* Re: [PATCH] udp: Fix typo in net/ipv4/udp.c
From: David Miller @ 2019-07-18 18:52 UTC (permalink / raw)
To: suyj.fnst; +Cc: kuznet, netdev, linux-kernel
In-Reply-To: <1563416363-7036-1-git-send-email-suyj.fnst@cn.fujitsu.com>
From: Su Yanjun <suyj.fnst@cn.fujitsu.com>
Date: Thu, 18 Jul 2019 10:19:23 +0800
> Signed-off-by: Su Yanjun <suyj.fnst@cn.fujitsu.com>
Several problems with this patch submission:
1) Please always make it clear when you post a new version of a patch
using indications in the Subject line such as "[PATCH v2]" or
similar.
Otherwise I assume it is a different change altogether and not
related to any other patch.
2) The targetted subsystem or tree must be clearly identified in the
subject line, for this you should say "[PATCH net v2]" since you
are targetting my networking bug fix GIT tree.
Thank you.
^ permalink raw reply
* Re: [PATCH bpf] tools/bpf: fix bpftool build with OUTPUT set
From: Jakub Kicinski @ 2019-07-18 18:51 UTC (permalink / raw)
To: Ilya Leoshkevich; +Cc: bpf, netdev, lmb, gor, heiko.carstens
In-Reply-To: <20190718142041.83342-1-iii@linux.ibm.com>
On Thu, 18 Jul 2019 16:20:41 +0200, Ilya Leoshkevich wrote:
> Hi Lorenz,
>
> I've been using the following patch for quite some time now.
> Please let me know if it works for you.
>
> Best regards,
> Ilya
>
> ---
>
> When OUTPUT is set, bpftool and libbpf put their objects into the same
> directory, and since some of them have the same names, the collision
> happens.
>
> Fix by invoking libbpf build in a manner similar to $(call descend) -
> descend itself cannot be used, since libbpf is a sibling, and not a
> child, of bpftool.
>
> Also, don't link bpftool with libbpf.a twice.
>
> Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
> ---
> tools/bpf/bpftool/Makefile | 17 ++++++-----------
> 1 file changed, 6 insertions(+), 11 deletions(-)
>
> diff --git a/tools/bpf/bpftool/Makefile b/tools/bpf/bpftool/Makefile
> index a7afea4dec47..2cbc3c166f44 100644
> --- a/tools/bpf/bpftool/Makefile
> +++ b/tools/bpf/bpftool/Makefile
> @@ -15,23 +15,18 @@ else
> endif
>
> BPF_DIR = $(srctree)/tools/lib/bpf/
> -
> -ifneq ($(OUTPUT),)
> - BPF_PATH = $(OUTPUT)
> -else
> - BPF_PATH = $(BPF_DIR)
> -endif
> -
> -LIBBPF = $(BPF_PATH)libbpf.a
> +BPF_PATH = $(objtree)/tools/lib/bpf
objtree won't be set for simple make in the directory. Perhaps we
should stick to using OUTPUT and srctree?
We should probably make a script with all the ways of calling make
should work. Otherwise we can lose track too easily.
# thru kbuild
make tools/bpf
T=$(mktemp -d)
make tools/bpf OUTPUT=$T
rm -rf $T
# from kernel source tree
make -C tools/bpf/bpftool
T=$(mktemp -d)
make -C tools/bpf/bpftool OUTPUT=$T
rm -rf $T
# from tools
cd tools/
make bpf
T=$(mktemp -d)
make bpf OUTPUT=$T
rm -rf $T
# from bpftool's dir
cd bpf/bpftool
make
T=$(mktemp -d)
make OUTPUT=$T
rm -rf $T
.. add your own.
> +LIBBPF = $(BPF_PATH)/libbpf.a
>
> BPFTOOL_VERSION := $(shell make --no-print-directory -sC ../../.. kernelversion)
>
> $(LIBBPF): FORCE
> - $(Q)$(MAKE) -C $(BPF_DIR) OUTPUT=$(OUTPUT) $(OUTPUT)libbpf.a
> + $(Q)mkdir -p $(BPF_PATH)
> + $(Q)$(MAKE) $(COMMAND_O) subdir=tools/lib/bpf -C $(BPF_DIR) $(LIBBPF)
>
> $(LIBBPF)-clean:
> $(call QUIET_CLEAN, libbpf)
> - $(Q)$(MAKE) -C $(BPF_DIR) OUTPUT=$(OUTPUT) clean >/dev/null
> + $(Q)$(MAKE) $(COMMAND_O) subdir=tools/lib/bpf -C $(BPF_DIR) clean >/dev/null
>
> prefix ?= /usr/local
> bash_compdir ?= /usr/share/bash-completion/completions
> @@ -112,7 +107,7 @@ $(OUTPUT)disasm.o: $(srctree)/kernel/bpf/disasm.c
> $(QUIET_CC)$(COMPILE.c) -MMD -o $@ $<
>
> $(OUTPUT)bpftool: $(OBJS) $(LIBBPF)
> - $(QUIET_LINK)$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $^ $(LIBS)
> + $(QUIET_LINK)$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJS) $(LIBS)
>
> $(OUTPUT)%.o: %.c
> $(QUIET_CC)$(COMPILE.c) -MMD -o $@ $<
^ permalink raw reply
* Re: [PATCH] net: bcmgenet: use promisc for unsupported filters
From: David Miller @ 2019-07-18 18:46 UTC (permalink / raw)
To: f.fainelli
Cc: justinpopo6, netdev, linux-kernel, bcm-kernel-feedback-list,
opendmb
In-Reply-To: <5f6422b5-e839-1600-6749-048a7e31ea96@gmail.com>
From: Florian Fainelli <f.fainelli@gmail.com>
Date: Wed, 17 Jul 2019 20:38:47 -0700
>
>
> On 7/17/2019 2:58 PM, justinpopo6@gmail.com wrote:
>> From: Justin Chen <justinpopo6@gmail.com>
>>
>> Currently we silently ignore filters if we cannot meet the filter
>> requirements. This will lead to the MAC dropping packets that are
>> expected to pass. A better solution would be to set the NIC to promisc
>> mode when the required filters cannot be met.
>>
>> Also correct the number of MDF filters supported. It should be 17,
>> not 16.
>>
>> Signed-off-by: Justin Chen <justinpopo6@gmail.com>
>
> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Applied and queued up for -stable.
^ permalink raw reply
* Re: [PATCH] Signed-off-by: Peter Kosyh <p.kosyh@gmail.com>
From: David Miller @ 2019-07-18 18:44 UTC (permalink / raw)
To: dsa; +Cc: p.kosyh, shrijeet, netdev, linux-kernel
In-Reply-To: <213bada2-fe81-3c14-1506-11abf0f3ca22@cumulusnetworks.com>
From: David Ahern <dsa@cumulusnetworks.com>
Date: Thu, 18 Jul 2019 08:02:45 -0600
> your subject line needs a proper Subject - a one-line summary of the
> change starting with 'vrf:'. See examples from 'git log drivers/net/vrf.c'
Indeed, you really need to fix this even for your second submission as it
had the same exact problem.
^ permalink raw reply
* Re: [PATCH v2] net/mlx5: Replace kfree with kvfree
From: Saeed Mahameed @ 2019-07-18 18:38 UTC (permalink / raw)
To: hslester96@gmail.com
Cc: linux-rdma@vger.kernel.org, davem@davemloft.net,
netdev@vger.kernel.org, leon@kernel.org,
linux-kernel@vger.kernel.org
In-Reply-To: <20190717101456.17401-1-hslester96@gmail.com>
On Wed, 2019-07-17 at 18:14 +0800, Chuhong Yuan wrote:
> Variable allocated by kvmalloc should not be freed by kfree.
> Because it may be allocated by vmalloc.
> So replace kfree with kvfree here.
>
> Fixes: 9b1f298236057 ("net/mlx5: Add support for FW fatal reporter
> dump")
> Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
Dave, i guess this can go to net.
Thanks,
Saeed.
> ---
> Changes in v2:
> - Add corresponding Fixes tag
>
> drivers/net/ethernet/mellanox/mlx5/core/health.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c
> b/drivers/net/ethernet/mellanox/mlx5/core/health.c
> index 2fe6923f7ce0..9314777d99e3 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
> @@ -597,7 +597,7 @@ mlx5_fw_fatal_reporter_dump(struct
> devlink_health_reporter *reporter,
> err = devlink_fmsg_arr_pair_nest_end(fmsg);
>
> free_data:
> - kfree(cr_data);
> + kvfree(cr_data);
> return err;
> }
>
^ permalink raw reply
* Re: [PATCH net-next] netfilter: nft_meta: Fix build error
From: Pablo Neira Ayuso @ 2019-07-18 18:30 UTC (permalink / raw)
To: YueHaibing
Cc: davem, kadlec, fw, roopa, nikolay, wenxu, linux-kernel, netdev,
bridge, coreteam, netfilter-devel
In-Reply-To: <20190709070126.29972-1-yuehaibing@huawei.com>
On Tue, Jul 09, 2019 at 03:01:26PM +0800, YueHaibing wrote:
> If NFT_BRIDGE_META is y and NF_TABLES is m, building fails:
>
> net/bridge/netfilter/nft_meta_bridge.o: In function `nft_meta_bridge_get_init':
> nft_meta_bridge.c:(.text+0xd0): undefined reference to `nft_parse_register'
> nft_meta_bridge.c:(.text+0xec): undefined reference to `nft_validate_register_store'
I took this one from Arnd instead:
https://patchwork.ozlabs.org/patch/1130262/
Thanks.
^ permalink raw reply
* Re: [EXT] [PATCH] bnx2x: Prevent load reordering in tx completion processing
From: Brian King @ 2019-07-18 18:26 UTC (permalink / raw)
To: Manish Chopra, GR-everest-linux-l2
Cc: Sudarsana Reddy Kalluru, Ariel Elior, netdev@vger.kernel.org
In-Reply-To: <DM6PR18MB2697C972B49EAD3AE7C3F37AABC80@DM6PR18MB2697.namprd18.prod.outlook.com>
On 7/18/19 5:12 AM, Manish Chopra wrote:
>> -----Original Message-----
>> From: Brian King <brking@linux.vnet.ibm.com>
>> Sent: Tuesday, July 16, 2019 3:12 AM
>> To: GR-everest-linux-l2 <GR-everest-linux-l2@marvell.com>
>> Cc: Sudarsana Reddy Kalluru <skalluru@marvell.com>; Ariel Elior
>> <aelior@marvell.com>; netdev@vger.kernel.org; Brian King
>> <brking@linux.vnet.ibm.com>
>> Subject: [EXT] [PATCH] bnx2x: Prevent load reordering in tx completion
>> processing
>>
>> External Email
>>
>> ----------------------------------------------------------------------
>> This patch fixes an issue seen on Power systems with bnx2x which results in
>> the skb is NULL WARN_ON in bnx2x_free_tx_pkt firing due to the skb pointer
>> getting loaded in bnx2x_free_tx_pkt prior to the hw_cons load in
>> bnx2x_tx_int. Adding a read memory barrier resolves the issue.
>>
>> Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
>> ---
>> drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 3 +++
>> 1 file changed, 3 insertions(+)
>>
>> diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
>> b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
>> index 656ed80..e2be5a6 100644
>> --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
>> +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
>> @@ -285,6 +285,9 @@ int bnx2x_tx_int(struct bnx2x *bp, struct
>> bnx2x_fp_txdata *txdata)
>> hw_cons = le16_to_cpu(*txdata->tx_cons_sb);
>> sw_cons = txdata->tx_pkt_cons;
>>
>> + /* Ensure subsequent loads occur after hw_cons */
>> + smp_rmb();
>> +
>> while (sw_cons != hw_cons) {
>> u16 pkt_cons;
>>
>> --
>> 1.8.3.1
>
> Could you please explain a bit in detail what could have caused skb to NULL exactly ?
> Curious that if skb would have been NULL for some reason it did not cause NULL pointer dereference in bnx2x_free_tx_pkt() on below call -
>
> prefetch(&skb->end);
>
> Which is prior to the said WARN_ON(!skb) in bnx2x_free_tx_pkt().
Right. In this case, that would end up passing an invalid address to prefetch. On a
Power processor, that turns into a dcbt instruction (data cache block touch), which
is a hint to the process that the subsequent code may access that data cache block.
Passing an invalid address to dcbt causes no harm. I just built a userspace program
to validate that doing something very similar to what is happening here, and the dcbt
executed with no errors, no segfault to the userspace process.
This is the scenario I think is occurring.
CPU[0]
bnx2x_start_xmit
[1] tx_buf->skb = skb; /* store skb pointer */
[2] ...
[3] wmb();
[4] DOORBELL_RELAXED
CPU[1]
bnx2x_tx_int
[5] hw_cons = le16_to_cpu(*txdata->tx_cons_sb);
[6] sw_cons = txdata->tx_pkt_cons;
[7] while (sw_cons != hw_cons) {
[8] pkt_cons = TX_BD(sw_cons);
[9] bnx2x_free_tx_pkt
[10] tx_buf = &txdata->tx_buf_ring[pkt_cons];
[11] skb = tx_buf->skb;
[12] prefetch(&skb->end);
[13] ...
[14] WARN_ON(!skb);
On CPU0 we are in the process of sending a TX buffer to the adapter. We have a wmb at [3]
to ensure that all stores to cacheable storage are coherent with respect to the
non-cacheable store at [4] to tell the adapter about the new TX buffer.
On CPU1, we have a potential race condition if the processor is aggressively reordering
loads. If we find ourselves in bnx2x_tx_int, while still in bn2x_start_xmit on CPU0,
its possible the processor could begin speculatively executing well into [11]. Since there
is no read barrier of any sort to tell the processor that the load of the skb pointer
cannot happen until the load of hw_cons has occurred, we could end up speculatively
loading the skb pointer before the write in [1] has completed with respect to CPU1.
Adding the smp_rmb between 6 and 7 ensures that the load at 5 occurs prior to the
load at 11.
This was reproducing consistently on a 16 socket Power 9 system built of four, four socket
nodes, with 10 bnx2x adapters and 26TB of memory. The NUMA effects can get to be exaggerated
a bit on these systems. With this patch, the system runs without issues. We've now been able to
run the workload multiple times with no issues.
Thanks,
Brian
--
Brian King
Power Linux I/O
IBM Linux Technology Center
^ permalink raw reply
* Re: [PATCH v2] net/netfilter: remove unnecessary spaces
From: Pablo Neira Ayuso @ 2019-07-18 18:26 UTC (permalink / raw)
To: yangxingwu
Cc: wensong, horms, ja, kadlec, fw, davem, netdev, lvs-devel,
netfilter-devel, coreteam, linux-kernel, joe
In-Reply-To: <20190716021301.27753-1-xingwu.yang@gmail.com>
Looks good, but you will have to wait until net-next reopens:
http://vger.kernel.org/~davem/net-next.html
Will keep this in my patchwork until that happens.
Thanks.
On Tue, Jul 16, 2019 at 10:13:01AM +0800, yangxingwu wrote:
> this patch removes extra spaces
>
> Signed-off-by: yangxingwu <xingwu.yang@gmail.com>
> ---
> net/netfilter/ipset/ip_set_hash_gen.h | 2 +-
> net/netfilter/ipset/ip_set_list_set.c | 2 +-
> net/netfilter/ipvs/ip_vs_core.c | 2 +-
> net/netfilter/ipvs/ip_vs_mh.c | 4 ++--
> net/netfilter/ipvs/ip_vs_proto_tcp.c | 2 +-
> net/netfilter/nf_conntrack_ftp.c | 2 +-
> net/netfilter/nf_conntrack_proto_tcp.c | 2 +-
> net/netfilter/nfnetlink_log.c | 4 ++--
> net/netfilter/nfnetlink_queue.c | 4 ++--
> net/netfilter/xt_IDLETIMER.c | 2 +-
> 10 files changed, 13 insertions(+), 13 deletions(-)
>
> diff --git a/net/netfilter/ipset/ip_set_hash_gen.h b/net/netfilter/ipset/ip_set_hash_gen.h
> index 10f6196..eb907d2 100644
> --- a/net/netfilter/ipset/ip_set_hash_gen.h
> +++ b/net/netfilter/ipset/ip_set_hash_gen.h
> @@ -954,7 +954,7 @@ struct htype {
> mtype_data_netmask(d, NCIDR_GET(h->nets[j].cidr[0]));
> #endif
> key = HKEY(d, h->initval, t->htable_bits);
> - n = rcu_dereference_bh(hbucket(t, key));
> + n = rcu_dereference_bh(hbucket(t, key));
> if (!n)
> continue;
> for (i = 0; i < n->pos; i++) {
> diff --git a/net/netfilter/ipset/ip_set_list_set.c b/net/netfilter/ipset/ip_set_list_set.c
> index 8ada318..5c2be76 100644
> --- a/net/netfilter/ipset/ip_set_list_set.c
> +++ b/net/netfilter/ipset/ip_set_list_set.c
> @@ -289,7 +289,7 @@ struct list_set {
> if (n &&
> !(SET_WITH_TIMEOUT(set) &&
> ip_set_timeout_expired(ext_timeout(n, set))))
> - n = NULL;
> + n = NULL;
>
> e = kzalloc(set->dsize, GFP_ATOMIC);
> if (!e)
> diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
> index 7138556..6b3ae76 100644
> --- a/net/netfilter/ipvs/ip_vs_core.c
> +++ b/net/netfilter/ipvs/ip_vs_core.c
> @@ -615,7 +615,7 @@ int ip_vs_leave(struct ip_vs_service *svc, struct sk_buff *skb,
> unsigned int flags = (svc->flags & IP_VS_SVC_F_ONEPACKET &&
> iph->protocol == IPPROTO_UDP) ?
> IP_VS_CONN_F_ONE_PACKET : 0;
> - union nf_inet_addr daddr = { .all = { 0, 0, 0, 0 } };
> + union nf_inet_addr daddr = { .all = { 0, 0, 0, 0 } };
>
> /* create a new connection entry */
> IP_VS_DBG(6, "%s(): create a cache_bypass entry\n", __func__);
> diff --git a/net/netfilter/ipvs/ip_vs_mh.c b/net/netfilter/ipvs/ip_vs_mh.c
> index 94d9d34..da0280c 100644
> --- a/net/netfilter/ipvs/ip_vs_mh.c
> +++ b/net/netfilter/ipvs/ip_vs_mh.c
> @@ -174,8 +174,8 @@ static int ip_vs_mh_populate(struct ip_vs_mh_state *s,
> return 0;
> }
>
> - table = kcalloc(BITS_TO_LONGS(IP_VS_MH_TAB_SIZE),
> - sizeof(unsigned long), GFP_KERNEL);
> + table = kcalloc(BITS_TO_LONGS(IP_VS_MH_TAB_SIZE),
> + sizeof(unsigned long), GFP_KERNEL);
> if (!table)
> return -ENOMEM;
>
> diff --git a/net/netfilter/ipvs/ip_vs_proto_tcp.c b/net/netfilter/ipvs/ip_vs_proto_tcp.c
> index 915ac82..c7b46a9 100644
> --- a/net/netfilter/ipvs/ip_vs_proto_tcp.c
> +++ b/net/netfilter/ipvs/ip_vs_proto_tcp.c
> @@ -710,7 +710,7 @@ static int __ip_vs_tcp_init(struct netns_ipvs *ipvs, struct ip_vs_proto_data *pd
> sizeof(tcp_timeouts));
> if (!pd->timeout_table)
> return -ENOMEM;
> - pd->tcp_state_table = tcp_states;
> + pd->tcp_state_table = tcp_states;
> return 0;
> }
>
> diff --git a/net/netfilter/nf_conntrack_ftp.c b/net/netfilter/nf_conntrack_ftp.c
> index 8c6c11b..26c1ff8 100644
> --- a/net/netfilter/nf_conntrack_ftp.c
> +++ b/net/netfilter/nf_conntrack_ftp.c
> @@ -162,7 +162,7 @@ static int try_rfc959(const char *data, size_t dlen,
> if (length == 0)
> return 0;
>
> - cmd->u3.ip = htonl((array[0] << 24) | (array[1] << 16) |
> + cmd->u3.ip = htonl((array[0] << 24) | (array[1] << 16) |
> (array[2] << 8) | array[3]);
> cmd->u.tcp.port = htons((array[4] << 8) | array[5]);
> return length;
> diff --git a/net/netfilter/nf_conntrack_proto_tcp.c b/net/netfilter/nf_conntrack_proto_tcp.c
> index 1e2cc83..48f3a67 100644
> --- a/net/netfilter/nf_conntrack_proto_tcp.c
> +++ b/net/netfilter/nf_conntrack_proto_tcp.c
> @@ -1225,7 +1225,7 @@ static int tcp_to_nlattr(struct sk_buff *skb, struct nlattr *nla,
> [CTA_PROTOINFO_TCP_WSCALE_ORIGINAL] = { .type = NLA_U8 },
> [CTA_PROTOINFO_TCP_WSCALE_REPLY] = { .type = NLA_U8 },
> [CTA_PROTOINFO_TCP_FLAGS_ORIGINAL] = { .len = sizeof(struct nf_ct_tcp_flags) },
> - [CTA_PROTOINFO_TCP_FLAGS_REPLY] = { .len = sizeof(struct nf_ct_tcp_flags) },
> + [CTA_PROTOINFO_TCP_FLAGS_REPLY] = { .len = sizeof(struct nf_ct_tcp_flags) },
> };
>
> #define TCP_NLATTR_SIZE ( \
> diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
> index 6dee4f9..d69e186 100644
> --- a/net/netfilter/nfnetlink_log.c
> +++ b/net/netfilter/nfnetlink_log.c
> @@ -651,7 +651,7 @@ static void nfulnl_instance_free_rcu(struct rcu_head *head)
> /* FIXME: do we want to make the size calculation conditional based on
> * what is actually present? way more branches and checks, but more
> * memory efficient... */
> - size = nlmsg_total_size(sizeof(struct nfgenmsg))
> + size = nlmsg_total_size(sizeof(struct nfgenmsg))
> + nla_total_size(sizeof(struct nfulnl_msg_packet_hdr))
> + nla_total_size(sizeof(u_int32_t)) /* ifindex */
> + nla_total_size(sizeof(u_int32_t)) /* ifindex */
> @@ -668,7 +668,7 @@ static void nfulnl_instance_free_rcu(struct rcu_head *head)
> + nla_total_size(sizeof(struct nfgenmsg)); /* NLMSG_DONE */
>
> if (in && skb_mac_header_was_set(skb)) {
> - size += nla_total_size(skb->dev->hard_header_len)
> + size += nla_total_size(skb->dev->hard_header_len)
> + nla_total_size(sizeof(u_int16_t)) /* hwtype */
> + nla_total_size(sizeof(u_int16_t)); /* hwlen */
> }
> diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
> index 89750f7..a1ef6e3 100644
> --- a/net/netfilter/nfnetlink_queue.c
> +++ b/net/netfilter/nfnetlink_queue.c
> @@ -394,7 +394,7 @@ static int nfqnl_put_bridge(struct nf_queue_entry *entry, struct sk_buff *skb)
> char *secdata = NULL;
> u32 seclen = 0;
>
> - size = nlmsg_total_size(sizeof(struct nfgenmsg))
> + size = nlmsg_total_size(sizeof(struct nfgenmsg))
> + nla_total_size(sizeof(struct nfqnl_msg_packet_hdr))
> + nla_total_size(sizeof(u_int32_t)) /* ifindex */
> + nla_total_size(sizeof(u_int32_t)) /* ifindex */
> @@ -453,7 +453,7 @@ static int nfqnl_put_bridge(struct nf_queue_entry *entry, struct sk_buff *skb)
> }
>
> if (queue->flags & NFQA_CFG_F_UID_GID) {
> - size += (nla_total_size(sizeof(u_int32_t)) /* uid */
> + size += (nla_total_size(sizeof(u_int32_t)) /* uid */
> + nla_total_size(sizeof(u_int32_t))); /* gid */
> }
>
> diff --git a/net/netfilter/xt_IDLETIMER.c b/net/netfilter/xt_IDLETIMER.c
> index 9cec9ea..f56d3ed 100644
> --- a/net/netfilter/xt_IDLETIMER.c
> +++ b/net/netfilter/xt_IDLETIMER.c
> @@ -283,7 +283,7 @@ static int __init idletimer_tg_init(void)
>
> idletimer_tg_kobj = &idletimer_tg_device->kobj;
>
> - err = xt_register_target(&idletimer_tg);
> + err = xt_register_target(&idletimer_tg);
> if (err < 0) {
> pr_debug("couldn't register xt target\n");
> goto out_dev;
> --
> 1.8.3.1
>
^ permalink raw reply
* Re: [PATCH] [net-next] netfilter: bridge: make NF_TABLES_BRIDGE tristate
From: Pablo Neira Ayuso @ 2019-07-18 18:24 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Jozsef Kadlecsik, Florian Westphal, Roopa Prabhu,
Nikolay Aleksandrov, David S. Miller, wenxu, netfilter-devel,
coreteam, bridge, netdev, linux-kernel
In-Reply-To: <20190710080835.296696-1-arnd@arndb.de>
On Wed, Jul 10, 2019 at 10:08:20AM +0200, Arnd Bergmann wrote:
> The new nft_meta_bridge code fails to link as built-in when NF_TABLES
> is a loadable module.
Applied, thanks Arnd.
^ permalink raw reply
* Re: [PATCH bpf] libbpf: fix missing __WORDSIZE definition
From: Andrii Nakryiko @ 2019-07-18 18:04 UTC (permalink / raw)
To: Arnaldo Carvalho de Melo
Cc: Andrii Nakryiko, bpf, Networking, Daniel Borkmann,
Alexei Starovoitov, Kernel Team, Jiri Olsa, Namhyung Kim,
Arnaldo Carvalho de Melo
In-Reply-To: <20190718175533.GG2093@redhat.com>
On Thu, Jul 18, 2019 at 10:55 AM Arnaldo Carvalho de Melo
<arnaldo.melo@gmail.com> wrote:
>
> Em Thu, Jul 18, 2019 at 10:25:13AM -0700, Andrii Nakryiko escreveu:
> > hashmap.h depends on __WORDSIZE being defined. It is defined by
> > glibc/musl in different headers. It's an explicit goal for musl to be
> > "non-detectable" at compilation time, so instead include glibc header if
> > glibc is explicitly detected and fall back to musl header otherwise.
> >
> > Fixes: e3b924224028 ("libbpf: add resizable non-thread safe internal hashmap")
> > Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> > Signed-off-by: Andrii Nakryiko <andriin@fb.com>
>
> I fixed this here differently, as below, I didn't send it because I'm still
> testing it, so far, with a few other fixes and cherry-picking "libbpf: fix ptr
> to u64 conversion warning on 32-bit platforms" that is still in the bpf tree
> and is needed for the cross build containers in my suite that are 32-bit, I
> have the results below, this builds perf + libbpf (where elfutils is available,
> which is in most cases, except the uCLibc containers due to missing argp-devel),
> with gcc and with clang:
>
> [perfbuilder@quaco linux-perf-tools-build]$ export PERF_TARBALL=http://192.168.124.1/perf/perf-5.2.0.tar.xz
> [perfbuilder@quaco linux-perf-tools-build]$ time dm
> 1 alpine:3.4 : Ok gcc (Alpine 5.3.0) 5.3.0, clang version 3.8.0 (tags/RELEASE_380/final)
> 2 alpine:3.5 : Ok gcc (Alpine 6.2.1) 6.2.1 20160822, clang version 3.8.1 (tags/RELEASE_381/final)
> 3 alpine:3.6 : Ok gcc (Alpine 6.3.0) 6.3.0, clang version 4.0.0 (tags/RELEASE_400/final)
> 4 alpine:3.7 : Ok gcc (Alpine 6.4.0) 6.4.0, Alpine clang version 5.0.0 (tags/RELEASE_500/final) (based on LLVM 5.0.0)
> 5 alpine:3.8 : Ok gcc (Alpine 6.4.0) 6.4.0, Alpine clang version 5.0.1 (tags/RELEASE_501/final) (based on LLVM 5.0.1)
> 6 alpine:3.9 : Ok gcc (Alpine 8.3.0) 8.3.0, Alpine clang version 5.0.1 (tags/RELEASE_502/final) (based on LLVM 5.0.1)
> 7 alpine:3.10 : Ok gcc (Alpine 8.3.0) 8.3.0, Alpine clang version 8.0.0 (tags/RELEASE_800/final) (based on LLVM 8.0.0)
> 8 alpine:edge : Ok gcc (Alpine 8.3.0) 8.3.0, Alpine clang version 7.0.1 (tags/RELEASE_701/final) (based on LLVM 7.0.1)
> 9 amazonlinux:1 : Ok gcc (GCC) 7.2.1 20170915 (Red Hat 7.2.1-2), clang version 3.6.2 (tags/RELEASE_362/final)
> 10 amazonlinux:2 : Ok gcc (GCC) 7.3.1 20180303 (Red Hat 7.3.1-5), clang version 7.0.1 (Amazon Linux 2 7.0.1-1.amzn2.0.2)
> 11 android-ndk:r12b-arm : Ok arm-linux-androideabi-gcc (GCC) 4.9.x 20150123 (prerelease)
> 12 android-ndk:r15c-arm : Ok arm-linux-androideabi-gcc (GCC) 4.9.x 20150123 (prerelease)
> 13 centos:5 : Ok gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-55)
> 14 centos:6 : Ok gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-23)
> 15 centos:7 : Ok gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-36), clang version 3.4.2 (tags/RELEASE_34/dot2-final)
> 16 clearlinux:latest : Ok gcc (Clear Linux OS for Intel Architecture) 9.1.1 20190628 gcc-9-branch@272773, clang version 8.0.0 (tags/RELEASE_800/final)
> 17 debian:8 : Ok gcc (Debian 4.9.2-10+deb8u2) 4.9.2, Debian clang version 3.5.0-10 (tags/RELEASE_350/final) (based on LLVM 3.5.0)
> 18 debian:9 : Ok gcc (Debian 6.3.0-18+deb9u1) 6.3.0 20170516, clang version 3.8.1-24 (tags/RELEASE_381/final)
> 19 debian:10 : Ok gcc (Debian 8.3.0-6) 8.3.0, clang version 7.0.1-8 (tags/RELEASE_701/final)
> 20 debian:experimental : Ok gcc (Debian 8.3.0-19) 8.3.0, clang version 7.0.1-8 (tags/RELEASE_701/final)
> 21 debian:experimental-x-arm64 : Ok aarch64-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0
> 22 debian:experimental-x-mips : Ok mips-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0
> 23 debian:experimental-x-mips64 : Ok mips64-linux-gnuabi64-gcc (Debian 8.3.0-7) 8.3.0
> 24 debian:experimental-x-mipsel : Ok mipsel-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0
> 25 fedora:20 : Ok gcc (GCC) 4.8.3 20140911 (Red Hat 4.8.3-7), clang version 3.4.2 (tags/RELEASE_34/dot2-final)
> 26 fedora:22 : Ok gcc (GCC) 5.3.1 20160406 (Red Hat 5.3.1-6), clang version 3.5.0 (tags/RELEASE_350/final)
> 27 fedora:23 : Ok gcc (GCC) 5.3.1 20160406 (Red Hat 5.3.1-6), clang version 3.7.0 (tags/RELEASE_370/final)
> 28 fedora:24 : Ok gcc (GCC) 6.3.1 20161221 (Red Hat 6.3.1-1), clang version 3.8.1 (tags/RELEASE_381/final)
> 29 fedora:24-x-ARC-uClibc : Ok arc-linux-gcc (ARCompact ISA Linux uClibc toolchain 2017.09-rc2) 7.1.1 20170710
>
> I've pushed it to a tmp.perf/core branch in my
> git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git tree, that has
> these:
>
> d5e1f2d60d41 (HEAD -> perf/core, acme.korg/tmp.perf/core) libbpf: fix ptr to u64 conversion warning on 32-bit platforms
> 7c08fd16f917 tools lib bpf: Avoid designated initializers for unnamed union members
> 4c9f83c95ad6 tools lib bpf: Avoid using 'link' as it shadows a global definition in some systems
> bdb07df4a0ad tools lib bpf: Fix endianness macro usage for some compilers
> 66dbf3caff52 tools lib bpf: Replace __WORDSIZE with BITS_PER_LONG to build on the musl libc
>
> Please take a look and check if everything is fine on your side. The HEAD I'll
> remove if Daniel thinks it should wait that landing via the BPF tree, I just put it
> there for the test builds.
>
> commit 66dbf3caff52be0d004bcb9ac4cea4c19eb75dfc
> Author: Arnaldo Carvalho de Melo <acme@redhat.com>
> Date: Thu Jul 18 09:46:28 2019 -0300
>
> tools lib bpf: Replace __WORDSIZE with BITS_PER_LONG to build on the musl libc
>
> BITS_PER_LONG is more generally available and equivalent to __WORDSIZE,
> so use it instead to keep it building in systems using the mustl libc
> where __WORDSIZE is in a different place than in glibc.
>
> And do this by explicitely adding the header where this definition is
> (asm/bitsperlong.h) instead of getting it indirectly.
>
> Cc: Adrian Hunter <adrian.hunter@intel.com>
> Cc: Alexei Starovoitov <ast@kernel.org>
> Cc: Andrii Nakryiko <andriin@fb.com>
> Cc: Jiri Olsa <jolsa@kernel.org>
> Cc: Namhyung Kim <namhyung@kernel.org>
> Fixes: e3b924224028 ("libbpf: add resizable non-thread safe internal hashmap")
> Link: https://lkml.kernel.org/n/tip-61vydgldzmmz5w2mf6rv3ryl@git.kernel.org
> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
>
> diff --git a/tools/lib/bpf/hashmap.h b/tools/lib/bpf/hashmap.h
> index 03748a742146..f1f37b574d9c 100644
> --- a/tools/lib/bpf/hashmap.h
> +++ b/tools/lib/bpf/hashmap.h
> @@ -10,12 +10,13 @@
>
> #include <stdbool.h>
> #include <stddef.h>
> +#include <asm/bitsperlong.h>
Relying on this header is problematic when syncing libbpf into Github
mirror. There we'll need to re-implement it anyway, and again,
probably through __WORDSIZE or some other tricks. So if we can do away
without kernel specific header that would be great.
> #include "libbpf_internal.h"
>
> static inline size_t hash_bits(size_t h, int bits)
> {
> /* shuffle bits and return requested number of upper bits */
> - return (h * 11400714819323198485llu) >> (__WORDSIZE - bits);
> + return (h * 11400714819323198485llu) >> (BITS_PER_LONG - bits);
> }
>
> typedef size_t (*hashmap_hash_fn)(const void *key, void *ctx);
^ permalink raw reply
* [PATCH net,v4 3/4] net: flow_offload: rename tc_setup_cb_t to flow_setup_cb_t
From: Pablo Neira Ayuso @ 2019-07-18 17:59 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, jiri, jakub.kicinski, pshelar
In-Reply-To: <20190718175931.13529-1-pablo@netfilter.org>
Rename this type definition and adapt users.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
v4: no changes.
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 2 +-
drivers/net/ethernet/mscc/ocelot_tc.c | 2 +-
include/net/flow_offload.h | 16 ++++++++++------
include/net/pkt_cls.h | 5 ++---
include/net/sch_generic.h | 6 ++----
net/core/flow_offload.c | 9 +++++----
net/dsa/slave.c | 2 +-
net/sched/cls_api.c | 2 +-
net/sched/cls_bpf.c | 2 +-
net/sched/cls_flower.c | 2 +-
net/sched/cls_matchall.c | 2 +-
net/sched/cls_u32.c | 6 +++---
12 files changed, 29 insertions(+), 27 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index a469035400cf..51cd0b6f1f3e 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -1679,7 +1679,7 @@ static int mlxsw_sp_setup_tc_block(struct mlxsw_sp_port *mlxsw_sp_port,
struct flow_block_offload *f)
{
struct flow_block_cb *block_cb;
- tc_setup_cb_t *cb;
+ flow_setup_cb_t *cb;
bool ingress;
int err;
diff --git a/drivers/net/ethernet/mscc/ocelot_tc.c b/drivers/net/ethernet/mscc/ocelot_tc.c
index abbcb66bf5ac..fba9512e9ca6 100644
--- a/drivers/net/ethernet/mscc/ocelot_tc.c
+++ b/drivers/net/ethernet/mscc/ocelot_tc.c
@@ -134,7 +134,7 @@ static int ocelot_setup_tc_block(struct ocelot_port *port,
struct flow_block_offload *f)
{
struct flow_block_cb *block_cb;
- tc_setup_cb_t *cb;
+ flow_setup_cb_t *cb;
int err;
netdev_dbg(port->dev, "tc_block command %d, binder_type %d\n",
diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index aa9b5287b231..98bf3af5c84d 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h
@@ -3,7 +3,6 @@
#include <linux/kernel.h>
#include <net/flow_dissector.h>
-#include <net/sch_generic.h>
struct flow_match {
struct flow_dissector *dissector;
@@ -261,23 +260,27 @@ struct flow_block_offload {
struct netlink_ext_ack *extack;
};
+enum tc_setup_type;
+typedef int flow_setup_cb_t(enum tc_setup_type type, void *type_data,
+ void *cb_priv);
+
struct flow_block_cb {
struct list_head driver_list;
struct list_head list;
- tc_setup_cb_t *cb;
+ flow_setup_cb_t *cb;
void *cb_ident;
void *cb_priv;
void (*release)(void *cb_priv);
unsigned int refcnt;
};
-struct flow_block_cb *flow_block_cb_alloc(tc_setup_cb_t *cb,
+struct flow_block_cb *flow_block_cb_alloc(flow_setup_cb_t *cb,
void *cb_ident, void *cb_priv,
void (*release)(void *cb_priv));
void flow_block_cb_free(struct flow_block_cb *block_cb);
struct flow_block_cb *flow_block_cb_lookup(struct flow_block_offload *offload,
- tc_setup_cb_t *cb, void *cb_ident);
+ flow_setup_cb_t *cb, void *cb_ident);
void *flow_block_cb_priv(struct flow_block_cb *block_cb);
void flow_block_cb_incref(struct flow_block_cb *block_cb);
@@ -295,11 +298,12 @@ static inline void flow_block_cb_remove(struct flow_block_cb *block_cb,
list_move(&block_cb->list, &offload->cb_list);
}
-bool flow_block_cb_is_busy(tc_setup_cb_t *cb, void *cb_ident,
+bool flow_block_cb_is_busy(flow_setup_cb_t *cb, void *cb_ident,
struct list_head *driver_block_list);
int flow_block_cb_setup_simple(struct flow_block_offload *f,
- struct list_head *driver_list, tc_setup_cb_t *cb,
+ struct list_head *driver_list,
+ flow_setup_cb_t *cb,
void *cb_ident, void *cb_priv, bool ingress_only);
enum flow_cls_command {
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index 841faadceb6e..e429809ca90d 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -6,7 +6,6 @@
#include <linux/workqueue.h>
#include <net/sch_generic.h>
#include <net/act_api.h>
-#include <net/flow_offload.h>
#include <net/net_namespace.h>
/* TC action not accessible from user space */
@@ -126,14 +125,14 @@ static inline struct Qdisc *tcf_block_q(struct tcf_block *block)
}
static inline
-int tc_setup_cb_block_register(struct tcf_block *block, tc_setup_cb_t *cb,
+int tc_setup_cb_block_register(struct tcf_block *block, flow_setup_cb_t *cb,
void *cb_priv)
{
return 0;
}
static inline
-void tc_setup_cb_block_unregister(struct tcf_block *block, tc_setup_cb_t *cb,
+void tc_setup_cb_block_unregister(struct tcf_block *block, flow_setup_cb_t *cb,
void *cb_priv)
{
}
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index 855167bbc372..9482e060483b 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -15,6 +15,7 @@
#include <linux/mutex.h>
#include <net/gen_stats.h>
#include <net/rtnetlink.h>
+#include <net/flow_offload.h>
struct Qdisc_ops;
struct qdisc_walker;
@@ -22,9 +23,6 @@ struct tcf_walker;
struct module;
struct bpf_flow_keys;
-typedef int tc_setup_cb_t(enum tc_setup_type type,
- void *type_data, void *cb_priv);
-
typedef int tc_indr_block_bind_cb_t(struct net_device *dev, void *cb_priv,
enum tc_setup_type type, void *type_data);
@@ -313,7 +311,7 @@ struct tcf_proto_ops {
void (*walk)(struct tcf_proto *tp,
struct tcf_walker *arg, bool rtnl_held);
int (*reoffload)(struct tcf_proto *tp, bool add,
- tc_setup_cb_t *cb, void *cb_priv,
+ flow_setup_cb_t *cb, void *cb_priv,
struct netlink_ext_ack *extack);
void (*bind_class)(void *, u32, unsigned long);
void * (*tmplt_create)(struct net *net,
diff --git a/net/core/flow_offload.c b/net/core/flow_offload.c
index 507de4b48815..a800fa78d96c 100644
--- a/net/core/flow_offload.c
+++ b/net/core/flow_offload.c
@@ -165,7 +165,7 @@ void flow_rule_match_enc_opts(const struct flow_rule *rule,
}
EXPORT_SYMBOL(flow_rule_match_enc_opts);
-struct flow_block_cb *flow_block_cb_alloc(tc_setup_cb_t *cb,
+struct flow_block_cb *flow_block_cb_alloc(flow_setup_cb_t *cb,
void *cb_ident, void *cb_priv,
void (*release)(void *cb_priv))
{
@@ -194,7 +194,7 @@ void flow_block_cb_free(struct flow_block_cb *block_cb)
EXPORT_SYMBOL(flow_block_cb_free);
struct flow_block_cb *flow_block_cb_lookup(struct flow_block_offload *f,
- tc_setup_cb_t *cb, void *cb_ident)
+ flow_setup_cb_t *cb, void *cb_ident)
{
struct flow_block_cb *block_cb;
@@ -226,7 +226,7 @@ unsigned int flow_block_cb_decref(struct flow_block_cb *block_cb)
}
EXPORT_SYMBOL(flow_block_cb_decref);
-bool flow_block_cb_is_busy(tc_setup_cb_t *cb, void *cb_ident,
+bool flow_block_cb_is_busy(flow_setup_cb_t *cb, void *cb_ident,
struct list_head *driver_block_list)
{
struct flow_block_cb *block_cb;
@@ -243,7 +243,8 @@ EXPORT_SYMBOL(flow_block_cb_is_busy);
int flow_block_cb_setup_simple(struct flow_block_offload *f,
struct list_head *driver_block_list,
- tc_setup_cb_t *cb, void *cb_ident, void *cb_priv,
+ flow_setup_cb_t *cb,
+ void *cb_ident, void *cb_priv,
bool ingress_only)
{
struct flow_block_cb *block_cb;
diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index 6ca9ec58f881..d697a64fb564 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -951,7 +951,7 @@ static int dsa_slave_setup_tc_block(struct net_device *dev,
struct flow_block_offload *f)
{
struct flow_block_cb *block_cb;
- tc_setup_cb_t *cb;
+ flow_setup_cb_t *cb;
if (f->binder_type == FLOW_BLOCK_BINDER_TYPE_CLSACT_INGRESS)
cb = dsa_slave_setup_tc_block_cb_ig;
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index d144233423c5..78f0f2815b8c 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -1514,7 +1514,7 @@ void tcf_block_put(struct tcf_block *block)
EXPORT_SYMBOL(tcf_block_put);
static int
-tcf_block_playback_offloads(struct tcf_block *block, tc_setup_cb_t *cb,
+tcf_block_playback_offloads(struct tcf_block *block, flow_setup_cb_t *cb,
void *cb_priv, bool add, bool offload_in_use,
struct netlink_ext_ack *extack)
{
diff --git a/net/sched/cls_bpf.c b/net/sched/cls_bpf.c
index 691f71830134..3f7a9c02b70c 100644
--- a/net/sched/cls_bpf.c
+++ b/net/sched/cls_bpf.c
@@ -651,7 +651,7 @@ static void cls_bpf_walk(struct tcf_proto *tp, struct tcf_walker *arg,
}
}
-static int cls_bpf_reoffload(struct tcf_proto *tp, bool add, tc_setup_cb_t *cb,
+static int cls_bpf_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
void *cb_priv, struct netlink_ext_ack *extack)
{
struct cls_bpf_head *head = rtnl_dereference(tp->root);
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 38d6e85693fc..054123742e32 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -1800,7 +1800,7 @@ fl_get_next_hw_filter(struct tcf_proto *tp, struct cls_fl_filter *f, bool add)
return NULL;
}
-static int fl_reoffload(struct tcf_proto *tp, bool add, tc_setup_cb_t *cb,
+static int fl_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
void *cb_priv, struct netlink_ext_ack *extack)
{
struct tcf_block *block = tp->chain->block;
diff --git a/net/sched/cls_matchall.c b/net/sched/cls_matchall.c
index a30d2f8feb32..455ea2793f9b 100644
--- a/net/sched/cls_matchall.c
+++ b/net/sched/cls_matchall.c
@@ -282,7 +282,7 @@ static void mall_walk(struct tcf_proto *tp, struct tcf_walker *arg,
arg->count++;
}
-static int mall_reoffload(struct tcf_proto *tp, bool add, tc_setup_cb_t *cb,
+static int mall_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
void *cb_priv, struct netlink_ext_ack *extack)
{
struct cls_mall_head *head = rtnl_dereference(tp->root);
diff --git a/net/sched/cls_u32.c b/net/sched/cls_u32.c
index be9e46c77e8b..8614088edd1b 100644
--- a/net/sched/cls_u32.c
+++ b/net/sched/cls_u32.c
@@ -1152,7 +1152,7 @@ static void u32_walk(struct tcf_proto *tp, struct tcf_walker *arg,
}
static int u32_reoffload_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht,
- bool add, tc_setup_cb_t *cb, void *cb_priv,
+ bool add, flow_setup_cb_t *cb, void *cb_priv,
struct netlink_ext_ack *extack)
{
struct tc_cls_u32_offload cls_u32 = {};
@@ -1172,7 +1172,7 @@ static int u32_reoffload_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht,
}
static int u32_reoffload_knode(struct tcf_proto *tp, struct tc_u_knode *n,
- bool add, tc_setup_cb_t *cb, void *cb_priv,
+ bool add, flow_setup_cb_t *cb, void *cb_priv,
struct netlink_ext_ack *extack)
{
struct tc_u_hnode *ht = rtnl_dereference(n->ht_down);
@@ -1213,7 +1213,7 @@ static int u32_reoffload_knode(struct tcf_proto *tp, struct tc_u_knode *n,
return 0;
}
-static int u32_reoffload(struct tcf_proto *tp, bool add, tc_setup_cb_t *cb,
+static int u32_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
void *cb_priv, struct netlink_ext_ack *extack)
{
struct tc_u_common *tp_c = tp->data;
--
2.11.0
^ permalink raw reply related
* [PATCH net,v4 4/4] net: flow_offload: add flow_block structure and use it
From: Pablo Neira Ayuso @ 2019-07-18 17:59 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, jiri, jakub.kicinski, pshelar
In-Reply-To: <20190718175931.13529-1-pablo@netfilter.org>
This object stores the flow block callbacks that are attached to this
block. Update flow_block_cb_lookup() to take this new object.
This patch restores the block sharing feature.
Fixes: da3eeb904ff4 ("net: flow_offload: add list handling functions")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
v3: add flow_block_init() - Jiri Pirko.
rename flow/flow_block/ - Jiri Pirko.
fix documentation in nft_stats - Jiri Pirko.
update flow_block_cb_lookup() argument to take flow_block
and adapt callers - Jiri Pirko.
v4: no changes.
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 2 +-
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 8 +++++---
drivers/net/ethernet/mscc/ocelot_flower.c | 8 ++++----
drivers/net/ethernet/mscc/ocelot_tc.c | 2 +-
drivers/net/ethernet/netronome/nfp/flower/offload.c | 5 +++--
include/net/flow_offload.h | 12 +++++++++++-
include/net/netfilter/nf_tables.h | 5 +++--
include/net/sch_generic.h | 2 +-
net/core/flow_offload.c | 6 +++---
net/dsa/slave.c | 2 +-
net/netfilter/nf_tables_api.c | 2 +-
net/netfilter/nf_tables_offload.c | 5 +++--
net/sched/cls_api.c | 10 +++++++---
13 files changed, 44 insertions(+), 25 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 2162412073c5..7f747cb1a4f4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -752,7 +752,7 @@ mlx5e_rep_indr_setup_tc_block(struct net_device *netdev,
if (!indr_priv)
return -ENOENT;
- block_cb = flow_block_cb_lookup(f,
+ block_cb = flow_block_cb_lookup(f->block,
mlx5e_rep_indr_setup_block_cb,
indr_priv);
if (!block_cb)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 51cd0b6f1f3e..650638152bbc 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -1604,7 +1604,8 @@ mlxsw_sp_setup_tc_block_flower_bind(struct mlxsw_sp_port *mlxsw_sp_port,
bool register_block = false;
int err;
- block_cb = flow_block_cb_lookup(f, mlxsw_sp_setup_tc_block_cb_flower,
+ block_cb = flow_block_cb_lookup(f->block,
+ mlxsw_sp_setup_tc_block_cb_flower,
mlxsw_sp);
if (!block_cb) {
acl_block = mlxsw_sp_acl_block_create(mlxsw_sp, f->net);
@@ -1656,7 +1657,8 @@ mlxsw_sp_setup_tc_block_flower_unbind(struct mlxsw_sp_port *mlxsw_sp_port,
struct flow_block_cb *block_cb;
int err;
- block_cb = flow_block_cb_lookup(f, mlxsw_sp_setup_tc_block_cb_flower,
+ block_cb = flow_block_cb_lookup(f->block,
+ mlxsw_sp_setup_tc_block_cb_flower,
mlxsw_sp);
if (!block_cb)
return;
@@ -1717,7 +1719,7 @@ static int mlxsw_sp_setup_tc_block(struct mlxsw_sp_port *mlxsw_sp_port,
case FLOW_BLOCK_UNBIND:
mlxsw_sp_setup_tc_block_flower_unbind(mlxsw_sp_port,
f, ingress);
- block_cb = flow_block_cb_lookup(f, cb, mlxsw_sp_port);
+ block_cb = flow_block_cb_lookup(f->block, cb, mlxsw_sp_port);
if (!block_cb)
return -ENOENT;
diff --git a/drivers/net/ethernet/mscc/ocelot_flower.c b/drivers/net/ethernet/mscc/ocelot_flower.c
index 6a11aea8b186..59487d446a09 100644
--- a/drivers/net/ethernet/mscc/ocelot_flower.c
+++ b/drivers/net/ethernet/mscc/ocelot_flower.c
@@ -316,8 +316,8 @@ int ocelot_setup_tc_block_flower_bind(struct ocelot_port *port,
if (f->binder_type == FLOW_BLOCK_BINDER_TYPE_CLSACT_EGRESS)
return -EOPNOTSUPP;
- block_cb = flow_block_cb_lookup(f, ocelot_setup_tc_block_cb_flower,
- port);
+ block_cb = flow_block_cb_lookup(f->block,
+ ocelot_setup_tc_block_cb_flower, port);
if (!block_cb) {
port_block = ocelot_port_block_create(port);
if (!port_block)
@@ -350,8 +350,8 @@ void ocelot_setup_tc_block_flower_unbind(struct ocelot_port *port,
{
struct flow_block_cb *block_cb;
- block_cb = flow_block_cb_lookup(f, ocelot_setup_tc_block_cb_flower,
- port);
+ block_cb = flow_block_cb_lookup(f->block,
+ ocelot_setup_tc_block_cb_flower, port);
if (!block_cb)
return;
diff --git a/drivers/net/ethernet/mscc/ocelot_tc.c b/drivers/net/ethernet/mscc/ocelot_tc.c
index fba9512e9ca6..16a6db71ca5e 100644
--- a/drivers/net/ethernet/mscc/ocelot_tc.c
+++ b/drivers/net/ethernet/mscc/ocelot_tc.c
@@ -169,7 +169,7 @@ static int ocelot_setup_tc_block(struct ocelot_port *port,
list_add_tail(&block_cb->driver_list, f->driver_block_list);
return 0;
case FLOW_BLOCK_UNBIND:
- block_cb = flow_block_cb_lookup(f, cb, port);
+ block_cb = flow_block_cb_lookup(f->block, cb, port);
if (!block_cb)
return -ENOENT;
diff --git a/drivers/net/ethernet/netronome/nfp/flower/offload.c b/drivers/net/ethernet/netronome/nfp/flower/offload.c
index 93ab0db6c504..e209f150c5f2 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/offload.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/offload.c
@@ -1327,7 +1327,8 @@ static int nfp_flower_setup_tc_block(struct net_device *netdev,
list_add_tail(&block_cb->driver_list, &nfp_block_cb_list);
return 0;
case FLOW_BLOCK_UNBIND:
- block_cb = flow_block_cb_lookup(f, nfp_flower_setup_tc_block_cb,
+ block_cb = flow_block_cb_lookup(f->block,
+ nfp_flower_setup_tc_block_cb,
repr);
if (!block_cb)
return -ENOENT;
@@ -1440,7 +1441,7 @@ nfp_flower_setup_indr_tc_block(struct net_device *netdev, struct nfp_app *app,
if (!cb_priv)
return -ENOENT;
- block_cb = flow_block_cb_lookup(f,
+ block_cb = flow_block_cb_lookup(f->block,
nfp_flower_setup_indr_block_cb,
cb_priv);
if (!block_cb)
diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index 98bf3af5c84d..2b555ba78a02 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h
@@ -248,6 +248,10 @@ enum flow_block_binder_type {
FLOW_BLOCK_BINDER_TYPE_CLSACT_EGRESS,
};
+struct flow_block {
+ struct list_head cb_list;
+};
+
struct netlink_ext_ack;
struct flow_block_offload {
@@ -255,6 +259,7 @@ struct flow_block_offload {
enum flow_block_binder_type binder_type;
bool block_shared;
struct net *net;
+ struct flow_block *block;
struct list_head cb_list;
struct list_head *driver_block_list;
struct netlink_ext_ack *extack;
@@ -279,7 +284,7 @@ struct flow_block_cb *flow_block_cb_alloc(flow_setup_cb_t *cb,
void (*release)(void *cb_priv));
void flow_block_cb_free(struct flow_block_cb *block_cb);
-struct flow_block_cb *flow_block_cb_lookup(struct flow_block_offload *offload,
+struct flow_block_cb *flow_block_cb_lookup(struct flow_block *block,
flow_setup_cb_t *cb, void *cb_ident);
void *flow_block_cb_priv(struct flow_block_cb *block_cb);
@@ -336,4 +341,9 @@ flow_cls_offload_flow_rule(struct flow_cls_offload *flow_cmd)
return flow_cmd->rule;
}
+static inline void flow_block_init(struct flow_block *flow_block)
+{
+ INIT_LIST_HEAD(&flow_block->cb_list);
+}
+
#endif /* _NET_FLOW_OFFLOAD_H */
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index 35dfdd9f69b3..9b624566b82d 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -11,6 +11,7 @@
#include <linux/rhashtable.h>
#include <net/netfilter/nf_flow_table.h>
#include <net/netlink.h>
+#include <net/flow_offload.h>
struct module;
@@ -951,7 +952,7 @@ struct nft_stats {
* @stats: per-cpu chain stats
* @chain: the chain
* @dev_name: device name that this base chain is attached to (if any)
- * @cb_list: list of flow block callbacks (for hardware offload)
+ * @flow_block: flow block (for hardware offload)
*/
struct nft_base_chain {
struct nf_hook_ops ops;
@@ -961,7 +962,7 @@ struct nft_base_chain {
struct nft_stats __percpu *stats;
struct nft_chain chain;
char dev_name[IFNAMSIZ];
- struct list_head cb_list;
+ struct flow_block flow_block;
};
static inline struct nft_base_chain *nft_base_chain(const struct nft_chain *chain)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index 9482e060483b..6b6b01234dd9 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -399,7 +399,7 @@ struct tcf_block {
refcount_t refcnt;
struct net *net;
struct Qdisc *q;
- struct list_head cb_list;
+ struct flow_block flow_block;
struct list_head owner_list;
bool keep_dst;
unsigned int offloadcnt; /* Number of oddloaded filters */
diff --git a/net/core/flow_offload.c b/net/core/flow_offload.c
index a800fa78d96c..d63b970784dc 100644
--- a/net/core/flow_offload.c
+++ b/net/core/flow_offload.c
@@ -193,12 +193,12 @@ void flow_block_cb_free(struct flow_block_cb *block_cb)
}
EXPORT_SYMBOL(flow_block_cb_free);
-struct flow_block_cb *flow_block_cb_lookup(struct flow_block_offload *f,
+struct flow_block_cb *flow_block_cb_lookup(struct flow_block *block,
flow_setup_cb_t *cb, void *cb_ident)
{
struct flow_block_cb *block_cb;
- list_for_each_entry(block_cb, f->driver_block_list, driver_list) {
+ list_for_each_entry(block_cb, &block->cb_list, list) {
if (block_cb->cb == cb &&
block_cb->cb_ident == cb_ident)
return block_cb;
@@ -268,7 +268,7 @@ int flow_block_cb_setup_simple(struct flow_block_offload *f,
list_add_tail(&block_cb->driver_list, driver_block_list);
return 0;
case FLOW_BLOCK_UNBIND:
- block_cb = flow_block_cb_lookup(f, cb, cb_ident);
+ block_cb = flow_block_cb_lookup(f->block, cb, cb_ident);
if (!block_cb)
return -ENOENT;
diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index d697a64fb564..33f41178afcc 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -975,7 +975,7 @@ static int dsa_slave_setup_tc_block(struct net_device *dev,
list_add_tail(&block_cb->driver_list, &dsa_slave_block_cb_list);
return 0;
case FLOW_BLOCK_UNBIND:
- block_cb = flow_block_cb_lookup(f, cb, dev);
+ block_cb = flow_block_cb_lookup(f->block, cb, dev);
if (!block_cb)
return -ENOENT;
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 014e06b0b5cf..605a7cfe7ca7 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -1662,7 +1662,7 @@ static int nf_tables_addchain(struct nft_ctx *ctx, u8 family, u8 genmask,
chain->flags |= NFT_BASE_CHAIN | flags;
basechain->policy = NF_ACCEPT;
- INIT_LIST_HEAD(&basechain->cb_list);
+ flow_block_init(&basechain->flow_block);
} else {
chain = kzalloc(sizeof(*chain), GFP_KERNEL);
if (chain == NULL)
diff --git a/net/netfilter/nf_tables_offload.c b/net/netfilter/nf_tables_offload.c
index 2c3302845f67..64f5fd5f240e 100644
--- a/net/netfilter/nf_tables_offload.c
+++ b/net/netfilter/nf_tables_offload.c
@@ -116,7 +116,7 @@ static int nft_setup_cb_call(struct nft_base_chain *basechain,
struct flow_block_cb *block_cb;
int err;
- list_for_each_entry(block_cb, &basechain->cb_list, list) {
+ list_for_each_entry(block_cb, &basechain->flow_block.cb_list, list) {
err = block_cb->cb(type, type_data, block_cb->cb_priv);
if (err < 0)
return err;
@@ -154,7 +154,7 @@ static int nft_flow_offload_rule(struct nft_trans *trans,
static int nft_flow_offload_bind(struct flow_block_offload *bo,
struct nft_base_chain *basechain)
{
- list_splice(&bo->cb_list, &basechain->cb_list);
+ list_splice(&bo->cb_list, &basechain->flow_block.cb_list);
return 0;
}
@@ -198,6 +198,7 @@ static int nft_flow_offload_chain(struct nft_trans *trans,
return -EOPNOTSUPP;
bo.command = cmd;
+ bo.block = &basechain->flow_block;
bo.binder_type = FLOW_BLOCK_BINDER_TYPE_CLSACT_INGRESS;
bo.extack = &extack;
INIT_LIST_HEAD(&bo.cb_list);
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 78f0f2815b8c..15796fd47fda 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -691,6 +691,8 @@ static void tc_indr_block_ing_cmd(struct tc_indr_block_dev *indr_dev,
if (!indr_dev->block)
return;
+ bo.block = &indr_dev->block->flow_block;
+
indr_block_cb->cb(indr_dev->dev, indr_block_cb->cb_priv, TC_SETUP_BLOCK,
&bo);
tcf_block_setup(indr_dev->block, &bo);
@@ -775,6 +777,7 @@ static void tc_indr_block_call(struct tcf_block *block, struct net_device *dev,
.command = command,
.binder_type = ei->binder_type,
.net = dev_net(dev),
+ .block = &block->flow_block,
.block_shared = tcf_block_shared(block),
.extack = extack,
};
@@ -810,6 +813,7 @@ static int tcf_block_offload_cmd(struct tcf_block *block,
bo.net = dev_net(dev);
bo.command = command;
bo.binder_type = ei->binder_type;
+ bo.block = &block->flow_block;
bo.block_shared = tcf_block_shared(block);
bo.extack = extack;
INIT_LIST_HEAD(&bo.cb_list);
@@ -987,8 +991,8 @@ static struct tcf_block *tcf_block_create(struct net *net, struct Qdisc *q,
return ERR_PTR(-ENOMEM);
}
mutex_init(&block->lock);
+ flow_block_init(&block->flow_block);
INIT_LIST_HEAD(&block->chain_list);
- INIT_LIST_HEAD(&block->cb_list);
INIT_LIST_HEAD(&block->owner_list);
INIT_LIST_HEAD(&block->chain0.filter_chain_list);
@@ -1570,7 +1574,7 @@ static int tcf_block_bind(struct tcf_block *block,
i++;
}
- list_splice(&bo->cb_list, &block->cb_list);
+ list_splice(&bo->cb_list, &block->flow_block.cb_list);
return 0;
@@ -3156,7 +3160,7 @@ int tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
if (block->nooffloaddevcnt && err_stop)
return -EOPNOTSUPP;
- list_for_each_entry(block_cb, &block->cb_list, list) {
+ list_for_each_entry(block_cb, &block->flow_block.cb_list, list) {
err = block_cb->cb(type, type_data, block_cb->cb_priv);
if (err) {
if (err_stop)
--
2.11.0
^ permalink raw reply related
* [PATCH net,v4 2/4] net: flow_offload: remove netns parameter from flow_block_cb_alloc()
From: Pablo Neira Ayuso @ 2019-07-18 17:59 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, jiri, jakub.kicinski, pshelar
In-Reply-To: <20190718175931.13529-1-pablo@netfilter.org>
No need to annotate the netns on the flow block callback object,
flow_block_cb_is_busy() already checks for used blocks.
Fixes: d63db30c8537 ("net: flow_offload: add flow_block_cb_alloc() and flow_block_cb_free()")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
v4: no changes.
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 3 +--
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 5 ++---
drivers/net/ethernet/mscc/ocelot_flower.c | 3 +--
drivers/net/ethernet/mscc/ocelot_tc.c | 2 +-
drivers/net/ethernet/netronome/nfp/flower/offload.c | 6 ++----
include/net/flow_offload.h | 3 +--
net/core/flow_offload.c | 9 +++------
net/dsa/slave.c | 2 +-
8 files changed, 12 insertions(+), 21 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 7245d287633d..2162412073c5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -735,8 +735,7 @@ mlx5e_rep_indr_setup_tc_block(struct net_device *netdev,
list_add(&indr_priv->list,
&rpriv->uplink_priv.tc_indr_block_priv_list);
- block_cb = flow_block_cb_alloc(f->net,
- mlx5e_rep_indr_setup_block_cb,
+ block_cb = flow_block_cb_alloc(mlx5e_rep_indr_setup_block_cb,
indr_priv, indr_priv,
mlx5e_rep_indr_tc_block_unbind);
if (IS_ERR(block_cb)) {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 4d34d42b3b0e..a469035400cf 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -1610,8 +1610,7 @@ mlxsw_sp_setup_tc_block_flower_bind(struct mlxsw_sp_port *mlxsw_sp_port,
acl_block = mlxsw_sp_acl_block_create(mlxsw_sp, f->net);
if (!acl_block)
return -ENOMEM;
- block_cb = flow_block_cb_alloc(f->net,
- mlxsw_sp_setup_tc_block_cb_flower,
+ block_cb = flow_block_cb_alloc(mlxsw_sp_setup_tc_block_cb_flower,
mlxsw_sp, acl_block,
mlxsw_sp_tc_block_flower_release);
if (IS_ERR(block_cb)) {
@@ -1702,7 +1701,7 @@ static int mlxsw_sp_setup_tc_block(struct mlxsw_sp_port *mlxsw_sp_port,
&mlxsw_sp_block_cb_list))
return -EBUSY;
- block_cb = flow_block_cb_alloc(f->net, cb, mlxsw_sp_port,
+ block_cb = flow_block_cb_alloc(cb, mlxsw_sp_port,
mlxsw_sp_port, NULL);
if (IS_ERR(block_cb))
return PTR_ERR(block_cb);
diff --git a/drivers/net/ethernet/mscc/ocelot_flower.c b/drivers/net/ethernet/mscc/ocelot_flower.c
index 7aaddc09c185..6a11aea8b186 100644
--- a/drivers/net/ethernet/mscc/ocelot_flower.c
+++ b/drivers/net/ethernet/mscc/ocelot_flower.c
@@ -323,8 +323,7 @@ int ocelot_setup_tc_block_flower_bind(struct ocelot_port *port,
if (!port_block)
return -ENOMEM;
- block_cb = flow_block_cb_alloc(f->net,
- ocelot_setup_tc_block_cb_flower,
+ block_cb = flow_block_cb_alloc(ocelot_setup_tc_block_cb_flower,
port, port_block,
ocelot_tc_block_unbind);
if (IS_ERR(block_cb)) {
diff --git a/drivers/net/ethernet/mscc/ocelot_tc.c b/drivers/net/ethernet/mscc/ocelot_tc.c
index 9e6464ffae5d..abbcb66bf5ac 100644
--- a/drivers/net/ethernet/mscc/ocelot_tc.c
+++ b/drivers/net/ethernet/mscc/ocelot_tc.c
@@ -156,7 +156,7 @@ static int ocelot_setup_tc_block(struct ocelot_port *port,
if (flow_block_cb_is_busy(cb, port, &ocelot_block_cb_list))
return -EBUSY;
- block_cb = flow_block_cb_alloc(f->net, cb, port, port, NULL);
+ block_cb = flow_block_cb_alloc(cb, port, port, NULL);
if (IS_ERR(block_cb))
return PTR_ERR(block_cb);
diff --git a/drivers/net/ethernet/netronome/nfp/flower/offload.c b/drivers/net/ethernet/netronome/nfp/flower/offload.c
index faa8ba012a37..93ab0db6c504 100644
--- a/drivers/net/ethernet/netronome/nfp/flower/offload.c
+++ b/drivers/net/ethernet/netronome/nfp/flower/offload.c
@@ -1318,8 +1318,7 @@ static int nfp_flower_setup_tc_block(struct net_device *netdev,
&nfp_block_cb_list))
return -EBUSY;
- block_cb = flow_block_cb_alloc(f->net,
- nfp_flower_setup_tc_block_cb,
+ block_cb = flow_block_cb_alloc(nfp_flower_setup_tc_block_cb,
repr, repr, NULL);
if (IS_ERR(block_cb))
return PTR_ERR(block_cb);
@@ -1424,8 +1423,7 @@ nfp_flower_setup_indr_tc_block(struct net_device *netdev, struct nfp_app *app,
cb_priv->app = app;
list_add(&cb_priv->list, &priv->indr_block_cb_priv);
- block_cb = flow_block_cb_alloc(f->net,
- nfp_flower_setup_indr_block_cb,
+ block_cb = flow_block_cb_alloc(nfp_flower_setup_indr_block_cb,
cb_priv, cb_priv,
nfp_flower_setup_indr_tc_release);
if (IS_ERR(block_cb)) {
diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index db337299e81e..aa9b5287b231 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h
@@ -264,7 +264,6 @@ struct flow_block_offload {
struct flow_block_cb {
struct list_head driver_list;
struct list_head list;
- struct net *net;
tc_setup_cb_t *cb;
void *cb_ident;
void *cb_priv;
@@ -272,7 +271,7 @@ struct flow_block_cb {
unsigned int refcnt;
};
-struct flow_block_cb *flow_block_cb_alloc(struct net *net, tc_setup_cb_t *cb,
+struct flow_block_cb *flow_block_cb_alloc(tc_setup_cb_t *cb,
void *cb_ident, void *cb_priv,
void (*release)(void *cb_priv));
void flow_block_cb_free(struct flow_block_cb *block_cb);
diff --git a/net/core/flow_offload.c b/net/core/flow_offload.c
index 76f8db3841d7..507de4b48815 100644
--- a/net/core/flow_offload.c
+++ b/net/core/flow_offload.c
@@ -165,7 +165,7 @@ void flow_rule_match_enc_opts(const struct flow_rule *rule,
}
EXPORT_SYMBOL(flow_rule_match_enc_opts);
-struct flow_block_cb *flow_block_cb_alloc(struct net *net, tc_setup_cb_t *cb,
+struct flow_block_cb *flow_block_cb_alloc(tc_setup_cb_t *cb,
void *cb_ident, void *cb_priv,
void (*release)(void *cb_priv))
{
@@ -175,7 +175,6 @@ struct flow_block_cb *flow_block_cb_alloc(struct net *net, tc_setup_cb_t *cb,
if (!block_cb)
return ERR_PTR(-ENOMEM);
- block_cb->net = net;
block_cb->cb = cb;
block_cb->cb_ident = cb_ident;
block_cb->cb_priv = cb_priv;
@@ -200,8 +199,7 @@ struct flow_block_cb *flow_block_cb_lookup(struct flow_block_offload *f,
struct flow_block_cb *block_cb;
list_for_each_entry(block_cb, f->driver_block_list, driver_list) {
- if (block_cb->net == f->net &&
- block_cb->cb == cb &&
+ if (block_cb->cb == cb &&
block_cb->cb_ident == cb_ident)
return block_cb;
}
@@ -261,8 +259,7 @@ int flow_block_cb_setup_simple(struct flow_block_offload *f,
if (flow_block_cb_is_busy(cb, cb_ident, driver_block_list))
return -EBUSY;
- block_cb = flow_block_cb_alloc(f->net, cb, cb_ident,
- cb_priv, NULL);
+ block_cb = flow_block_cb_alloc(cb, cb_ident, cb_priv, NULL);
if (IS_ERR(block_cb))
return PTR_ERR(block_cb);
diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index 614c38ece104..6ca9ec58f881 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -967,7 +967,7 @@ static int dsa_slave_setup_tc_block(struct net_device *dev,
if (flow_block_cb_is_busy(cb, dev, &dsa_slave_block_cb_list))
return -EBUSY;
- block_cb = flow_block_cb_alloc(f->net, cb, dev, dev, NULL);
+ block_cb = flow_block_cb_alloc(cb, dev, dev, NULL);
if (IS_ERR(block_cb))
return PTR_ERR(block_cb);
--
2.11.0
^ permalink raw reply related
* [PATCH net,v4 1/4] net: openvswitch: rename flow_stats to sw_flow_stats
From: Pablo Neira Ayuso @ 2019-07-18 17:59 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, jiri, jakub.kicinski, pshelar
In-Reply-To: <20190718175931.13529-1-pablo@netfilter.org>
There is a flow_stats structure defined in include/net/flow_offload.h
and a follow up patch adds #include <net/flow_offload.h> to
net/sch_generic.h.
This breaks compilation since OVS codebase includes net/sock.h which
pulls in linux/filter.h which includes net/sch_generic.h.
In file included from ./include/net/sch_generic.h:18:0,
from ./include/linux/filter.h:25,
from ./include/net/sock.h:59,
from ./include/linux/tcp.h:19,
from net/openvswitch/datapath.c:24
This definition takes precedence to OVS since it is placed in the
networking core, so rename flow_stats in OVS to sw_flow_stats since
this structure is contained in the sw_flow object.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
v4: more detailed explanation - David Miller.
replace a few more spots using flow_stats in ovs codebase.
net/openvswitch/flow.c | 8 ++++----
net/openvswitch/flow.h | 4 ++--
net/openvswitch/flow_table.c | 8 ++++----
3 files changed, 10 insertions(+), 10 deletions(-)
diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
index dca3b1e2acf0..bc89e16e0505 100644
--- a/net/openvswitch/flow.c
+++ b/net/openvswitch/flow.c
@@ -59,7 +59,7 @@ u64 ovs_flow_used_time(unsigned long flow_jiffies)
void ovs_flow_stats_update(struct sw_flow *flow, __be16 tcp_flags,
const struct sk_buff *skb)
{
- struct flow_stats *stats;
+ struct sw_flow_stats *stats;
unsigned int cpu = smp_processor_id();
int len = skb->len + (skb_vlan_tag_present(skb) ? VLAN_HLEN : 0);
@@ -87,7 +87,7 @@ void ovs_flow_stats_update(struct sw_flow *flow, __be16 tcp_flags,
if (likely(flow->stats_last_writer != -1) &&
likely(!rcu_access_pointer(flow->stats[cpu]))) {
/* Try to allocate CPU-specific stats. */
- struct flow_stats *new_stats;
+ struct sw_flow_stats *new_stats;
new_stats =
kmem_cache_alloc_node(flow_stats_cache,
@@ -134,7 +134,7 @@ void ovs_flow_stats_get(const struct sw_flow *flow,
/* We open code this to make sure cpu 0 is always considered */
for (cpu = 0; cpu < nr_cpu_ids; cpu = cpumask_next(cpu, &flow->cpu_used_mask)) {
- struct flow_stats *stats = rcu_dereference_ovsl(flow->stats[cpu]);
+ struct sw_flow_stats *stats = rcu_dereference_ovsl(flow->stats[cpu]);
if (stats) {
/* Local CPU may write on non-local stats, so we must
@@ -158,7 +158,7 @@ void ovs_flow_stats_clear(struct sw_flow *flow)
/* We open code this to make sure cpu 0 is always considered */
for (cpu = 0; cpu < nr_cpu_ids; cpu = cpumask_next(cpu, &flow->cpu_used_mask)) {
- struct flow_stats *stats = ovsl_dereference(flow->stats[cpu]);
+ struct sw_flow_stats *stats = ovsl_dereference(flow->stats[cpu]);
if (stats) {
spin_lock_bh(&stats->lock);
diff --git a/net/openvswitch/flow.h b/net/openvswitch/flow.h
index 3e2cc2202d66..a5506e2d4b7a 100644
--- a/net/openvswitch/flow.h
+++ b/net/openvswitch/flow.h
@@ -194,7 +194,7 @@ struct sw_flow_actions {
struct nlattr actions[];
};
-struct flow_stats {
+struct sw_flow_stats {
u64 packet_count; /* Number of packets matched. */
u64 byte_count; /* Number of bytes matched. */
unsigned long used; /* Last used time (in jiffies). */
@@ -216,7 +216,7 @@ struct sw_flow {
struct cpumask cpu_used_mask;
struct sw_flow_mask *mask;
struct sw_flow_actions __rcu *sf_acts;
- struct flow_stats __rcu *stats[]; /* One for each CPU. First one
+ struct sw_flow_stats __rcu *stats[]; /* One for each CPU. First one
* is allocated at flow creation time,
* the rest are allocated on demand
* while holding the 'stats[0].lock'.
diff --git a/net/openvswitch/flow_table.c b/net/openvswitch/flow_table.c
index 988fd8a94e43..cf3582c5ed70 100644
--- a/net/openvswitch/flow_table.c
+++ b/net/openvswitch/flow_table.c
@@ -66,7 +66,7 @@ void ovs_flow_mask_key(struct sw_flow_key *dst, const struct sw_flow_key *src,
struct sw_flow *ovs_flow_alloc(void)
{
struct sw_flow *flow;
- struct flow_stats *stats;
+ struct sw_flow_stats *stats;
flow = kmem_cache_zalloc(flow_cache, GFP_KERNEL);
if (!flow)
@@ -110,7 +110,7 @@ static void flow_free(struct sw_flow *flow)
for (cpu = 0; cpu < nr_cpu_ids; cpu = cpumask_next(cpu, &flow->cpu_used_mask))
if (flow->stats[cpu])
kmem_cache_free(flow_stats_cache,
- (struct flow_stats __force *)flow->stats[cpu]);
+ (struct sw_flow_stats __force *)flow->stats[cpu]);
kmem_cache_free(flow_cache, flow);
}
@@ -712,13 +712,13 @@ int ovs_flow_init(void)
flow_cache = kmem_cache_create("sw_flow", sizeof(struct sw_flow)
+ (nr_cpu_ids
- * sizeof(struct flow_stats *)),
+ * sizeof(struct sw_flow_stats *)),
0, 0, NULL);
if (flow_cache == NULL)
return -ENOMEM;
flow_stats_cache
- = kmem_cache_create("sw_flow_stats", sizeof(struct flow_stats),
+ = kmem_cache_create("sw_flow_stats", sizeof(struct sw_flow_stats),
0, SLAB_HWCACHE_ALIGN, NULL);
if (flow_stats_cache == NULL) {
kmem_cache_destroy(flow_cache);
--
2.11.0
^ permalink raw reply related
* [PATCH net,v4 0/4] flow_offload fixes
From: Pablo Neira Ayuso @ 2019-07-18 17:59 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, jiri, jakub.kicinski, pshelar
Hi,
The following patchset contains fixes for the flow_offload infrastructure:
1) Fix possible build breakage before patch 3/4. Both the flow_offload
infrastructure and OVS define the flow_stats structure. Patch 3/4 in
this batch indirectly pulls in the flow_stats definition from
include/net/flow_offload.h into OVS, leading to structure redefinition
compile-time errors.
2) Remove netns parameter from flow_block_cb_alloc(), this is not
required as Jiri suggests. The flow_block_cb_is_busy() function uses
the per-driver block list to check for used blocks which was the
original intention for this parameter.
3) Rename tc_setup_cb_t to flow_setup_cb_t. This callback is not
exclusive of tc anymore, this might confuse the reader as Jiri
suggests, fix this semantic inconsistency.
4) Fix block sharing feature: Add flow_block structure and use it,
update flow_block_cb_lookup() to use this flow_block object.
Please, apply, thanks.
Pablo Neira Ayuso (4):
net: openvswitch: rename flow_stats to sw_flow_stats
net: flow_offload: remove netns parameter from flow_block_cb_alloc()
net: flow_offload: rename tc_setup_cb_t to flow_setup_cb_t
net: flow_offload: add flow_block structure and use it
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 5 ++--
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 15 +++++------
drivers/net/ethernet/mscc/ocelot_flower.c | 11 ++++----
drivers/net/ethernet/mscc/ocelot_tc.c | 6 ++---
.../net/ethernet/netronome/nfp/flower/offload.c | 11 ++++----
include/net/flow_offload.h | 29 ++++++++++++++++------
include/net/netfilter/nf_tables.h | 5 ++--
include/net/pkt_cls.h | 5 ++--
include/net/sch_generic.h | 8 +++---
net/core/flow_offload.c | 22 ++++++++--------
net/dsa/slave.c | 6 ++---
net/netfilter/nf_tables_api.c | 2 +-
net/netfilter/nf_tables_offload.c | 5 ++--
net/openvswitch/flow.c | 8 +++---
net/openvswitch/flow.h | 4 +--
net/openvswitch/flow_table.c | 8 +++---
net/sched/cls_api.c | 12 ++++++---
net/sched/cls_bpf.c | 2 +-
net/sched/cls_flower.c | 2 +-
net/sched/cls_matchall.c | 2 +-
net/sched/cls_u32.c | 6 ++---
21 files changed, 93 insertions(+), 81 deletions(-)
--
2.11.0
^ permalink raw reply
* Re: [PATCH bpf] libbpf: fix missing __WORDSIZE definition
From: Arnaldo Carvalho de Melo @ 2019-07-18 17:55 UTC (permalink / raw)
To: Andrii Nakryiko
Cc: bpf, netdev, daniel, ast, andrii.nakryiko, kernel-team, Jiri Olsa,
Namhyung Kim, Arnaldo Carvalho de Melo
In-Reply-To: <20190718172513.2394157-1-andriin@fb.com>
Em Thu, Jul 18, 2019 at 10:25:13AM -0700, Andrii Nakryiko escreveu:
> hashmap.h depends on __WORDSIZE being defined. It is defined by
> glibc/musl in different headers. It's an explicit goal for musl to be
> "non-detectable" at compilation time, so instead include glibc header if
> glibc is explicitly detected and fall back to musl header otherwise.
>
> Fixes: e3b924224028 ("libbpf: add resizable non-thread safe internal hashmap")
> Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> Signed-off-by: Andrii Nakryiko <andriin@fb.com>
I fixed this here differently, as below, I didn't send it because I'm still
testing it, so far, with a few other fixes and cherry-picking "libbpf: fix ptr
to u64 conversion warning on 32-bit platforms" that is still in the bpf tree
and is needed for the cross build containers in my suite that are 32-bit, I
have the results below, this builds perf + libbpf (where elfutils is available,
which is in most cases, except the uCLibc containers due to missing argp-devel),
with gcc and with clang:
[perfbuilder@quaco linux-perf-tools-build]$ export PERF_TARBALL=http://192.168.124.1/perf/perf-5.2.0.tar.xz
[perfbuilder@quaco linux-perf-tools-build]$ time dm
1 alpine:3.4 : Ok gcc (Alpine 5.3.0) 5.3.0, clang version 3.8.0 (tags/RELEASE_380/final)
2 alpine:3.5 : Ok gcc (Alpine 6.2.1) 6.2.1 20160822, clang version 3.8.1 (tags/RELEASE_381/final)
3 alpine:3.6 : Ok gcc (Alpine 6.3.0) 6.3.0, clang version 4.0.0 (tags/RELEASE_400/final)
4 alpine:3.7 : Ok gcc (Alpine 6.4.0) 6.4.0, Alpine clang version 5.0.0 (tags/RELEASE_500/final) (based on LLVM 5.0.0)
5 alpine:3.8 : Ok gcc (Alpine 6.4.0) 6.4.0, Alpine clang version 5.0.1 (tags/RELEASE_501/final) (based on LLVM 5.0.1)
6 alpine:3.9 : Ok gcc (Alpine 8.3.0) 8.3.0, Alpine clang version 5.0.1 (tags/RELEASE_502/final) (based on LLVM 5.0.1)
7 alpine:3.10 : Ok gcc (Alpine 8.3.0) 8.3.0, Alpine clang version 8.0.0 (tags/RELEASE_800/final) (based on LLVM 8.0.0)
8 alpine:edge : Ok gcc (Alpine 8.3.0) 8.3.0, Alpine clang version 7.0.1 (tags/RELEASE_701/final) (based on LLVM 7.0.1)
9 amazonlinux:1 : Ok gcc (GCC) 7.2.1 20170915 (Red Hat 7.2.1-2), clang version 3.6.2 (tags/RELEASE_362/final)
10 amazonlinux:2 : Ok gcc (GCC) 7.3.1 20180303 (Red Hat 7.3.1-5), clang version 7.0.1 (Amazon Linux 2 7.0.1-1.amzn2.0.2)
11 android-ndk:r12b-arm : Ok arm-linux-androideabi-gcc (GCC) 4.9.x 20150123 (prerelease)
12 android-ndk:r15c-arm : Ok arm-linux-androideabi-gcc (GCC) 4.9.x 20150123 (prerelease)
13 centos:5 : Ok gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-55)
14 centos:6 : Ok gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-23)
15 centos:7 : Ok gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-36), clang version 3.4.2 (tags/RELEASE_34/dot2-final)
16 clearlinux:latest : Ok gcc (Clear Linux OS for Intel Architecture) 9.1.1 20190628 gcc-9-branch@272773, clang version 8.0.0 (tags/RELEASE_800/final)
17 debian:8 : Ok gcc (Debian 4.9.2-10+deb8u2) 4.9.2, Debian clang version 3.5.0-10 (tags/RELEASE_350/final) (based on LLVM 3.5.0)
18 debian:9 : Ok gcc (Debian 6.3.0-18+deb9u1) 6.3.0 20170516, clang version 3.8.1-24 (tags/RELEASE_381/final)
19 debian:10 : Ok gcc (Debian 8.3.0-6) 8.3.0, clang version 7.0.1-8 (tags/RELEASE_701/final)
20 debian:experimental : Ok gcc (Debian 8.3.0-19) 8.3.0, clang version 7.0.1-8 (tags/RELEASE_701/final)
21 debian:experimental-x-arm64 : Ok aarch64-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0
22 debian:experimental-x-mips : Ok mips-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0
23 debian:experimental-x-mips64 : Ok mips64-linux-gnuabi64-gcc (Debian 8.3.0-7) 8.3.0
24 debian:experimental-x-mipsel : Ok mipsel-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0
25 fedora:20 : Ok gcc (GCC) 4.8.3 20140911 (Red Hat 4.8.3-7), clang version 3.4.2 (tags/RELEASE_34/dot2-final)
26 fedora:22 : Ok gcc (GCC) 5.3.1 20160406 (Red Hat 5.3.1-6), clang version 3.5.0 (tags/RELEASE_350/final)
27 fedora:23 : Ok gcc (GCC) 5.3.1 20160406 (Red Hat 5.3.1-6), clang version 3.7.0 (tags/RELEASE_370/final)
28 fedora:24 : Ok gcc (GCC) 6.3.1 20161221 (Red Hat 6.3.1-1), clang version 3.8.1 (tags/RELEASE_381/final)
29 fedora:24-x-ARC-uClibc : Ok arc-linux-gcc (ARCompact ISA Linux uClibc toolchain 2017.09-rc2) 7.1.1 20170710
I've pushed it to a tmp.perf/core branch in my
git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git tree, that has
these:
d5e1f2d60d41 (HEAD -> perf/core, acme.korg/tmp.perf/core) libbpf: fix ptr to u64 conversion warning on 32-bit platforms
7c08fd16f917 tools lib bpf: Avoid designated initializers for unnamed union members
4c9f83c95ad6 tools lib bpf: Avoid using 'link' as it shadows a global definition in some systems
bdb07df4a0ad tools lib bpf: Fix endianness macro usage for some compilers
66dbf3caff52 tools lib bpf: Replace __WORDSIZE with BITS_PER_LONG to build on the musl libc
Please take a look and check if everything is fine on your side. The HEAD I'll
remove if Daniel thinks it should wait that landing via the BPF tree, I just put it
there for the test builds.
commit 66dbf3caff52be0d004bcb9ac4cea4c19eb75dfc
Author: Arnaldo Carvalho de Melo <acme@redhat.com>
Date: Thu Jul 18 09:46:28 2019 -0300
tools lib bpf: Replace __WORDSIZE with BITS_PER_LONG to build on the musl libc
BITS_PER_LONG is more generally available and equivalent to __WORDSIZE,
so use it instead to keep it building in systems using the mustl libc
where __WORDSIZE is in a different place than in glibc.
And do this by explicitely adding the header where this definition is
(asm/bitsperlong.h) instead of getting it indirectly.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Fixes: e3b924224028 ("libbpf: add resizable non-thread safe internal hashmap")
Link: https://lkml.kernel.org/n/tip-61vydgldzmmz5w2mf6rv3ryl@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
diff --git a/tools/lib/bpf/hashmap.h b/tools/lib/bpf/hashmap.h
index 03748a742146..f1f37b574d9c 100644
--- a/tools/lib/bpf/hashmap.h
+++ b/tools/lib/bpf/hashmap.h
@@ -10,12 +10,13 @@
#include <stdbool.h>
#include <stddef.h>
+#include <asm/bitsperlong.h>
#include "libbpf_internal.h"
static inline size_t hash_bits(size_t h, int bits)
{
/* shuffle bits and return requested number of upper bits */
- return (h * 11400714819323198485llu) >> (__WORDSIZE - bits);
+ return (h * 11400714819323198485llu) >> (BITS_PER_LONG - bits);
}
typedef size_t (*hashmap_hash_fn)(const void *key, void *ctx);
^ permalink raw reply related
* Re: [PATCH v3 net-next 13/19] ionic: Add initial ethtool support
From: Shannon Nelson @ 2019-07-18 17:54 UTC (permalink / raw)
To: Andrew Lunn; +Cc: netdev
In-Reply-To: <20190718172823.GH25635@lunn.ch>
On 7/18/19 10:28 AM, Andrew Lunn wrote:
> On Thu, Jul 18, 2019 at 10:05:23AM -0700, Shannon Nelson wrote:
>> On 7/17/19 8:21 PM, Andrew Lunn wrote:
>>> On Tue, Jul 09, 2019 at 03:42:39PM -0700, Shannon Nelson wrote:
>>>> On 7/8/19 7:27 PM, Andrew Lunn wrote:
>>>>>> +static int ionic_get_module_eeprom(struct net_device *netdev,
>>>>>> + struct ethtool_eeprom *ee,
>>>>>> + u8 *data)
>>>>>> +{
>>>>>> + struct lif *lif = netdev_priv(netdev);
>>>>>> + struct ionic_dev *idev = &lif->ionic->idev;
>>>>>> + struct xcvr_status *xcvr;
>>>>>> + u32 len;
>>>>>> +
>>>>>> + /* copy the module bytes into data */
>>>>>> + xcvr = &idev->port_info->status.xcvr;
>>>>>> + len = min_t(u32, sizeof(xcvr->sprom), ee->len);
>>>>>> + memcpy(data, xcvr->sprom, len);
>>>>> Hi Shannon
>>>>>
>>>>> This also looks odd. Where is the call into the firmware to get the
>>>>> eeprom contents? Even though it is called 'eeprom', the data is not
>>>>> static. It contains real time diagnostic values, temperature, transmit
>>>>> power, receiver power, voltages etc.
>>>> idev->port_info is a memory mapped space that the device keeps up-to-date.
>>> Hi Shannon
>>>
>>> It at least needs a comment. How frequently does the device update
>>> this chunk of memory? It would be good to comment about that as
>>> well. Or do MMIO reads block while i2c operations occur to update the
>>> memory?
>> The device keeps this updated when changes happen internally so that there
>> is no need to block on MMIO read.
> Hi Shannon
>
> I'm thinking about the diagnostic page. RX and TX power, temperature,
> alarms etc. These are real time values, so you should read them on
> demand, or at last only cache them for a short time.
>
>
They *are* read on demand. The port_info and lif_info structs are dma
mapped spaces that the device keeps up-to-date with PCI writes in the
background so that the driver can do a quick memory read for current data.
sln
^ permalink raw reply
* Re: [PATCH iproute2 net-next v4 5/6] tc: etf: Add documentation for skip-skb-check.
From: David Ahern @ 2019-07-18 17:36 UTC (permalink / raw)
To: Vedang Patel, netdev
Cc: jhs, xiyou.wangcong, jiri, stephen, vinicius.gomes,
leandro.maciel.dorileo, jakub.kicinski, m-karicheri2
In-Reply-To: <1563306789-2908-5-git-send-email-vedang.patel@intel.com>
On 7/16/19 1:53 PM, Vedang Patel wrote:
> Document the newly added option (skip-skb-check) on the etf man-page.
>
> Signed-off-by: Vedang Patel <vedang.patel@intel.com>
> ---
> man/man8/tc-etf.8 | 10 ++++++++++
> 1 file changed, 10 insertions(+)
>
> diff --git a/man/man8/tc-etf.8 b/man/man8/tc-etf.8
> index 30a12de7d2c7..2e01a591dbaa 100644
> --- a/man/man8/tc-etf.8
> +++ b/man/man8/tc-etf.8
> @@ -106,6 +106,16 @@ referred to as "Launch Time" or "Time-Based Scheduling" by the
> documentation of network interface controllers.
> The default is for this option to be disabled.
>
> +.TP
> +skip_skb_check
patch 1 adds skip_sock_check.
> +.br
> +.BR etf(8)
> +currently drops any packet which does not have a socket associated with it or
> +if the socket does not have SO_TXTIME socket option set. But, this will not
> +work if the launchtime is set by another entity inside the kernel (e.g. some
> +other Qdisc). Setting the skip_skb_check will skip checking for a socket
> +associated with the packet.
> +
> .SH EXAMPLES
>
> ETF is used to enforce a Quality of Service. It controls when each
>
^ permalink raw reply
* Re: [PATCH iproute2 net-next v4 1/6] Update kernel headers
From: David Ahern @ 2019-07-18 17:32 UTC (permalink / raw)
To: Vedang Patel, netdev
Cc: jhs, xiyou.wangcong, jiri, stephen, vinicius.gomes,
leandro.maciel.dorileo, jakub.kicinski, m-karicheri2
In-Reply-To: <1563306789-2908-1-git-send-email-vedang.patel@intel.com>
On 7/16/19 1:53 PM, Vedang Patel wrote:
> The type for txtime-delay parameter will change from s32 to u32. So,
> make the corresponding change in the ABI file as well.
>
> Signed-off-by: Vedang Patel <vedang.patel@intel.com>
> ---
> include/uapi/linux/pkt_sched.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h
> index 1f623252abe8..18f185299f47 100644
> --- a/include/uapi/linux/pkt_sched.h
> +++ b/include/uapi/linux/pkt_sched.h
> @@ -1174,7 +1174,7 @@ enum {
> TCA_TAPRIO_ATTR_SCHED_CYCLE_TIME, /* s64 */
> TCA_TAPRIO_ATTR_SCHED_CYCLE_TIME_EXTENSION, /* s64 */
> TCA_TAPRIO_ATTR_FLAGS, /* u32 */
> - TCA_TAPRIO_ATTR_TXTIME_DELAY, /* s32 */
> + TCA_TAPRIO_ATTR_TXTIME_DELAY, /* u32 */
> __TCA_TAPRIO_ATTR_MAX,
> };
>
>
kernel uapi headers are synced from the kernel. You will need to make
this change to the kernel header and it will make its way to iproute2
^ permalink raw reply
* Re: [PATCH v2 bpf] libbpf: fix missing __WORDSIZE definition
From: Andrii Nakryiko @ 2019-07-18 17:31 UTC (permalink / raw)
To: Arnaldo Carvalho de Melo
Cc: bpf, Networking, Alexei Starovoitov, Daniel Borkmann,
Arnaldo Carvalho de Melo, Kernel Team, Andrii Nakryiko
In-Reply-To: <20190718173021.2418606-1-andriin@fb.com>
On Thu, Jul 18, 2019 at 10:30 AM Andrii Nakryiko <andriin@fb.com> wrote:
>
> hashmap.h depends on __WORDSIZE being defined. It is defined by
> glibc/musl in different headers. It's an explicit goal for musl to be
> "non-detectable" at compilation time, so instead include glibc header if
> glibc is explicitly detected and fall back to musl header otherwise.
>
> Fixes: e3b924224028 ("libbpf: add resizable non-thread safe internal hashmap")
> Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Arnaldo,
As it compiled fine for me even before and I don't have a musl set up,
could you please verify this fixes the issue for perf for all your
test environments?
Thanks!
> Signed-off-by: Andrii Nakryiko <andriin@fb.com>
> ---
> tools/lib/bpf/hashmap.h | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/tools/lib/bpf/hashmap.h b/tools/lib/bpf/hashmap.h
> index 03748a742146..bae8879cdf58 100644
> --- a/tools/lib/bpf/hashmap.h
> +++ b/tools/lib/bpf/hashmap.h
> @@ -10,6 +10,11 @@
>
> #include <stdbool.h>
> #include <stddef.h>
> +#ifdef __GLIBC__
> +#include <bits/wordsize.h>
> +#else
> +#include <bits/reg.h>
> +#endif
> #include "libbpf_internal.h"
>
> static inline size_t hash_bits(size_t h, int bits)
> --
> 2.17.1
>
^ permalink raw reply
* [PATCH v2 bpf] libbpf: fix missing __WORDSIZE definition
From: Andrii Nakryiko @ 2019-07-18 17:30 UTC (permalink / raw)
To: bpf, netdev, ast, daniel, acme
Cc: andrii.nakryiko, kernel-team, Andrii Nakryiko
hashmap.h depends on __WORDSIZE being defined. It is defined by
glibc/musl in different headers. It's an explicit goal for musl to be
"non-detectable" at compilation time, so instead include glibc header if
glibc is explicitly detected and fall back to musl header otherwise.
Fixes: e3b924224028 ("libbpf: add resizable non-thread safe internal hashmap")
Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
---
tools/lib/bpf/hashmap.h | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/tools/lib/bpf/hashmap.h b/tools/lib/bpf/hashmap.h
index 03748a742146..bae8879cdf58 100644
--- a/tools/lib/bpf/hashmap.h
+++ b/tools/lib/bpf/hashmap.h
@@ -10,6 +10,11 @@
#include <stdbool.h>
#include <stddef.h>
+#ifdef __GLIBC__
+#include <bits/wordsize.h>
+#else
+#include <bits/reg.h>
+#endif
#include "libbpf_internal.h"
static inline size_t hash_bits(size_t h, int bits)
--
2.17.1
^ permalink raw reply related
* Re: [PATCH v3 net-next 13/19] ionic: Add initial ethtool support
From: Andrew Lunn @ 2019-07-18 17:28 UTC (permalink / raw)
To: Shannon Nelson; +Cc: netdev
In-Reply-To: <35761cbb-aa2d-d698-7368-6ebe25607fe0@pensando.io>
On Thu, Jul 18, 2019 at 10:05:23AM -0700, Shannon Nelson wrote:
> On 7/17/19 8:21 PM, Andrew Lunn wrote:
> >On Tue, Jul 09, 2019 at 03:42:39PM -0700, Shannon Nelson wrote:
> >>On 7/8/19 7:27 PM, Andrew Lunn wrote:
> >>>>+static int ionic_get_module_eeprom(struct net_device *netdev,
> >>>>+ struct ethtool_eeprom *ee,
> >>>>+ u8 *data)
> >>>>+{
> >>>>+ struct lif *lif = netdev_priv(netdev);
> >>>>+ struct ionic_dev *idev = &lif->ionic->idev;
> >>>>+ struct xcvr_status *xcvr;
> >>>>+ u32 len;
> >>>>+
> >>>>+ /* copy the module bytes into data */
> >>>>+ xcvr = &idev->port_info->status.xcvr;
> >>>>+ len = min_t(u32, sizeof(xcvr->sprom), ee->len);
> >>>>+ memcpy(data, xcvr->sprom, len);
> >>>Hi Shannon
> >>>
> >>>This also looks odd. Where is the call into the firmware to get the
> >>>eeprom contents? Even though it is called 'eeprom', the data is not
> >>>static. It contains real time diagnostic values, temperature, transmit
> >>>power, receiver power, voltages etc.
> >>idev->port_info is a memory mapped space that the device keeps up-to-date.
> >Hi Shannon
> >
> >It at least needs a comment. How frequently does the device update
> >this chunk of memory? It would be good to comment about that as
> >well. Or do MMIO reads block while i2c operations occur to update the
> >memory?
>
> The device keeps this updated when changes happen internally so that there
> is no need to block on MMIO read.
Hi Shannon
I'm thinking about the diagnostic page. RX and TX power, temperature,
alarms etc. These are real time values, so you should read them on
demand, or at last only cache them for a short time.
Andrew
^ permalink raw reply
* Re: [PATCH bpf] libbpf: fix missing __WORDSIZE definition
From: Andrii Nakryiko @ 2019-07-18 17:26 UTC (permalink / raw)
To: Andrii Nakryiko
Cc: Arnaldo Carvalho de Melo, bpf, Networking, Daniel Borkmann,
Alexei Starovoitov, Kernel Team
In-Reply-To: <20190718172513.2394157-1-andriin@fb.com>
On Thu, Jul 18, 2019 at 10:25 AM Andrii Nakryiko <andriin@fb.com> wrote:
>
> hashmap.h depends on __WORDSIZE being defined. It is defined by
> glibc/musl in different headers. It's an explicit goal for musl to be
> "non-detectable" at compilation time, so instead include glibc header if
> glibc is explicitly detected and fall back to musl header otherwise.
>
> Fixes: e3b924224028 ("libbpf: add resizable non-thread safe internal hashmap")
> Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> Signed-off-by: Andrii Nakryiko <andriin@fb.com>
> ---
> tools/lib/bpf/hashmap.h | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/tools/lib/bpf/hashmap.h b/tools/lib/bpf/hashmap.h
> index 03748a742146..46a8cb667994 100644
> --- a/tools/lib/bpf/hashmap.h
> +++ b/tools/lib/bpf/hashmap.h
> @@ -10,7 +10,12 @@
>
> #include <stdbool.h>
> #include <stddef.h>
> +#ifdef __GLIBC__
> +#include <bits/wordsize.h>
> +#else
> +#include <bits/reg.h>
> #include "libbpf_internal.h"
> +#endif
Disregard this version, #endif on the wrong line. Fixing in v2.
>
> static inline size_t hash_bits(size_t h, int bits)
> {
> --
> 2.17.1
>
^ permalink raw reply
* [PATCH bpf] libbpf: fix missing __WORDSIZE definition
From: Andrii Nakryiko @ 2019-07-18 17:25 UTC (permalink / raw)
To: acme, bpf, netdev, daniel, ast
Cc: andrii.nakryiko, kernel-team, Andrii Nakryiko
hashmap.h depends on __WORDSIZE being defined. It is defined by
glibc/musl in different headers. It's an explicit goal for musl to be
"non-detectable" at compilation time, so instead include glibc header if
glibc is explicitly detected and fall back to musl header otherwise.
Fixes: e3b924224028 ("libbpf: add resizable non-thread safe internal hashmap")
Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
---
tools/lib/bpf/hashmap.h | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/tools/lib/bpf/hashmap.h b/tools/lib/bpf/hashmap.h
index 03748a742146..46a8cb667994 100644
--- a/tools/lib/bpf/hashmap.h
+++ b/tools/lib/bpf/hashmap.h
@@ -10,7 +10,12 @@
#include <stdbool.h>
#include <stddef.h>
+#ifdef __GLIBC__
+#include <bits/wordsize.h>
+#else
+#include <bits/reg.h>
#include "libbpf_internal.h"
+#endif
static inline size_t hash_bits(size_t h, int bits)
{
--
2.17.1
^ permalink raw reply related
* Re: [PATCH] net: dsa: sja1105: Fix missing unlock on error in sk_buff()
From: Florian Fainelli @ 2019-07-18 17:24 UTC (permalink / raw)
To: Wei Yongjun, Andrew Lunn, Vivien Didelot, Vladimir Oltean
Cc: netdev, kernel-janitors
In-Reply-To: <20190717062956.127446-1-weiyongjun1@huawei.com>
On 7/16/19 11:29 PM, Wei Yongjun wrote:
> Add the missing unlock before return from function sk_buff()
> in the error handling case.
>
> Fixes: f3097be21bf1 ("net: dsa: sja1105: Add a state machine for RX timestamping")
> Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
--
Florian
^ 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