
+
+ vsi = devm_kzalloc(&pf->pdev->dev, sizeof(*vsi), GFP_KERNEL);
+ if (!vsi)
+ goto unlock_pf;
+
+ vsi->type = type;
+ vsi->back = pf;
+ set_bit(__ICE_DOWN, vsi->state);
+ vsi->idx = pf->next_vsi;
+ vsi->work_lmt = ICE_DFLT_IRQ_WORK;
+
+ ice_vsi_set_num_qs(vsi);
+
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ if (ice_vsi_alloc_arrays(vsi, true))
+ goto err_rings;
+
+ /* Setup default MSIX irq handler for VSI */
+ vsi->irq_handler = ice_msix_clean_rings;
+ break;
+ default:
+ dev_warn(&pf->pdev->dev, "Unknown VSI type %d\n", vsi->type);
+ goto unlock_pf;
+ }
+
+ /* fill VSI slot in the PF struct */
+ pf->vsi[pf->next_vsi] = vsi;
+
+ /* prepare pf->next_vsi for next use */
+ pf->next_vsi = ice_get_free_slot(pf->vsi, pf->num_alloc_vsi,
+ pf->next_vsi);
+ goto unlock_pf;
+
+err_rings:
+ devm_kfree(&pf->pdev->dev, vsi);
+ vsi = NULL;
+unlock_pf:
+ mutex_unlock(&pf->sw_mutex);
+ return vsi;
+}
+
/**
* ice_vsi_get_qs_contig - Assign a contiguous chunk of queues to VSI
* @vsi: the VSI getting queues
@@ -533,7 +597,7 @@ static int ice_vsi_get_qs_scatter(struct ice_vsi *vsi)
*
* Returns 0 on success and a negative value on error
*/
-int ice_vsi_get_qs(struct ice_vsi *vsi)
+static int ice_vsi_get_qs(struct ice_vsi *vsi)
{
int ret = 0;
@@ -602,7 +666,7 @@ static void ice_rss_clean(struct ice_vsi *vsi)
* ice_vsi_set_rss_params - Setup RSS capabilities per VSI type
* @vsi: the VSI being configured
*/
-void ice_vsi_set_rss_params(struct ice_vsi *vsi)
+static void ice_vsi_set_rss_params(struct ice_vsi *vsi)
{
struct ice_hw_common_caps *cap;
struct ice_pf *pf = vsi->back;
@@ -793,7 +857,7 @@ static void ice_set_rss_vsi_ctx(struct ice_vsi_ctx *ctxt, struct ice_vsi *vsi)
* This initializes a VSI context depending on the VSI type to be added and
* passes it down to the add_vsi aq command to create a new VSI.
*/
-int ice_vsi_init(struct ice_vsi *vsi)
+static int ice_vsi_init(struct ice_vsi *vsi)
{
struct ice_vsi_ctx ctxt = { 0 };
struct ice_pf *pf = vsi->back;
@@ -922,7 +986,7 @@ static int ice_vsi_alloc_q_vector(struct ice_vsi *vsi, int v_idx)
* We allocate one q_vector per queue interrupt. If allocation fails we
* return -ENOMEM.
*/
-int ice_vsi_alloc_q_vectors(struct ice_vsi *vsi)
+static int ice_vsi_alloc_q_vectors(struct ice_vsi *vsi)
{
struct ice_pf *pf = vsi->back;
int v_idx = 0, num_q_vectors;
@@ -970,7 +1034,7 @@ int ice_vsi_alloc_q_vectors(struct ice_vsi *vsi)
*
* Returns 0 on success or negative on failure
*/
-int ice_vsi_setup_vector_base(struct ice_vsi *vsi)
+static int ice_vsi_setup_vector_base(struct ice_vsi *vsi)
{
struct ice_pf *pf = vsi->back;
int num_q_vectors = 0;
@@ -1038,7 +1102,7 @@ static void ice_vsi_clear_rings(struct ice_vsi *vsi)
* ice_vsi_alloc_rings - Allocates Tx and Rx rings for the VSI
* @vsi: VSI which is having rings allocated
*/
-int ice_vsi_alloc_rings(struct ice_vsi *vsi)
+static int ice_vsi_alloc_rings(struct ice_vsi *vsi)
{
struct ice_pf *pf = vsi->back;
int i;
@@ -1096,7 +1160,7 @@ int ice_vsi_alloc_rings(struct ice_vsi *vsi)
* through the MSI-X enabling code. On a constrained vector budget, we map Tx
* and Rx rings to the vector as "efficiently" as possible.
*/
-void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
+static void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
{
int q_vectors = vsi->num_q_vectors;
int tx_rings_rem, rx_rings_rem;
@@ -1142,6 +1206,69 @@ void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
}
}
+/**
+ * ice_vsi_cfg_rss_lut_key - Configure RSS params for a VSI
+ * @vsi: VSI to be configured
+ */
+static int ice_vsi_cfg_rss_lut_key(struct ice_vsi *vsi)
+{
+ u8 seed[ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE];
+ struct ice_aqc_get_set_rss_keys *key;
+ struct ice_pf *pf = vsi->back;
+ enum ice_status status;
+ int err = 0;
+ u8 *lut;
+
+ vsi->rss_size = min_t(int, vsi->rss_size, vsi->num_rxq);
+
+ lut = devm_kzalloc(&pf->pdev->dev, vsi->rss_table_size, GFP_KERNEL);
+ if (!lut)
+ return -ENOMEM;
+
+ if (vsi->rss_lut_user)
+ memcpy(lut, vsi->rss_lut_user, vsi->rss_table_size);
+ else
+ ice_fill_rss_lut(lut, vsi->rss_table_size, vsi->rss_size);
+
+ status = ice_aq_set_rss_lut(&pf->hw, vsi->vsi_num, vsi->rss_lut_type,
+ lut, vsi->rss_table_size);
+
+ if (status) {
+ dev_err(&vsi->back->pdev->dev,
+ "set_rss_lut failed, error %d\n", status);
+ err = -EIO;
+ goto ice_vsi_cfg_rss_exit;
+ }
+
+ key = devm_kzalloc(&vsi->back->pdev->dev, sizeof(*key), GFP_KERNEL);
+ if (!key) {
+ err = -ENOMEM;
+ goto ice_vsi_cfg_rss_exit;
+ }
+
+ if (vsi->rss_hkey_user)
+ memcpy(seed, vsi->rss_hkey_user,
+ ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE);
+ else
+ netdev_rss_key_fill((void *)seed,
+ ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE);
+ memcpy(&key->standard_rss_key, seed,
+ ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE);
+
+ status = ice_aq_set_rss_key(&pf->hw, vsi->vsi_num, key);
+
+ if (status) {
+ dev_err(&vsi->back->pdev->dev, "set_rss_key failed, error %d\n",
+ status);
+ err = -EIO;
+ }
+
+ devm_kfree(&pf->pdev->dev, key);
+ice_vsi_cfg_rss_exit:
+ devm_kfree(&pf->pdev->dev, lut);
+ return err;
+}
+
/**
* ice_add_mac_to_list - Add a mac address filter entry to the list
* @vsi: the VSI to be forwarded to
@@ -1722,6 +1849,112 @@ int ice_cfg_vlan_pruning(struct ice_vsi *vsi, bool ena)
return -EIO;
}
+/**
+ * ice_vsi_setup - Set up a VSI by a given type
+ * @pf: board private structure
+ * @pi: pointer to the port_info instance
+ * @type: VSI type
+ * @vf_id: defines VF id to which this VSI connects. This field is meant to be
+ * used only for ICE_VSI_VF VSI type. For other VSI types, should
+ * fill-in ICE_INVAL_VFID as input.
+ *
+ * This allocates the sw VSI structure and its queue resources.
+ *
+ * Returns pointer to the successfully allocated and configured VSI sw struct on
+ * success, NULL on failure.
+ */
+struct ice_vsi *
+ice_vsi_setup(struct ice_pf *pf, struct ice_port_info *pi,
+ enum ice_vsi_type type, u16 __always_unused vf_id)
+{
+ u16 max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 };
+ struct device *dev = &pf->pdev->dev;
+ struct ice_vsi *vsi;
+ int ret, i;
+
+ vsi = ice_vsi_alloc(pf, type);
+ if (!vsi) {
+ dev_err(dev, "could not allocate VSI\n");
+ return NULL;
+ }
+
+ vsi->port_info = pi;
+ vsi->vsw = pf->first_sw;
+
+ if (ice_vsi_get_qs(vsi)) {
+ dev_err(dev, "Failed to allocate queues. vsi->idx = %d\n",
+ vsi->idx);
+ goto unroll_get_qs;
+ }
+
+ /* set RSS capabilities */
+ ice_vsi_set_rss_params(vsi);
+
+ /* create the VSI */
+ ret = ice_vsi_init(vsi);
+ if (ret)
+ goto unroll_get_qs;
+
+ switch (vsi->type) {
+ case ICE_VSI_PF:
+ ret = ice_vsi_alloc_q_vectors(vsi);
+ if (ret)
+ goto unroll_vsi_init;
+
+ ret = ice_vsi_setup_vector_base(vsi);
+ if (ret)
+ goto unroll_alloc_q_vector;
+
+ ret = ice_vsi_alloc_rings(vsi);
+ if (ret)
+ goto unroll_vector_base;
+
+ ice_vsi_map_rings_to_vectors(vsi);
+
+ /* Do not exit if configuring RSS had an issue, at least
+ * receive traffic on first queue. Hence no need to capture
+ * return value
+ */
+ if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
+ ice_vsi_cfg_rss_lut_key(vsi);
+ break;
+ default:
+ /* if VSI type is not recognized, clean up the resources and
+ * exit
+ */
+ goto unroll_vsi_init;
+ }
+
+ ice_vsi_set_tc_cfg(vsi);
+
+ /* configure VSI nodes based on number of queues and TC's */
+ for (i = 0; i < vsi->tc_cfg.numtc; i++)
+ max_txqs[i] = vsi->num_txq;
+
+ ret = ice_cfg_vsi_lan(vsi->port_info, vsi->vsi_num,
+ vsi->tc_cfg.ena_tc, max_txqs);
+ if (ret) {
+ dev_info(&pf->pdev->dev, "Failed VSI lan queue config\n");
+ goto unroll_vector_base;
+ }
+
+ return vsi;
+
+unroll_vector_base:
+ ice_free_res(vsi->back->irq_tracker, vsi->base_vector, vsi->idx);
+unroll_alloc_q_vector:
+ ice_vsi_free_q_vectors(vsi);
+unroll_vsi_init:
+ ice_vsi_delete(vsi);
+unroll_get_qs:
+ ice_vsi_put_qs(vsi);
+ pf->q_left_tx += vsi->alloc_txq;
+ pf->q_left_rx += vsi->alloc_rxq;
+ ice_vsi_clear(vsi);
+
+ return NULL;
+}
+
/**
* ice_vsi_release_msix - Clear the queue to Interrupt mapping in HW
* @vsi: the VSI being cleaned up
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
index aaab3fc4b018..a76cde895bf3 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -6,26 +6,6 @@
#include "ice.h"
-int ice_vsi_setup_vector_base(struct ice_vsi *vsi);
-
-int ice_vsi_alloc_q_vectors(struct ice_vsi *vsi);
-
-int ice_vsi_get_qs(struct ice_vsi *vsi);
-
-void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi);
-
-int ice_vsi_alloc_rings(struct ice_vsi *vsi);
-
-void ice_vsi_set_rss_params(struct ice_vsi *vsi);
-
-void ice_vsi_set_num_qs(struct ice_vsi *vsi);
-
-int ice_get_free_slot(void *array, int size, int curr);
-
-int ice_vsi_init(struct ice_vsi *vsi);
-
-int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors);
-
int ice_add_mac_to_list(struct ice_vsi *vsi, struct list_head *add_list,
const u8 *macaddr);
@@ -59,6 +39,10 @@ void ice_vsi_delete(struct ice_vsi *vsi);
int ice_vsi_clear(struct ice_vsi *vsi);
+struct ice_vsi *
+ice_vsi_setup(struct ice_pf *pf, struct ice_port_info *pi,
+ enum ice_vsi_type type, u16 vf_id);
+
int ice_vsi_release(struct ice_vsi *vsi);
void ice_vsi_close(struct ice_vsi *vsi);
@@ -84,5 +68,7 @@ void ice_vsi_free_rx_rings(struct ice_vsi *vsi);
void ice_vsi_free_tx_rings(struct ice_vsi *vsi);
+int ice_vsi_cfg_tc(struct ice_vsi *vsi, u8 ena_tc);
+
irqreturn_t ice_msix_clean_rings(int __always_unused irq, void *data);
#endif /* !_ICE_LIB_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 3927b18b45a0..58cb2edd1c74 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1312,70 +1312,6 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data)
return ret;
}
-/**
- * ice_vsi_alloc - Allocates the next available struct vsi in the PF
- * @pf: board private structure
- * @type: type of VSI
- *
- * returns a pointer to a VSI on success, NULL on failure.
- */
-static struct ice_vsi *ice_vsi_alloc(struct ice_pf *pf, enum ice_vsi_type type)
-{
- struct ice_vsi *vsi = NULL;
-
- /* Need to protect the allocation of the VSIs at the PF level */
- mutex_lock(&pf->sw_mutex);
-
- /* If we have already allocated our maximum number of VSIs,
- * pf->next_vsi will be ICE_NO_VSI. If not, pf->next_vsi index
- * is available to be populated
- */
- if (pf->next_vsi == ICE_NO_VSI) {
- dev_dbg(&pf->pdev->dev, "out of VSI slots!\n");
- goto unlock_pf;
- }
-
- vsi = devm_kzalloc(&pf->pdev->dev, sizeof(*vsi), GFP_KERNEL);
- if (!vsi)
- goto unlock_pf;
-
- vsi->type = type;
- vsi->back = pf;
- set_bit(__ICE_DOWN, vsi->state);
- vsi->idx = pf->next_vsi;
- vsi->work_lmt = ICE_DFLT_IRQ_WORK;
-
- ice_vsi_set_num_qs(vsi);
-
- switch (vsi->type) {
- case ICE_VSI_PF:
- if (ice_vsi_alloc_arrays(vsi, true))
- goto err_rings;
-
- /* Setup default MSIX irq handler for VSI */
- vsi->irq_handler = ice_msix_clean_rings;
- break;
- default:
- dev_warn(&pf->pdev->dev, "Unknown VSI type %d\n", vsi->type);
- goto unlock_pf;
- }
-
- /* fill VSI slot in the PF struct */
- pf->vsi[pf->next_vsi] = vsi;
-
- /* prepare pf->next_vsi for next use */
- pf->next_vsi = ice_get_free_slot(pf->vsi, pf->num_alloc_vsi,
- pf->next_vsi);
- goto unlock_pf;
-
-err_rings:
- devm_kfree(&pf->pdev->dev, vsi);
- vsi = NULL;
-unlock_pf:
- mutex_unlock(&pf->sw_mutex);
- return vsi;
-}
-
/**
* ice_free_irq_msix_misc - Unroll misc vector setup
* @pf: board private structure
@@ -1597,194 +1533,6 @@ void ice_fill_rss_lut(u8 *lut, u16 rss_table_size, u16 rss_size)
lut[i] = i % rss_size;
}
-/**
- * ice_vsi_cfg_rss - Configure RSS params for a VSI
- * @vsi: VSI to be configured
- */
-static int ice_vsi_cfg_rss(struct ice_vsi *vsi)
-{
- u8 seed[ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE];
- struct ice_aqc_get_set_rss_keys *key;
- struct ice_pf *pf = vsi->back;
- enum ice_status status;
- int err = 0;
- u8 *lut;
-
- vsi->rss_size = min_t(int, vsi->rss_size, vsi->num_rxq);
-
- lut = devm_kzalloc(&pf->pdev->dev, vsi->rss_table_size, GFP_KERNEL);
- if (!lut)
- return -ENOMEM;
-
- if (vsi->rss_lut_user)
- memcpy(lut, vsi->rss_lut_user, vsi->rss_table_size);
- else
- ice_fill_rss_lut(lut, vsi->rss_table_size, vsi->rss_size);
-
- status = ice_aq_set_rss_lut(&pf->hw, vsi->vsi_num, vsi->rss_lut_type,
- lut, vsi->rss_table_size);
-
- if (status) {
- dev_err(&vsi->back->pdev->dev,
- "set_rss_lut failed, error %d\n", status);
- err = -EIO;
- goto ice_vsi_cfg_rss_exit;
- }
-
- key = devm_kzalloc(&vsi->back->pdev->dev, sizeof(*key), GFP_KERNEL);
- if (!key) {
- err = -ENOMEM;
- goto ice_vsi_cfg_rss_exit;
- }
-
- if (vsi->rss_hkey_user)
- memcpy(seed, vsi->rss_hkey_user,
- ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE);
- else
- netdev_rss_key_fill((void *)seed,
- ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE);
- memcpy(&key->standard_rss_key, seed,
- ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE);
-
- status = ice_aq_set_rss_key(&pf->hw, vsi->vsi_num, key);
-
- if (status) {
- dev_err(&vsi->back->pdev->dev, "set_rss_key failed, error %d\n",
- status);
- err = -EIO;
- }
-
- devm_kfree(&pf->pdev->dev, key);
-ice_vsi_cfg_rss_exit:
- devm_kfree(&pf->pdev->dev, lut);
- return err;
-}
-
-/**
- * ice_vsi_setup - Set up a VSI by a given type
- * @pf: board private structure
- * @pi: pointer to the port_info instance
- * @type: VSI type
- * @vf_id: defines VF id to which this VSI connects. This field is meant to be
- * used only for ICE_VSI_VF VSI type. For other VSI types, should
- * fill-in ICE_INVAL_VFID as input.
- *
- * This allocates the sw VSI structure and its queue resources.
- *
- * Returns pointer to the successfully allocated and configured VSI sw struct on
- * success, NULL on failure.
- */
-static struct ice_vsi *
-ice_vsi_setup(struct ice_pf *pf, struct ice_port_info *pi,
- enum ice_vsi_type type, u16 __always_unused vf_id)
-{
- u16 max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 };
- struct device *dev = &pf->pdev->dev;
- struct ice_vsi *vsi;
- int ret, i;
-
- vsi = ice_vsi_alloc(pf, type);
- if (!vsi) {
- dev_err(dev, "could not allocate VSI\n");
- return NULL;
- }
-
- vsi->port_info = pi;
- vsi->vsw = pf->first_sw;
-
- if (ice_vsi_get_qs(vsi)) {
- dev_err(dev, "Failed to allocate queues. vsi->idx = %d\n",
- vsi->idx);
- goto err_get_qs;
- }
-
- /* set RSS capabilities */
- ice_vsi_set_rss_params(vsi);
-
- /* create the VSI */
- ret = ice_vsi_init(vsi);
- if (ret)
- goto err_vsi;
-
- switch (vsi->type) {
- case ICE_VSI_PF:
- ret = ice_cfg_netdev(vsi);
- if (ret)
- goto err_cfg_netdev;
-
- ret = register_netdev(vsi->netdev);
- if (ret)
- goto err_register_netdev;
-
- netif_carrier_off(vsi->netdev);
-
- /* make sure transmit queues start off as stopped */
- netif_tx_stop_all_queues(vsi->netdev);
- ret = ice_vsi_alloc_q_vectors(vsi);
- if (ret)
- goto err_msix;
-
- ret = ice_vsi_setup_vector_base(vsi);
- if (ret)
- goto err_rings;
-
- ret = ice_vsi_alloc_rings(vsi);
- if (ret)
- goto err_rings;
-
- ice_vsi_map_rings_to_vectors(vsi);
-
- /* Do not exit if configuring RSS had an issue, at least
- * receive traffic on first queue. Hence no need to capture
- * return value
- */
- if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
- ice_vsi_cfg_rss(vsi);
- break;
- default:
- /* if vsi type is not recognized, clean up the resources and
- * exit
- */
- goto err_rings;
- }
-
- ice_vsi_set_tc_cfg(vsi);
-
- /* configure VSI nodes based on number of queues and TC's */
- for (i = 0; i < vsi->tc_cfg.numtc; i++)
- max_txqs[i] = vsi->num_txq;
-
- ret = ice_cfg_vsi_lan(vsi->port_info, vsi->vsi_num,
- vsi->tc_cfg.ena_tc, max_txqs);
- if (ret) {
- dev_info(&pf->pdev->dev, "Failed VSI lan queue config\n");
- goto err_rings;
- }
-
- return vsi;
-
-err_rings:
- ice_vsi_free_q_vectors(vsi);
-err_msix:
- if (vsi->netdev && vsi->netdev->reg_state == NETREG_REGISTERED)
- unregister_netdev(vsi->netdev);
-err_register_netdev:
- if (vsi->netdev) {
- free_netdev(vsi->netdev);
- vsi->netdev = NULL;
- }
-err_cfg_netdev:
- ice_vsi_delete(vsi);
-err_vsi:
- ice_vsi_put_qs(vsi);
-err_get_qs:
- pf->q_left_tx += vsi->alloc_txq;
- pf->q_left_rx += vsi->alloc_rxq;
- ice_vsi_clear(vsi);
-
- return NULL;
-}
-
/**
* ice_pf_vsi_setup - Set up a PF VSI
* @pf: board private structure
--
2.17.1
^ permalink raw reply related
* [net-next 5/8] ice: Move common functions out of ice_main.c part 5/7
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem
Cc: Anirudh Venkataramanan, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <20181001211431.14325-1-jeffrey.t.kirsher@intel.com>
From: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
This patch continues the code move out of ice_main.c
The following top level functions (and related dependency functions) were
moved to ice_lib.c:
ice_vsi_clear
ice_vsi_close
ice_vsi_free_arrays
ice_vsi_map_rings_to_vectors
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_lib.c | 133 ++++++++++++++++++++++
drivers/net/ethernet/intel/ice/ice_lib.h | 8 ++
drivers/net/ethernet/intel/ice/ice_main.c | 132 ---------------------
3 files changed, 141 insertions(+), 132 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index df20d68c92ab..6ba82337d017 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -341,6 +341,71 @@ void ice_vsi_delete(struct ice_vsi *vsi)
vsi->vsi_num);
}
+/**
+ * ice_vsi_free_arrays - clean up VSI resources
+ * @vsi: pointer to VSI being cleared
+ * @free_qvectors: bool to specify if q_vectors should be deallocated
+ */
+void ice_vsi_free_arrays(struct ice_vsi *vsi, bool free_qvectors)
+{
+ struct ice_pf *pf = vsi->back;
+
+ /* free the ring and vector containers */
+ if (free_qvectors && vsi->q_vectors) {
+ devm_kfree(&pf->pdev->dev, vsi->q_vectors);
+ vsi->q_vectors = NULL;
+ }
+ if (vsi->tx_rings) {
+ devm_kfree(&pf->pdev->dev, vsi->tx_rings);
+ vsi->tx_rings = NULL;
+ }
+ if (vsi->rx_rings) {
+ devm_kfree(&pf->pdev->dev, vsi->rx_rings);
+ vsi->rx_rings = NULL;
+ }
+}
+
+/**
+ * ice_vsi_clear - clean up and deallocate the provided VSI
+ * @vsi: pointer to VSI being cleared
+ *
+ * This deallocates the VSI's queue resources, removes it from the PF's
+ * VSI array if necessary, and deallocates the VSI
+ *
+ * Returns 0 on success, negative on failure
+ */
+int ice_vsi_clear(struct ice_vsi *vsi)
+{
+ struct ice_pf *pf = NULL;
+
+ if (!vsi)
+ return 0;
+
+ if (!vsi->back)
+ return -EINVAL;
+
+ pf = vsi->back;
+
+ if (!pf->vsi[vsi->idx] || pf->vsi[vsi->idx] != vsi) {
+ dev_dbg(&pf->pdev->dev, "vsi does not exist at pf->vsi[%d]\n",
+ vsi->idx);
+ return -EINVAL;
+ }
+
+ mutex_lock(&pf->sw_mutex);
+ /* updates the PF for this cleared VSI */
+
+ pf->vsi[vsi->idx] = NULL;
+ if (vsi->idx < pf->next_vsi)
+ pf->next_vsi = vsi->idx;
+
+ ice_vsi_free_arrays(vsi, true);
+ mutex_unlock(&pf->sw_mutex);
+ devm_kfree(&pf->pdev->dev, vsi);
+
+ return 0;
+}
+
/**
* ice_msix_clean_rings - MSIX mode Interrupt Handler
* @irq: interrupt number
@@ -700,6 +765,60 @@ int ice_vsi_alloc_rings(struct ice_vsi *vsi)
return -ENOMEM;
}
+/**
+ * ice_vsi_map_rings_to_vectors - Map VSI rings to interrupt vectors
+ * @vsi: the VSI being configured
+ *
+ * This function maps descriptor rings to the queue-specific vectors allotted
+ * through the MSI-X enabling code. On a constrained vector budget, we map Tx
+ * and Rx rings to the vector as "efficiently" as possible.
+ */
+void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
+{
+ int q_vectors = vsi->num_q_vectors;
+ int tx_rings_rem, rx_rings_rem;
+ int v_id;
+
+ /* initially assigning remaining rings count to VSIs num queue value */
+ tx_rings_rem = vsi->num_txq;
+ rx_rings_rem = vsi->num_rxq;
+
+ for (v_id = 0; v_id < q_vectors; v_id++) {
+ struct ice_q_vector *q_vector = vsi->q_vectors[v_id];
+ int tx_rings_per_v, rx_rings_per_v, q_id, q_base;
+
+ /* Tx rings mapping to vector */
+ tx_rings_per_v = DIV_ROUND_UP(tx_rings_rem, q_vectors - v_id);
+ q_vector->num_ring_tx = tx_rings_per_v;
+ q_vector->tx.ring = NULL;
+ q_base = vsi->num_txq - tx_rings_rem;
+
+ for (q_id = q_base; q_id < (q_base + tx_rings_per_v); q_id++) {
+ struct ice_ring *tx_ring = vsi->tx_rings[q_id];
+
+ tx_ring->q_vector = q_vector;
+ tx_ring->next = q_vector->tx.ring;
+ q_vector->tx.ring = tx_ring;
+ }
+ tx_rings_rem -= tx_rings_per_v;
+
+ /* Rx rings mapping to vector */
+ rx_rings_per_v = DIV_ROUND_UP(rx_rings_rem, q_vectors - v_id);
+ q_vector->num_ring_rx = rx_rings_per_v;
+ q_vector->rx.ring = NULL;
+ q_base = vsi->num_rxq - rx_rings_rem;
+
+ for (q_id = q_base; q_id < (q_base + rx_rings_per_v); q_id++) {
+ struct ice_ring *rx_ring = vsi->rx_rings[q_id];
+
+ rx_ring->q_vector = q_vector;
+ rx_ring->next = q_vector->rx.ring;
+ q_vector->rx.ring = rx_ring;
+ }
+ rx_rings_rem -= rx_rings_per_v;
+ }
+}
+
/**
* ice_add_mac_to_list - Add a mac address filter entry to the list
* @vsi: the VSI to be forwarded to
@@ -1385,6 +1504,20 @@ void ice_vsi_free_rx_rings(struct ice_vsi *vsi)
ice_free_rx_ring(vsi->rx_rings[i]);
}
+/**
+ * ice_vsi_close - Shut down a VSI
+ * @vsi: the VSI being shut down
+ */
+void ice_vsi_close(struct ice_vsi *vsi)
+{
+ if (!test_and_set_bit(__ICE_DOWN, vsi->state))
+ ice_down(vsi);
+
+ ice_vsi_free_irq(vsi);
+ ice_vsi_free_tx_rings(vsi);
+ ice_vsi_free_rx_rings(vsi);
+}
+
/**
* ice_free_res - free a block of resources
* @res: pointer to the resource
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
index ab5a4e1edcf8..002bbca8e7ea 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -6,6 +6,8 @@
#include "ice.h"
+void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi);
+
int ice_vsi_alloc_rings(struct ice_vsi *vsi);
void ice_vsi_set_rss_params(struct ice_vsi *vsi);
@@ -16,6 +18,8 @@ int ice_get_free_slot(void *array, int size, int curr);
int ice_vsi_init(struct ice_vsi *vsi);
+void ice_vsi_free_arrays(struct ice_vsi *vsi, bool free_qvectors);
+
void ice_vsi_clear_rings(struct ice_vsi *vsi);
int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors);
@@ -51,6 +55,10 @@ int ice_cfg_vlan_pruning(struct ice_vsi *vsi, bool ena);
void ice_vsi_delete(struct ice_vsi *vsi);
+int ice_vsi_clear(struct ice_vsi *vsi);
+
+void ice_vsi_close(struct ice_vsi *vsi);
+
int ice_free_res(struct ice_res_tracker *res, u16 index, u16 id);
int
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 73a6574c0d7d..f458ff285a34 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1313,60 +1313,6 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data)
return ret;
}
-/**
- * ice_vsi_map_rings_to_vectors - Map VSI rings to interrupt vectors
- * @vsi: the VSI being configured
- *
- * This function maps descriptor rings to the queue-specific vectors allotted
- * through the MSI-X enabling code. On a constrained vector budget, we map Tx
- * and Rx rings to the vector as "efficiently" as possible.
- */
-static void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
-{
- int q_vectors = vsi->num_q_vectors;
- int tx_rings_rem, rx_rings_rem;
- int v_id;
-
- /* initially assigning remaining rings count to VSIs num queue value */
- tx_rings_rem = vsi->num_txq;
- rx_rings_rem = vsi->num_rxq;
-
- for (v_id = 0; v_id < q_vectors; v_id++) {
- struct ice_q_vector *q_vector = vsi->q_vectors[v_id];
- int tx_rings_per_v, rx_rings_per_v, q_id, q_base;
-
- /* Tx rings mapping to vector */
- tx_rings_per_v = DIV_ROUND_UP(tx_rings_rem, q_vectors - v_id);
- q_vector->num_ring_tx = tx_rings_per_v;
- q_vector->tx.ring = NULL;
- q_base = vsi->num_txq - tx_rings_rem;
-
- for (q_id = q_base; q_id < (q_base + tx_rings_per_v); q_id++) {
- struct ice_ring *tx_ring = vsi->tx_rings[q_id];
-
- tx_ring->q_vector = q_vector;
- tx_ring->next = q_vector->tx.ring;
- q_vector->tx.ring = tx_ring;
- }
- tx_rings_rem -= tx_rings_per_v;
-
- /* Rx rings mapping to vector */
- rx_rings_per_v = DIV_ROUND_UP(rx_rings_rem, q_vectors - v_id);
- q_vector->num_ring_rx = rx_rings_per_v;
- q_vector->rx.ring = NULL;
- q_base = vsi->num_rxq - rx_rings_rem;
-
- for (q_id = q_base; q_id < (q_base + rx_rings_per_v); q_id++) {
- struct ice_ring *rx_ring = vsi->rx_rings[q_id];
-
- rx_ring->q_vector = q_vector;
- rx_ring->next = q_vector->rx.ring;
- q_vector->rx.ring = rx_ring;
- }
- rx_rings_rem -= rx_rings_per_v;
- }
-}
-
/**
* ice_vsi_alloc - Allocates the next available struct vsi in the PF
* @pf: board private structure
@@ -1770,71 +1716,6 @@ static int ice_cfg_netdev(struct ice_vsi *vsi)
return 0;
}
-/**
- * ice_vsi_free_arrays - clean up vsi resources
- * @vsi: pointer to VSI being cleared
- * @free_qvectors: bool to specify if q_vectors should be deallocated
- */
-static void ice_vsi_free_arrays(struct ice_vsi *vsi, bool free_qvectors)
-{
- struct ice_pf *pf = vsi->back;
-
- /* free the ring and vector containers */
- if (free_qvectors && vsi->q_vectors) {
- devm_kfree(&pf->pdev->dev, vsi->q_vectors);
- vsi->q_vectors = NULL;
- }
- if (vsi->tx_rings) {
- devm_kfree(&pf->pdev->dev, vsi->tx_rings);
- vsi->tx_rings = NULL;
- }
- if (vsi->rx_rings) {
- devm_kfree(&pf->pdev->dev, vsi->rx_rings);
- vsi->rx_rings = NULL;
- }
-}
-
-/**
- * ice_vsi_clear - clean up and deallocate the provided vsi
- * @vsi: pointer to VSI being cleared
- *
- * This deallocates the vsi's queue resources, removes it from the PF's
- * VSI array if necessary, and deallocates the VSI
- *
- * Returns 0 on success, negative on failure
- */
-static int ice_vsi_clear(struct ice_vsi *vsi)
-{
- struct ice_pf *pf = NULL;
-
- if (!vsi)
- return 0;
-
- if (!vsi->back)
- return -EINVAL;
-
- pf = vsi->back;
-
- if (!pf->vsi[vsi->idx] || pf->vsi[vsi->idx] != vsi) {
- dev_dbg(&pf->pdev->dev, "vsi does not exist at pf->vsi[%d]\n",
- vsi->idx);
- return -EINVAL;
- }
-
- mutex_lock(&pf->sw_mutex);
- /* updates the PF for this cleared vsi */
-
- pf->vsi[vsi->idx] = NULL;
- if (vsi->idx < pf->next_vsi)
- pf->next_vsi = vsi->idx;
-
- ice_vsi_free_arrays(vsi, true);
- mutex_unlock(&pf->sw_mutex);
- devm_kfree(&pf->pdev->dev, vsi);
-
- return 0;
-}
-
/**
* ice_vsi_alloc_q_vector - Allocate memory for a single interrupt vector
* @vsi: the VSI being configured
@@ -3709,19 +3590,6 @@ static int ice_vsi_open(struct ice_vsi *vsi)
return err;
}
-/**
- * ice_vsi_close - Shut down a VSI
- * @vsi: the VSI being shut down
- */
-static void ice_vsi_close(struct ice_vsi *vsi)
-{
- if (!test_and_set_bit(__ICE_DOWN, vsi->state))
- ice_down(vsi);
-
- ice_vsi_free_irq(vsi);
- ice_vsi_free_tx_rings(vsi);
- ice_vsi_free_rx_rings(vsi);
-}
/**
* ice_rss_clean - Delete RSS related VSI structures that hold user inputs
--
2.17.1
^ permalink raw reply related
* [net-next 8/8] ice: Change pf state behavior to protect reset path
From: Jeff Kirsher @ 2018-10-01 21:14 UTC (permalink / raw)
To: davem
Cc: Dave Ertman, netdev, nhorman, sassmann, jogreene,
Anirudh Venkataramanan, Jeff Kirsher
In-Reply-To: <20181001211431.14325-1-jeffrey.t.kirsher@intel.com>
From: Dave Ertman <david.m.ertman@intel.com>
Currently, there is no bit, or set of bits, that protect the entirety
of the reset path.
If the reset is originated by the driver, then the relevant
one of the following bits will be set when the reset is scheduled:
__ICE_PFR_REQ
__ICE_CORER_REQ
__ICE_GLOBR_REQ
This bit will not be cleared until after the rebuild has completed.
If the reset is originated by the FW, then the first the driver knows of
it will be the reception of the OICR interrupt. The __ICE_RESET_OICR_RECV
bit will be set in the interrupt handler. This will also be the indicator
in a SW originated reset that we have completed the pre-OICR tasks and
have informed the FW that a reset was requested.
To utilize these bits, change the function:
ice_is_reset_recovery_pending()
to be:
ice_is_reset_in_progress()
The new function will check all of the above bits in the pf->state and
will return a true if one or more of these bits are set.
Signed-off-by: Dave Ertman <david.m.ertman@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice.h | 2 +-
drivers/net/ethernet/intel/ice/ice_lib.c | 13 ++++---
drivers/net/ethernet/intel/ice/ice_lib.h | 2 +-
drivers/net/ethernet/intel/ice/ice_main.c | 44 +++++++++++------------
4 files changed, 31 insertions(+), 30 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index e84a612ffa71..9cce4cb91401 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -124,7 +124,7 @@ enum ice_state {
__ICE_DOWN,
__ICE_NEEDS_RESTART,
__ICE_PREPARED_FOR_RESET, /* set by driver when prepared */
- __ICE_RESET_RECOVERY_PENDING, /* set by driver when reset starts */
+ __ICE_RESET_OICR_RECV, /* set by driver after rcv reset OICR */
__ICE_PFR_REQ, /* set by driver and peers */
__ICE_CORER_REQ, /* set by driver and peers */
__ICE_GLOBR_REQ, /* set by driver and peers */
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 21e3a3e70329..95588fe0e22f 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -2250,7 +2250,7 @@ int ice_vsi_release(struct ice_vsi *vsi)
* currently. This is done to avoid check_flush_dependency() warning
* on this wq
*/
- if (vsi->netdev && !ice_is_reset_recovery_pending(pf->state)) {
+ if (vsi->netdev && !ice_is_reset_in_progress(pf->state)) {
unregister_netdev(vsi->netdev);
free_netdev(vsi->netdev);
vsi->netdev = NULL;
@@ -2280,7 +2280,7 @@ int ice_vsi_release(struct ice_vsi *vsi)
* free VSI netdev when PF is not in reset recovery pending state,\
* for ex: during rmmod.
*/
- if (!ice_is_reset_recovery_pending(pf->state))
+ if (!ice_is_reset_in_progress(pf->state))
ice_vsi_clear(vsi);
return 0;
@@ -2367,10 +2367,13 @@ int ice_vsi_rebuild(struct ice_vsi *vsi)
}
/**
- * ice_is_reset_recovery_pending - schedule a reset
+ * ice_is_reset_in_progress - check for a reset in progress
* @state: pf state field
*/
-bool ice_is_reset_recovery_pending(unsigned long *state)
+bool ice_is_reset_in_progress(unsigned long *state)
{
- return test_bit(__ICE_RESET_RECOVERY_PENDING, state);
+ return test_bit(__ICE_RESET_OICR_RECV, state) ||
+ test_bit(__ICE_PFR_REQ, state) ||
+ test_bit(__ICE_CORER_REQ, state) ||
+ test_bit(__ICE_GLOBR_REQ, state);
}
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h
index a76cde895bf3..4265464ee3d3 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_lib.h
@@ -54,7 +54,7 @@ ice_get_res(struct ice_pf *pf, struct ice_res_tracker *res, u16 needed, u16 id);
int ice_vsi_rebuild(struct ice_vsi *vsi);
-bool ice_is_reset_recovery_pending(unsigned long *state);
+bool ice_is_reset_in_progress(unsigned long *state);
void ice_vsi_free_q_vectors(struct ice_vsi *vsi);
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 58cb2edd1c74..a3513acd272b 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -364,21 +364,17 @@ static void ice_do_reset(struct ice_pf *pf, enum ice_reset_req reset_type)
dev_dbg(dev, "reset_type 0x%x requested\n", reset_type);
WARN_ON(in_interrupt());
- /* PFR is a bit of a special case because it doesn't result in an OICR
- * interrupt. Set pending bit here which otherwise gets set in the
- * OICR handler.
- */
- if (reset_type == ICE_RESET_PFR)
- set_bit(__ICE_RESET_RECOVERY_PENDING, pf->state);
-
ice_prepare_for_reset(pf);
/* trigger the reset */
if (ice_reset(hw, reset_type)) {
dev_err(dev, "reset %d failed\n", reset_type);
set_bit(__ICE_RESET_FAILED, pf->state);
- clear_bit(__ICE_RESET_RECOVERY_PENDING, pf->state);
+ clear_bit(__ICE_RESET_OICR_RECV, pf->state);
clear_bit(__ICE_PREPARED_FOR_RESET, pf->state);
+ clear_bit(__ICE_PFR_REQ, pf->state);
+ clear_bit(__ICE_CORER_REQ, pf->state);
+ clear_bit(__ICE_GLOBR_REQ, pf->state);
return;
}
@@ -389,8 +385,8 @@ static void ice_do_reset(struct ice_pf *pf, enum ice_reset_req reset_type)
if (reset_type == ICE_RESET_PFR) {
pf->pfr_count++;
ice_rebuild(pf);
- clear_bit(__ICE_RESET_RECOVERY_PENDING, pf->state);
clear_bit(__ICE_PREPARED_FOR_RESET, pf->state);
+ clear_bit(__ICE_PFR_REQ, pf->state);
}
}
@@ -405,14 +401,14 @@ static void ice_reset_subtask(struct ice_pf *pf)
/* When a CORER/GLOBR/EMPR is about to happen, the hardware triggers an
* OICR interrupt. The OICR handler (ice_misc_intr) determines what type
* of reset is pending and sets bits in pf->state indicating the reset
- * type and __ICE_RESET_RECOVERY_PENDING. So, if the latter bit is set
+ * type and __ICE_RESET_OICR_RECV. So, if the latter bit is set
* prepare for pending reset if not already (for PF software-initiated
* global resets the software should already be prepared for it as
* indicated by __ICE_PREPARED_FOR_RESET; for global resets initiated
* by firmware or software on other PFs, that bit is not set so prepare
* for the reset now), poll for reset done, rebuild and return.
*/
- if (ice_is_reset_recovery_pending(pf->state)) {
+ if (test_bit(__ICE_RESET_OICR_RECV, pf->state)) {
clear_bit(__ICE_GLOBR_RECV, pf->state);
clear_bit(__ICE_CORER_RECV, pf->state);
if (!test_bit(__ICE_PREPARED_FOR_RESET, pf->state))
@@ -428,19 +424,22 @@ static void ice_reset_subtask(struct ice_pf *pf)
/* clear bit to resume normal operations, but
* ICE_NEEDS_RESTART bit is set incase rebuild failed
*/
- clear_bit(__ICE_RESET_RECOVERY_PENDING, pf->state);
+ clear_bit(__ICE_RESET_OICR_RECV, pf->state);
clear_bit(__ICE_PREPARED_FOR_RESET, pf->state);
+ clear_bit(__ICE_PFR_REQ, pf->state);
+ clear_bit(__ICE_CORER_REQ, pf->state);
+ clear_bit(__ICE_GLOBR_REQ, pf->state);
}
return;
}
/* No pending resets to finish processing. Check for new resets */
- if (test_and_clear_bit(__ICE_PFR_REQ, pf->state))
+ if (test_bit(__ICE_PFR_REQ, pf->state))
reset_type = ICE_RESET_PFR;
- if (test_and_clear_bit(__ICE_CORER_REQ, pf->state))
+ if (test_bit(__ICE_CORER_REQ, pf->state))
reset_type = ICE_RESET_CORER;
- if (test_and_clear_bit(__ICE_GLOBR_REQ, pf->state))
+ if (test_bit(__ICE_GLOBR_REQ, pf->state))
reset_type = ICE_RESET_GLOBR;
/* If no valid reset type requested just return */
if (reset_type == ICE_RESET_INVAL)
@@ -1029,7 +1028,7 @@ static void ice_service_task(struct work_struct *work)
ice_reset_subtask(pf);
/* bail if a reset/recovery cycle is pending or rebuild failed */
- if (ice_is_reset_recovery_pending(pf->state) ||
+ if (ice_is_reset_in_progress(pf->state) ||
test_bit(__ICE_SUSPENDED, pf->state) ||
test_bit(__ICE_NEEDS_RESTART, pf->state)) {
ice_service_task_complete(pf);
@@ -1250,8 +1249,7 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data)
* We also make note of which reset happened so that peer
* devices/drivers can be informed.
*/
- if (!test_and_set_bit(__ICE_RESET_RECOVERY_PENDING,
- pf->state)) {
+ if (!test_and_set_bit(__ICE_RESET_OICR_RECV, pf->state)) {
if (reset == ICE_RESET_CORER)
set_bit(__ICE_CORER_RECV, pf->state);
else if (reset == ICE_RESET_GLOBR)
@@ -1265,7 +1263,7 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data)
* is received and set back to false after the driver
* has determined that the hardware is out of reset.
*
- * __ICE_RESET_RECOVERY_PENDING in pf->state indicates
+ * __ICE_RESET_OICR_RECV in pf->state indicates
* that a post reset rebuild is required before the
* driver is operational again. This is set above.
*
@@ -1355,7 +1353,7 @@ static int ice_req_irq_msix_misc(struct ice_pf *pf)
* lost during reset. Note that this function is called only during
* rebuild path and not while reset is in progress.
*/
- if (ice_is_reset_recovery_pending(pf->state))
+ if (ice_is_reset_in_progress(pf->state))
goto skip_req_irq;
/* reserve one vector in irq_tracker for misc interrupts */
@@ -1637,7 +1635,7 @@ static int ice_setup_pf_sw(struct ice_pf *pf)
struct ice_vsi *vsi;
int status = 0;
- if (ice_is_reset_recovery_pending(pf->state))
+ if (ice_is_reset_in_progress(pf->state))
return -EBUSY;
vsi = ice_pf_vsi_setup(pf, pf->hw.port_info);
@@ -2203,7 +2201,7 @@ static int ice_set_mac_address(struct net_device *netdev, void *pi)
}
if (test_bit(__ICE_DOWN, pf->state) ||
- ice_is_reset_recovery_pending(pf->state)) {
+ ice_is_reset_in_progress(pf->state)) {
netdev_err(netdev, "can't set mac %pM. device not ready\n",
mac);
return -EBUSY;
@@ -3256,7 +3254,7 @@ static int ice_change_mtu(struct net_device *netdev, int new_mtu)
}
/* if a reset is in progress, wait for some time for it to complete */
do {
- if (ice_is_reset_recovery_pending(pf->state)) {
+ if (ice_is_reset_in_progress(pf->state)) {
count++;
usleep_range(1000, 2000);
} else {
--
2.17.1
^ permalink raw reply related
* [PATCH net] tcp/dccp: fix lockdep issue when SYN is backlogged
From: Eric Dumazet @ 2018-10-01 22:02 UTC (permalink / raw)
To: David S . Miller; +Cc: netdev, Eric Dumazet, Eric Dumazet
In normal SYN processing, packets are handled without listener
lock and in RCU protected ingress path.
But syzkaller is known to be able to trick us and SYN
packets might be processed in process context, after being
queued into socket backlog.
In commit 06f877d613be ("tcp/dccp: fix other lockdep splats
accessing ireq_opt") I made a very stupid fix, that happened
to work mostly because of the regular path being RCU protected.
Really the thing protecting ireq->ireq_opt is RCU read lock,
and the pseudo request refcnt is not relevant.
This patch extends what I did in commit 449809a66c1d ("tcp/dccp:
block BH for SYN processing") by adding an extra rcu_read_{lock|unlock}
pair in the paths that might be taken when processing SYN from
socket backlog (thus possibly in process context)
Fixes: 06f877d613be ("tcp/dccp: fix other lockdep splats accessing ireq_opt")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
---
include/net/inet_sock.h | 3 +--
net/dccp/input.c | 4 +++-
net/ipv4/tcp_input.c | 4 +++-
3 files changed, 7 insertions(+), 4 deletions(-)
diff --git a/include/net/inet_sock.h b/include/net/inet_sock.h
index e03b93360f332b3e3232873ac1cbd0ee7478fabb..a8cd5cf9ff5b6ddc50bd2e70d3f9103afa32a3b5 100644
--- a/include/net/inet_sock.h
+++ b/include/net/inet_sock.h
@@ -132,8 +132,7 @@ static inline int inet_request_bound_dev_if(const struct sock *sk,
static inline struct ip_options_rcu *ireq_opt_deref(const struct inet_request_sock *ireq)
{
- return rcu_dereference_check(ireq->ireq_opt,
- refcount_read(&ireq->req.rsk_refcnt) > 0);
+ return rcu_dereference(ireq->ireq_opt);
}
struct inet_cork {
diff --git a/net/dccp/input.c b/net/dccp/input.c
index d28d46bff6ab43441f34284ec975c1e052a774d0..85d6c879383da8994c6b20cd1e49e0f667a07482 100644
--- a/net/dccp/input.c
+++ b/net/dccp/input.c
@@ -606,11 +606,13 @@ int dccp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
if (sk->sk_state == DCCP_LISTEN) {
if (dh->dccph_type == DCCP_PKT_REQUEST) {
/* It is possible that we process SYN packets from backlog,
- * so we need to make sure to disable BH right there.
+ * so we need to make sure to disable BH and RCU right there.
*/
+ rcu_read_lock();
local_bh_disable();
acceptable = inet_csk(sk)->icsk_af_ops->conn_request(sk, skb) >= 0;
local_bh_enable();
+ rcu_read_unlock();
if (!acceptable)
return 1;
consume_skb(skb);
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 4cf2f7bb2802ad4ae968b5a6dfb9d005ed619c76..47e08c1b5bc3e14e6ae2851b7ec8de91a3eb4a35 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -6009,11 +6009,13 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb)
if (th->fin)
goto discard;
/* It is possible that we process SYN packets from backlog,
- * so we need to make sure to disable BH right there.
+ * so we need to make sure to disable BH and RCU right there.
*/
+ rcu_read_lock();
local_bh_disable();
acceptable = icsk->icsk_af_ops->conn_request(sk, skb) >= 0;
local_bh_enable();
+ rcu_read_unlock();
if (!acceptable)
return 1;
--
2.19.0.605.g01d371f741-goog
^ permalink raw reply related
* Re: [PATCH net-next v2] tcp: up initial rmem to 128KB and SYN rwin to around 64KB
From: Yuchung Cheng @ 2018-10-01 22:18 UTC (permalink / raw)
To: David Miller
Cc: Eric Dumazet, netdev, Neal Cardwell, Wei Wang,
Soheil Hassas Yeganeh
In-Reply-To: <20180929.112314.687721909787322730.davem@davemloft.net>
On Sat, Sep 29, 2018 at 11:23 AM, David Miller <davem@davemloft.net> wrote:
>
> From: Yuchung Cheng <ycheng@google.com>
> Date: Fri, 28 Sep 2018 13:09:02 -0700
>
> > Previously TCP initial receive buffer is ~87KB by default and
> > the initial receive window is ~29KB (20 MSS). This patch changes
> > the two numbers to 128KB and ~64KB (rounding down to the multiples
> > of MSS) respectively. The patch also simplifies the calculations s.t.
> > the two numbers are directly controlled by sysctl tcp_rmem[1]:
> >
> > 1) Initial receiver buffer budget (sk_rcvbuf): while this should
> > be configured via sysctl tcp_rmem[1], previously tcp_fixup_rcvbuf()
> > always override and set a larger size when a new connection
> > establishes.
> >
> > 2) Initial receive window in SYN: previously it is set to 20
> > packets if MSS <= 1460. The number 20 was based on the initial
> > congestion window of 10: the receiver needs twice amount to
> > avoid being limited by the receive window upon out-of-order
> > delivery in the first window burst. But since this only
> > applies if the receiving MSS <= 1460, connection using large MTU
> > (e.g. to utilize receiver zero-copy) may be limited by the
> > receive window.
> >
> > This patch also lowers the initial bytes expected to receive in
> > the receiver buffer autotuning algorithm - otherwise the receiver
> > may take two to three rounds to increase the buffer to the
> > appropriate level (2x sender congestion window).
> >
> > With this patch TCP memory configuration is more straight-forward and
> > more properly sized to modern high-speed networks by default. Several
> > popular stacks have been announcing 64KB rwin in SYNs as well.
> >
> > Signed-off-by: Yuchung Cheng <ycheng@google.com>
> > Signed-off-by: Wei Wang <weiwan@google.com>
> > Signed-off-by: Neal Cardwell <ncardwell@google.com>
> > Signed-off-by: Eric Dumazet <edumazet@google.com>
> > Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
>
> Applied, thanks.
Hi David: thanks for taking this patch - I didn't notice this earlier
but it seems patch v1 was applied instead of v2? should I submit a
v2-v1-diff patch?
^ permalink raw reply
* Re: [PATCH net-next v2] tcp: up initial rmem to 128KB and SYN rwin to around 64KB
From: Eric Dumazet @ 2018-10-01 22:20 UTC (permalink / raw)
To: Yuchung Cheng
Cc: David Miller, netdev, Neal Cardwell, Wei Wang,
Soheil Hassas Yeganeh
In-Reply-To: <CAK6E8=dCC+5bP9gUrM6E42d9iYr2HNznQ74vhM=3xoD9ms90rA@mail.gmail.com>
On Mon, Oct 1, 2018 at 3:18 PM Yuchung Cheng <ycheng@google.com> wrote:
> Hi David: thanks for taking this patch - I didn't notice this earlier
> but it seems patch v1 was applied instead of v2? should I submit a
> v2-v1-diff patch?
Yes, please, send an additional patch.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox