* Re: [PATCH] net: phy: replace bool members in struct phy_device with bit-fields
From: David Miller @ 2018-05-24 19:36 UTC (permalink / raw)
To: hkallweit1; +Cc: f.fainelli, andrew, netdev
In-Reply-To: <3c59ea3d-f707-b991-1f88-8540891488b9@gmail.com>
From: Heiner Kallweit <hkallweit1@gmail.com>
Date: Wed, 23 May 2018 08:05:20 +0200
> In struct phy_device we have a number of flags being defined as type
> bool. Similar to e.g. struct pci_dev we can save some space by using
> bit-fields.
>
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Applied to net-next, thanks.
^ permalink raw reply
* Re: [PATCH net] vhost: synchronize IOTLB message with dev cleanup
From: David Miller @ 2018-05-24 19:35 UTC (permalink / raw)
To: jasowang; +Cc: mst, kvm, virtualization, netdev, linux-kernel
In-Reply-To: <1526990337-24892-1-git-send-email-jasowang@redhat.com>
From: Jason Wang <jasowang@redhat.com>
Date: Tue, 22 May 2018 19:58:57 +0800
> DaeRyong Jeong reports a race between vhost_dev_cleanup() and
> vhost_process_iotlb_msg():
>
> Thread interleaving:
> CPU0 (vhost_process_iotlb_msg) CPU1 (vhost_dev_cleanup)
> (In the case of both VHOST_IOTLB_UPDATE and
> VHOST_IOTLB_INVALIDATE)
> ===== =====
> vhost_umem_clean(dev->iotlb);
> if (!dev->iotlb) {
> ret = -EFAULT;
> break;
> }
> dev->iotlb = NULL;
>
> The reason is we don't synchronize between them, fixing by protecting
> vhost_process_iotlb_msg() with dev mutex.
>
> Reported-by: DaeRyong Jeong <threeearcat@gmail.com>
> Fixes: 6b1e6cc7855b0 ("vhost: new device IOTLB API")
> Signed-off-by: Jason Wang <jasowang@redhat.com>
Michael, please review.
^ permalink raw reply
* Re: [PATCH net-next 0/8] nfp: offload LAG for tc flower egress
From: Or Gerlitz @ 2018-05-24 19:26 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David Miller, Linux Netdev List, oss-drivers, Jiri Pirko,
Jay Vosburgh, Veaceslav Falico, Andy Gospodarek
In-Reply-To: <20180524114929.0fb4e38f@cakuba>
On Thu, May 24, 2018 at 9:49 PM, Jakub Kicinski
<jakub.kicinski@netronome.com> wrote:
> On Thu, 24 May 2018 20:04:56 +0300, Or Gerlitz wrote:
>> Does this apply also to non-uplink representors? if yes, what is the use case?
>>
>> We are looking on supporting uplink lag in sriov switchdev scheme - we refer to
>> it as "vf lag" -- b/c the netdev and rdma devices seen by the VF are actually
>> subject to HA and/or LAG - I wasn't sure if/how you limit this series
>> to uplink reprs
>
> I don't think we have a limitation on the output port within the LAG.
> But keep in mind in our devices all ports belong to the same eswitch/PF
> so bonding uplink ports is generally sufficient, I'm not sure VF
> bonding adds much HA. IOW AFAIK we support VF bonding because HW can do
> it easily, not because we have a strong use case for it.
To make it clear, vf lag is code name for uplink lag, I think we want
to say that
we provide the VM a lagged VF, anyway, again, the lag is done on the uplink reps
not on the vf reps. Unlike the uplink port which is physical one, the
vf vport is virtual
one, what could be the benefit to bond two vports?
^ permalink raw reply
* Re: [PATCH] rtlwifi: remove duplicate code
From: Joe Perches @ 2018-05-24 19:24 UTC (permalink / raw)
To: Gustavo A. R. Silva, Ping-Ke Shih, Kalle Valo, David S. Miller
Cc: linux-wireless, netdev, linux-kernel
In-Reply-To: <20180524185450.GA2875@embeddedor.com>
On Thu, 2018-05-24 at 13:54 -0500, Gustavo A. R. Silva wrote:
> Remove and refactor some code in order to avoid having identical code
> for different branches.
True and nice tool and patch submittal thanks.
> Notice that the logic has been there since 2014.
But perhaps the original logic is a defective copy/paste
and it should be corrected instead.
Can anyone from realtek verify this?
> Addresses-Coverity-ID: 1426199 ("Identical code for different branches")
> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
> ---
> .../realtek/rtlwifi/btcoexist/halbtc8723b2ant.c | 23 ++++------------------
> 1 file changed, 4 insertions(+), 19 deletions(-)
>
> diff --git a/drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtc8723b2ant.c b/drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtc8723b2ant.c
> index 279fe01..df3facc 100644
> --- a/drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtc8723b2ant.c
> +++ b/drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtc8723b2ant.c
> @@ -2876,25 +2876,10 @@ static void btc8723b2ant_action_hid(struct btc_coexist *btcoexist)
> btc8723b2ant_ps_tdma(btcoexist, NORMAL_EXEC, true, 13);
>
> /* sw mechanism */
> - if (BTC_WIFI_BW_HT40 == wifi_bw) {
> - if ((wifi_rssi_state == BTC_RSSI_STATE_HIGH) ||
> - (wifi_rssi_state == BTC_RSSI_STATE_STAY_HIGH)) {
> - btc8723b2ant_sw_mechanism(btcoexist, true, true,
> - false, false);
> - } else {
> - btc8723b2ant_sw_mechanism(btcoexist, true, true,
> - false, false);
> - }
> - } else {
> - if ((wifi_rssi_state == BTC_RSSI_STATE_HIGH) ||
> - (wifi_rssi_state == BTC_RSSI_STATE_STAY_HIGH)) {
> - btc8723b2ant_sw_mechanism(btcoexist, false, true,
> - false, false);
> - } else {
> - btc8723b2ant_sw_mechanism(btcoexist, false, true,
> - false, false);
> - }
> - }
> + if (wifi_bw == BTC_WIFI_BW_HT40)
> + btc8723b2ant_sw_mechanism(btcoexist, true, true, false, false);
> + else
> + btc8723b2ant_sw_mechanism(btcoexist, false, true, false, false);
> }
>
> /* A2DP only / PAN(EDR) only/ A2DP+PAN(HS) */
^ permalink raw reply
* Re: Poor TCP performance with XPS enabled after scrubbing skb
From: Flavio Leitner @ 2018-05-24 19:17 UTC (permalink / raw)
To: Eric Dumazet; +Cc: netdev, Paolo Abeni
In-Reply-To: <c8f6c2f8-c590-2912-c7af-8bce717480b6@gmail.com>
On Tue, May 15, 2018 at 02:08:09PM -0700, Eric Dumazet wrote:
>
>
> On 05/15/2018 12:31 PM, Flavio Leitner wrote:
> > Hi,
> >
> > There is a significant throughput issue (~50% drop) for a single TCP
> > stream when the skb is scrubbed and XPS is enabled.
> >
> > If I turn CONFIG_XPS off, then the issue never happens and the test
> > reaches line rate. The same happens if I echo 0 to tx-*/xps_cpus.
> >
> > It looks like that when the skb is scrubbed, there is no more reference
> > to the struct sock,
>
> And this is really the problem here, since it breaks back pressure (and TCP Small queues)
>
> I am not sure why skb_orphan() is used in this scrubbing really.
>
veth originally called skb_orphan() on veth_xmit() most probably
because there was no TX completion. Then the code got generalized to
dev_forward_skb() and later on moved to skb_scrub_packet().
The issue is that we call skb_scrub_packet() on TX and RX paths and
that is done while crossing netns. It doesn't look correct to keep
the ->sk because I suspect that iptables/selinux/bpf, or some code
path that I am probably missing could expose/use the wrong ->sk, for
example.
However, netdev_pick_tx() can't store the queue mapping without ->sk.
The hack in the first email relies on the headers (skb_tx_hash) to
always selected the same TX queue, which solves the original problem
but not the TCP small queues you mentioned.
--
Flavio
^ permalink raw reply
* [PATCH] rtlwifi: remove duplicate code
From: Gustavo A. R. Silva @ 2018-05-24 18:54 UTC (permalink / raw)
To: Ping-Ke Shih, Kalle Valo, David S. Miller
Cc: linux-wireless, netdev, linux-kernel, Gustavo A. R. Silva
Remove and refactor some code in order to avoid having identical code
for different branches.
Notice that the logic has been there since 2014.
Addresses-Coverity-ID: 1426199 ("Identical code for different branches")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
---
.../realtek/rtlwifi/btcoexist/halbtc8723b2ant.c | 23 ++++------------------
1 file changed, 4 insertions(+), 19 deletions(-)
diff --git a/drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtc8723b2ant.c b/drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtc8723b2ant.c
index 279fe01..df3facc 100644
--- a/drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtc8723b2ant.c
+++ b/drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtc8723b2ant.c
@@ -2876,25 +2876,10 @@ static void btc8723b2ant_action_hid(struct btc_coexist *btcoexist)
btc8723b2ant_ps_tdma(btcoexist, NORMAL_EXEC, true, 13);
/* sw mechanism */
- if (BTC_WIFI_BW_HT40 == wifi_bw) {
- if ((wifi_rssi_state == BTC_RSSI_STATE_HIGH) ||
- (wifi_rssi_state == BTC_RSSI_STATE_STAY_HIGH)) {
- btc8723b2ant_sw_mechanism(btcoexist, true, true,
- false, false);
- } else {
- btc8723b2ant_sw_mechanism(btcoexist, true, true,
- false, false);
- }
- } else {
- if ((wifi_rssi_state == BTC_RSSI_STATE_HIGH) ||
- (wifi_rssi_state == BTC_RSSI_STATE_STAY_HIGH)) {
- btc8723b2ant_sw_mechanism(btcoexist, false, true,
- false, false);
- } else {
- btc8723b2ant_sw_mechanism(btcoexist, false, true,
- false, false);
- }
- }
+ if (wifi_bw == BTC_WIFI_BW_HT40)
+ btc8723b2ant_sw_mechanism(btcoexist, true, true, false, false);
+ else
+ btc8723b2ant_sw_mechanism(btcoexist, false, true, false, false);
}
/* A2DP only / PAN(EDR) only/ A2DP+PAN(HS) */
--
2.7.4
^ permalink raw reply related
* Re: [PATCH net-next 0/8] nfp: offload LAG for tc flower egress
From: Jakub Kicinski @ 2018-05-24 18:53 UTC (permalink / raw)
To: Samudrala, Sridhar
Cc: Or Gerlitz, David Miller, Linux Netdev List, oss-drivers,
Jiri Pirko, Jay Vosburgh, Veaceslav Falico, Andy Gospodarek
In-Reply-To: <185db5ee-4a86-6479-46e6-4c48f9516a90@intel.com>
On Thu, 24 May 2018 11:23:00 -0700, Samudrala, Sridhar wrote:
> On 5/24/2018 10:04 AM, Or Gerlitz wrote:
> > On Thu, May 24, 2018 at 5:22 AM, Jakub Kicinski
> > <jakub.kicinski@netronome.com> wrote:
> >> Hi!
> >>
> >> This series from John adds bond offload to the nfp driver. Patch 5
> >> exposes the hash type for NETDEV_LAG_TX_TYPE_HASH to make sure nfp
> >> hashing matches that of the software LAG. This may be unnecessarily
> >> conservative, let's see what LAG maintainers think :)
> >>
> >> John says:
> >>
> >> This patchset sets up the infrastructure and offloads output actions for
> >> when a TC flower rule attempts to egress a packet to a LAG port.
> >>
> >> Firstly it adds some of the infrastructure required to the flower app and
> >> to the nfp core. This includes the ability to change the MAC address of a
> >> repr, a function for combining lookup and write to a FW symbol, and the
> >> addition of private data to a repr on a per app basis.
> >>
> >> Patch 6 continues by implementing notifiers that track Linux bonds and
> >> communicates to the FW those which enslave reprs, along with the current
> >> state of reprs within the bond.
> >>
> >> Patch 7 ensures bonds are synchronised with FW by receiving and acting
> >> upon cmsgs sent to the kernel. These may request that a bond message is
> >> retransmitted when FW can process it, or may request a full sync of the
> >> bonds defined in the kernel.
> >>
> >> Patch 8 offloads a flower action when that action requires egressing to a
> >> pre-defined Linux bond.
> > Does this apply also to non-uplink representors? if yes, what is the use case?
> >
> > We are looking on supporting uplink lag in sriov switchdev scheme - we refer to
> > it as "vf lag" -- b/c the netdev and rdma devices seen by the VF are actually
> > subject to HA and/or LAG - I wasn't sure if/how you limit this series
> > to uplink reprs
>
> Also, does this patchset support offloading LAG when using vxlan based tunnels?
>
> When using OVS offloading with vxlan, the encap rule that gets offloaded via tc-flower
> has egress port as vxlan device and the decap rule has the in-port as vxlan device, not
> the actual egress port. How are you addressing this issue?
It is very much on our radar, I think we will send out a related RFC
later today :)
But to be honest I think you can just install an egress callback on the
bond and that will pretty much work today. You don't have to "own" the
egress device to install a egdev callback on it.
^ permalink raw reply
* Re: [PATCH net-next 0/8] nfp: offload LAG for tc flower egress
From: Jakub Kicinski @ 2018-05-24 18:49 UTC (permalink / raw)
To: Or Gerlitz
Cc: David Miller, Linux Netdev List, oss-drivers, Jiri Pirko,
Jay Vosburgh, Veaceslav Falico, Andy Gospodarek
In-Reply-To: <CAJ3xEMj48Gvox-hCyrGEXNtcr7g_9+drxAN6jbaOSGLEHaappA@mail.gmail.com>
On Thu, 24 May 2018 20:04:56 +0300, Or Gerlitz wrote:
> On Thu, May 24, 2018 at 5:22 AM, Jakub Kicinski wrote:
> > Hi!
> >
> > This series from John adds bond offload to the nfp driver. Patch 5
> > exposes the hash type for NETDEV_LAG_TX_TYPE_HASH to make sure nfp
> > hashing matches that of the software LAG. This may be unnecessarily
> > conservative, let's see what LAG maintainers think :)
> >
> > John says:
> >
> > This patchset sets up the infrastructure and offloads output actions for
> > when a TC flower rule attempts to egress a packet to a LAG port.
> >
> > Firstly it adds some of the infrastructure required to the flower app and
> > to the nfp core. This includes the ability to change the MAC address of a
> > repr, a function for combining lookup and write to a FW symbol, and the
> > addition of private data to a repr on a per app basis.
> >
> > Patch 6 continues by implementing notifiers that track Linux bonds and
> > communicates to the FW those which enslave reprs, along with the current
> > state of reprs within the bond.
> >
> > Patch 7 ensures bonds are synchronised with FW by receiving and acting
> > upon cmsgs sent to the kernel. These may request that a bond message is
> > retransmitted when FW can process it, or may request a full sync of the
> > bonds defined in the kernel.
> >
> > Patch 8 offloads a flower action when that action requires egressing to a
> > pre-defined Linux bond.
>
> Does this apply also to non-uplink representors? if yes, what is the use case?
>
> We are looking on supporting uplink lag in sriov switchdev scheme - we refer to
> it as "vf lag" -- b/c the netdev and rdma devices seen by the VF are actually
> subject to HA and/or LAG - I wasn't sure if/how you limit this series
> to uplink reprs
I don't think we have a limitation on the output port within the LAG.
But keep in mind in our devices all ports belong to the same eswitch/PF
so bonding uplink ports is generally sufficient, I'm not sure VF
bonding adds much HA. IOW AFAIK we support VF bonding because HW can do
it easily, not because we have a strong use case for it.
^ permalink raw reply
* Re: [PATCH net-next 0/8] nfp: offload LAG for tc flower egress
From: Samudrala, Sridhar @ 2018-05-24 18:23 UTC (permalink / raw)
To: Or Gerlitz, Jakub Kicinski
Cc: David Miller, Linux Netdev List, oss-drivers, Jiri Pirko,
Jay Vosburgh, Veaceslav Falico, Andy Gospodarek
In-Reply-To: <CAJ3xEMj48Gvox-hCyrGEXNtcr7g_9+drxAN6jbaOSGLEHaappA@mail.gmail.com>
On 5/24/2018 10:04 AM, Or Gerlitz wrote:
> On Thu, May 24, 2018 at 5:22 AM, Jakub Kicinski
> <jakub.kicinski@netronome.com> wrote:
>> Hi!
>>
>> This series from John adds bond offload to the nfp driver. Patch 5
>> exposes the hash type for NETDEV_LAG_TX_TYPE_HASH to make sure nfp
>> hashing matches that of the software LAG. This may be unnecessarily
>> conservative, let's see what LAG maintainers think :)
>>
>> John says:
>>
>> This patchset sets up the infrastructure and offloads output actions for
>> when a TC flower rule attempts to egress a packet to a LAG port.
>>
>> Firstly it adds some of the infrastructure required to the flower app and
>> to the nfp core. This includes the ability to change the MAC address of a
>> repr, a function for combining lookup and write to a FW symbol, and the
>> addition of private data to a repr on a per app basis.
>>
>> Patch 6 continues by implementing notifiers that track Linux bonds and
>> communicates to the FW those which enslave reprs, along with the current
>> state of reprs within the bond.
>>
>> Patch 7 ensures bonds are synchronised with FW by receiving and acting
>> upon cmsgs sent to the kernel. These may request that a bond message is
>> retransmitted when FW can process it, or may request a full sync of the
>> bonds defined in the kernel.
>>
>> Patch 8 offloads a flower action when that action requires egressing to a
>> pre-defined Linux bond.
> Does this apply also to non-uplink representors? if yes, what is the use case?
>
> We are looking on supporting uplink lag in sriov switchdev scheme - we refer to
> it as "vf lag" -- b/c the netdev and rdma devices seen by the VF are actually
> subject to HA and/or LAG - I wasn't sure if/how you limit this series
> to uplink reprs
Also, does this patchset support offloading LAG when using vxlan based tunnels?
When using OVS offloading with vxlan, the encap rule that gets offloaded via tc-flower
has egress port as vxlan device and the decap rule has the in-port as vxlan device, not
the actual egress port. How are you addressing this issue?
^ permalink raw reply
* [PATCH bpf-next v5 7/7] tools/bpftool: add perf subcommand
From: Yonghong Song @ 2018-05-24 18:21 UTC (permalink / raw)
To: peterz, ast, daniel, netdev; +Cc: kernel-team
In-Reply-To: <20180524182158.456462-1-yhs@fb.com>
The new command "bpftool perf [show | list]" will traverse
all processes under /proc, and if any fd is associated
with a perf event, it will print out related perf event
information. Documentation is also added.
Below is an example to show the results using bcc commands.
Running the following 4 bcc commands:
kprobe: trace.py '__x64_sys_nanosleep'
kretprobe: trace.py 'r::__x64_sys_nanosleep'
tracepoint: trace.py 't:syscalls:sys_enter_nanosleep'
uprobe: trace.py 'p:/home/yhs/a.out:main'
The bpftool command line and result:
$ bpftool perf
pid 21711 fd 5: prog_id 5 kprobe func __x64_sys_write offset 0
pid 21765 fd 5: prog_id 7 kretprobe func __x64_sys_nanosleep offset 0
pid 21767 fd 5: prog_id 8 tracepoint sys_enter_nanosleep
pid 21800 fd 5: prog_id 9 uprobe filename /home/yhs/a.out offset 1159
$ bpftool -j perf
[{"pid":21711,"fd":5,"prog_id":5,"fd_type":"kprobe","func":"__x64_sys_write","offset":0}, \
{"pid":21765,"fd":5,"prog_id":7,"fd_type":"kretprobe","func":"__x64_sys_nanosleep","offset":0}, \
{"pid":21767,"fd":5,"prog_id":8,"fd_type":"tracepoint","tracepoint":"sys_enter_nanosleep"}, \
{"pid":21800,"fd":5,"prog_id":9,"fd_type":"uprobe","filename":"/home/yhs/a.out","offset":1159}]
$ bpftool prog
5: kprobe name probe___x64_sys tag e495a0c82f2c7a8d gpl
loaded_at 2018-05-15T04:46:37-0700 uid 0
xlated 200B not jited memlock 4096B map_ids 4
7: kprobe name probe___x64_sys tag f2fdee479a503abf gpl
loaded_at 2018-05-15T04:48:32-0700 uid 0
xlated 200B not jited memlock 4096B map_ids 7
8: tracepoint name tracepoint__sys tag 5390badef2395fcf gpl
loaded_at 2018-05-15T04:48:48-0700 uid 0
xlated 200B not jited memlock 4096B map_ids 8
9: kprobe name probe_main_1 tag 0a87bdc2e2953b6d gpl
loaded_at 2018-05-15T04:49:52-0700 uid 0
xlated 200B not jited memlock 4096B map_ids 9
$ ps ax | grep "python ./trace.py"
21711 pts/0 T 0:03 python ./trace.py __x64_sys_write
21765 pts/0 S+ 0:00 python ./trace.py r::__x64_sys_nanosleep
21767 pts/2 S+ 0:00 python ./trace.py t:syscalls:sys_enter_nanosleep
21800 pts/3 S+ 0:00 python ./trace.py p:/home/yhs/a.out:main
22374 pts/1 S+ 0:00 grep --color=auto python ./trace.py
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
---
tools/bpf/bpftool/Documentation/bpftool-perf.rst | 81 ++++++++
tools/bpf/bpftool/Documentation/bpftool.rst | 5 +-
tools/bpf/bpftool/bash-completion/bpftool | 9 +
tools/bpf/bpftool/main.c | 3 +-
tools/bpf/bpftool/main.h | 1 +
tools/bpf/bpftool/perf.c | 246 +++++++++++++++++++++++
6 files changed, 343 insertions(+), 2 deletions(-)
create mode 100644 tools/bpf/bpftool/Documentation/bpftool-perf.rst
create mode 100644 tools/bpf/bpftool/perf.c
diff --git a/tools/bpf/bpftool/Documentation/bpftool-perf.rst b/tools/bpf/bpftool/Documentation/bpftool-perf.rst
new file mode 100644
index 0000000..e3eb0ea
--- /dev/null
+++ b/tools/bpf/bpftool/Documentation/bpftool-perf.rst
@@ -0,0 +1,81 @@
+================
+bpftool-perf
+================
+-------------------------------------------------------------------------------
+tool for inspection of perf related bpf prog attachments
+-------------------------------------------------------------------------------
+
+:Manual section: 8
+
+SYNOPSIS
+========
+
+ **bpftool** [*OPTIONS*] **perf** *COMMAND*
+
+ *OPTIONS* := { [{ **-j** | **--json** }] [{ **-p** | **--pretty** }] }
+
+ *COMMANDS* :=
+ { **show** | **list** | **help** }
+
+PERF COMMANDS
+=============
+
+| **bpftool** **perf { show | list }**
+| **bpftool** **perf help**
+
+DESCRIPTION
+===========
+ **bpftool perf { show | list }**
+ List all raw_tracepoint, tracepoint, kprobe attachment in the system.
+
+ Output will start with process id and file descriptor in that process,
+ followed by bpf program id, attachment information, and attachment point.
+ The attachment point for raw_tracepoint/tracepoint is the trace probe name.
+ The attachment point for k[ret]probe is either symbol name and offset,
+ or a kernel virtual address.
+ The attachment point for u[ret]probe is the file name and the file offset.
+
+ **bpftool perf help**
+ Print short help message.
+
+OPTIONS
+=======
+ -h, --help
+ Print short generic help message (similar to **bpftool help**).
+
+ -v, --version
+ Print version number (similar to **bpftool version**).
+
+ -j, --json
+ Generate JSON output. For commands that cannot produce JSON, this
+ option has no effect.
+
+ -p, --pretty
+ Generate human-readable JSON output. Implies **-j**.
+
+EXAMPLES
+========
+
+| **# bpftool perf**
+
+::
+
+ pid 21711 fd 5: prog_id 5 kprobe func __x64_sys_write offset 0
+ pid 21765 fd 5: prog_id 7 kretprobe func __x64_sys_nanosleep offset 0
+ pid 21767 fd 5: prog_id 8 tracepoint sys_enter_nanosleep
+ pid 21800 fd 5: prog_id 9 uprobe filename /home/yhs/a.out offset 1159
+
+|
+| **# bpftool -j perf**
+
+::
+
+ [{"pid":21711,"fd":5,"prog_id":5,"fd_type":"kprobe","func":"__x64_sys_write","offset":0}, \
+ {"pid":21765,"fd":5,"prog_id":7,"fd_type":"kretprobe","func":"__x64_sys_nanosleep","offset":0}, \
+ {"pid":21767,"fd":5,"prog_id":8,"fd_type":"tracepoint","tracepoint":"sys_enter_nanosleep"}, \
+ {"pid":21800,"fd":5,"prog_id":9,"fd_type":"uprobe","filename":"/home/yhs/a.out","offset":1159}]
+
+
+SEE ALSO
+========
+ **bpftool**\ (8), **bpftool-prog**\ (8), **bpftool-map**\ (8)
diff --git a/tools/bpf/bpftool/Documentation/bpftool.rst b/tools/bpf/bpftool/Documentation/bpftool.rst
index 564cb0d..b6f5d56 100644
--- a/tools/bpf/bpftool/Documentation/bpftool.rst
+++ b/tools/bpf/bpftool/Documentation/bpftool.rst
@@ -16,7 +16,7 @@ SYNOPSIS
**bpftool** **version**
- *OBJECT* := { **map** | **program** | **cgroup** }
+ *OBJECT* := { **map** | **program** | **cgroup** | **perf** }
*OPTIONS* := { { **-V** | **--version** } | { **-h** | **--help** }
| { **-j** | **--json** } [{ **-p** | **--pretty** }] }
@@ -30,6 +30,8 @@ SYNOPSIS
*CGROUP-COMMANDS* := { **show** | **list** | **attach** | **detach** | **help** }
+ *PERF-COMMANDS* := { **show** | **list** | **help** }
+
DESCRIPTION
===========
*bpftool* allows for inspection and simple modification of BPF objects
@@ -56,3 +58,4 @@ OPTIONS
SEE ALSO
========
**bpftool-map**\ (8), **bpftool-prog**\ (8), **bpftool-cgroup**\ (8)
+ **bpftool-perf**\ (8)
diff --git a/tools/bpf/bpftool/bash-completion/bpftool b/tools/bpf/bpftool/bash-completion/bpftool
index b301c9b..7bc198d 100644
--- a/tools/bpf/bpftool/bash-completion/bpftool
+++ b/tools/bpf/bpftool/bash-completion/bpftool
@@ -448,6 +448,15 @@ _bpftool()
;;
esac
;;
+ perf)
+ case $command in
+ *)
+ [[ $prev == $object ]] && \
+ COMPREPLY=( $( compgen -W 'help \
+ show list' -- "$cur" ) )
+ ;;
+ esac
+ ;;
esac
} &&
complete -F _bpftool bpftool
diff --git a/tools/bpf/bpftool/main.c b/tools/bpf/bpftool/main.c
index 1ec852d..eea7f14 100644
--- a/tools/bpf/bpftool/main.c
+++ b/tools/bpf/bpftool/main.c
@@ -87,7 +87,7 @@ static int do_help(int argc, char **argv)
" %s batch file FILE\n"
" %s version\n"
"\n"
- " OBJECT := { prog | map | cgroup }\n"
+ " OBJECT := { prog | map | cgroup | perf }\n"
" " HELP_SPEC_OPTIONS "\n"
"",
bin_name, bin_name, bin_name);
@@ -216,6 +216,7 @@ static const struct cmd cmds[] = {
{ "prog", do_prog },
{ "map", do_map },
{ "cgroup", do_cgroup },
+ { "perf", do_perf },
{ "version", do_version },
{ 0 }
};
diff --git a/tools/bpf/bpftool/main.h b/tools/bpf/bpftool/main.h
index 6173cd9..63fdb31 100644
--- a/tools/bpf/bpftool/main.h
+++ b/tools/bpf/bpftool/main.h
@@ -119,6 +119,7 @@ int do_prog(int argc, char **arg);
int do_map(int argc, char **arg);
int do_event_pipe(int argc, char **argv);
int do_cgroup(int argc, char **arg);
+int do_perf(int argc, char **arg);
int prog_parse_fd(int *argc, char ***argv);
int map_parse_fd_and_info(int *argc, char ***argv, void *info, __u32 *info_len);
diff --git a/tools/bpf/bpftool/perf.c b/tools/bpf/bpftool/perf.c
new file mode 100644
index 0000000..ac6b1a1
--- /dev/null
+++ b/tools/bpf/bpftool/perf.c
@@ -0,0 +1,246 @@
+// SPDX-License-Identifier: GPL-2.0+
+// Copyright (C) 2018 Facebook
+// Author: Yonghong Song <yhs@fb.com>
+
+#define _GNU_SOURCE
+#include <ctype.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+#include <ftw.h>
+
+#include <bpf.h>
+
+#include "main.h"
+
+/* 0: undecided, 1: supported, 2: not supported */
+static int perf_query_supported;
+static bool has_perf_query_support(void)
+{
+ __u64 probe_offset, probe_addr;
+ __u32 len, prog_id, fd_type;
+ char buf[256];
+ int fd;
+
+ if (perf_query_supported)
+ goto out;
+
+ fd = open(bin_name, O_RDONLY);
+ if (fd < 0) {
+ p_err("perf_query_support: %s", strerror(errno));
+ goto out;
+ }
+
+ /* the following query will fail as no bpf attachment,
+ * the expected errno is ENOTSUPP
+ */
+ errno = 0;
+ len = sizeof(buf);
+ bpf_task_fd_query(getpid(), fd, 0, buf, &len, &prog_id,
+ &fd_type, &probe_offset, &probe_addr);
+
+ if (errno == 524 /* ENOTSUPP */) {
+ perf_query_supported = 1;
+ goto close_fd;
+ }
+
+ perf_query_supported = 2;
+ p_err("perf_query_support: %s", strerror(errno));
+ fprintf(stderr,
+ "HINT: non root or kernel doesn't support TASK_FD_QUERY\n");
+
+close_fd:
+ close(fd);
+out:
+ return perf_query_supported == 1;
+}
+
+static void print_perf_json(int pid, int fd, __u32 prog_id, __u32 fd_type,
+ char *buf, __u64 probe_offset, __u64 probe_addr)
+{
+ jsonw_start_object(json_wtr);
+ jsonw_int_field(json_wtr, "pid", pid);
+ jsonw_int_field(json_wtr, "fd", fd);
+ jsonw_uint_field(json_wtr, "prog_id", prog_id);
+ switch (fd_type) {
+ case BPF_FD_TYPE_RAW_TRACEPOINT:
+ jsonw_string_field(json_wtr, "fd_type", "raw_tracepoint");
+ jsonw_string_field(json_wtr, "tracepoint", buf);
+ break;
+ case BPF_FD_TYPE_TRACEPOINT:
+ jsonw_string_field(json_wtr, "fd_type", "tracepoint");
+ jsonw_string_field(json_wtr, "tracepoint", buf);
+ break;
+ case BPF_FD_TYPE_KPROBE:
+ jsonw_string_field(json_wtr, "fd_type", "kprobe");
+ if (buf[0] != '\0') {
+ jsonw_string_field(json_wtr, "func", buf);
+ jsonw_lluint_field(json_wtr, "offset", probe_offset);
+ } else {
+ jsonw_lluint_field(json_wtr, "addr", probe_addr);
+ }
+ break;
+ case BPF_FD_TYPE_KRETPROBE:
+ jsonw_string_field(json_wtr, "fd_type", "kretprobe");
+ if (buf[0] != '\0') {
+ jsonw_string_field(json_wtr, "func", buf);
+ jsonw_lluint_field(json_wtr, "offset", probe_offset);
+ } else {
+ jsonw_lluint_field(json_wtr, "addr", probe_addr);
+ }
+ break;
+ case BPF_FD_TYPE_UPROBE:
+ jsonw_string_field(json_wtr, "fd_type", "uprobe");
+ jsonw_string_field(json_wtr, "filename", buf);
+ jsonw_lluint_field(json_wtr, "offset", probe_offset);
+ break;
+ case BPF_FD_TYPE_URETPROBE:
+ jsonw_string_field(json_wtr, "fd_type", "uretprobe");
+ jsonw_string_field(json_wtr, "filename", buf);
+ jsonw_lluint_field(json_wtr, "offset", probe_offset);
+ break;
+ }
+ jsonw_end_object(json_wtr);
+}
+
+static void print_perf_plain(int pid, int fd, __u32 prog_id, __u32 fd_type,
+ char *buf, __u64 probe_offset, __u64 probe_addr)
+{
+ printf("pid %d fd %d: prog_id %u ", pid, fd, prog_id);
+ switch (fd_type) {
+ case BPF_FD_TYPE_RAW_TRACEPOINT:
+ printf("raw_tracepoint %s\n", buf);
+ break;
+ case BPF_FD_TYPE_TRACEPOINT:
+ printf("tracepoint %s\n", buf);
+ break;
+ case BPF_FD_TYPE_KPROBE:
+ if (buf[0] != '\0')
+ printf("kprobe func %s offset %llu\n", buf,
+ probe_offset);
+ else
+ printf("kprobe addr %llu\n", probe_addr);
+ break;
+ case BPF_FD_TYPE_KRETPROBE:
+ if (buf[0] != '\0')
+ printf("kretprobe func %s offset %llu\n", buf,
+ probe_offset);
+ else
+ printf("kretprobe addr %llu\n", probe_addr);
+ break;
+ case BPF_FD_TYPE_UPROBE:
+ printf("uprobe filename %s offset %llu\n", buf, probe_offset);
+ break;
+ case BPF_FD_TYPE_URETPROBE:
+ printf("uretprobe filename %s offset %llu\n", buf,
+ probe_offset);
+ break;
+ }
+}
+
+static int show_proc(const char *fpath, const struct stat *sb,
+ int tflag, struct FTW *ftwbuf)
+{
+ __u64 probe_offset, probe_addr;
+ __u32 len, prog_id, fd_type;
+ int err, pid = 0, fd = 0;
+ const char *pch;
+ char buf[4096];
+
+ /* prefix always /proc */
+ pch = fpath + 5;
+ if (*pch == '\0')
+ return 0;
+
+ /* pid should be all numbers */
+ pch++;
+ while (isdigit(*pch)) {
+ pid = pid * 10 + *pch - '0';
+ pch++;
+ }
+ if (*pch == '\0')
+ return 0;
+ if (*pch != '/')
+ return FTW_SKIP_SUBTREE;
+
+ /* check /proc/<pid>/fd directory */
+ pch++;
+ if (strncmp(pch, "fd", 2))
+ return FTW_SKIP_SUBTREE;
+ pch += 2;
+ if (*pch == '\0')
+ return 0;
+ if (*pch != '/')
+ return FTW_SKIP_SUBTREE;
+
+ /* check /proc/<pid>/fd/<fd_num> */
+ pch++;
+ while (isdigit(*pch)) {
+ fd = fd * 10 + *pch - '0';
+ pch++;
+ }
+ if (*pch != '\0')
+ return FTW_SKIP_SUBTREE;
+
+ /* query (pid, fd) for potential perf events */
+ len = sizeof(buf);
+ err = bpf_task_fd_query(pid, fd, 0, buf, &len, &prog_id, &fd_type,
+ &probe_offset, &probe_addr);
+ if (err < 0)
+ return 0;
+
+ if (json_output)
+ print_perf_json(pid, fd, prog_id, fd_type, buf, probe_offset,
+ probe_addr);
+ else
+ print_perf_plain(pid, fd, prog_id, fd_type, buf, probe_offset,
+ probe_addr);
+
+ return 0;
+}
+
+static int do_show(int argc, char **argv)
+{
+ int flags = FTW_ACTIONRETVAL | FTW_PHYS;
+ int err = 0, nopenfd = 16;
+
+ if (!has_perf_query_support())
+ return -1;
+
+ if (json_output)
+ jsonw_start_array(json_wtr);
+ if (nftw("/proc", show_proc, nopenfd, flags) == -1) {
+ p_err("%s", strerror(errno));
+ err = -1;
+ }
+ if (json_output)
+ jsonw_end_array(json_wtr);
+
+ return err;
+}
+
+static int do_help(int argc, char **argv)
+{
+ fprintf(stderr,
+ "Usage: %s %s { show | list | help }\n"
+ "",
+ bin_name, argv[-2]);
+
+ return 0;
+}
+
+static const struct cmd cmds[] = {
+ { "show", do_show },
+ { "list", do_show },
+ { "help", do_help },
+ { 0 }
+};
+
+int do_perf(int argc, char **argv)
+{
+ return cmd_select(cmds, argc, argv, do_help);
+}
--
2.9.5
^ permalink raw reply related
* [PATCH bpf-next v5 6/7] tools/bpf: add two BPF_TASK_FD_QUERY tests in test_progs
From: Yonghong Song @ 2018-05-24 18:21 UTC (permalink / raw)
To: peterz, ast, daniel, netdev; +Cc: kernel-team
In-Reply-To: <20180524182158.456462-1-yhs@fb.com>
The new tests are added to query perf_event information
for raw_tracepoint and tracepoint attachment. For tracepoint,
both syscalls and non-syscalls tracepoints are queries as
they are treated slightly differently inside the kernel.
Signed-off-by: Yonghong Song <yhs@fb.com>
---
tools/testing/selftests/bpf/test_progs.c | 158 +++++++++++++++++++++++++++++++
1 file changed, 158 insertions(+)
diff --git a/tools/testing/selftests/bpf/test_progs.c b/tools/testing/selftests/bpf/test_progs.c
index 3ecf733..0ef6820 100644
--- a/tools/testing/selftests/bpf/test_progs.c
+++ b/tools/testing/selftests/bpf/test_progs.c
@@ -1542,6 +1542,162 @@ static void test_get_stack_raw_tp(void)
bpf_object__close(obj);
}
+static void test_task_fd_query_rawtp(void)
+{
+ const char *file = "./test_get_stack_rawtp.o";
+ __u64 probe_offset, probe_addr;
+ __u32 len, prog_id, fd_type;
+ struct bpf_object *obj;
+ int efd, err, prog_fd;
+ __u32 duration = 0;
+ char buf[256];
+
+ err = bpf_prog_load(file, BPF_PROG_TYPE_RAW_TRACEPOINT, &obj, &prog_fd);
+ if (CHECK(err, "prog_load raw tp", "err %d errno %d\n", err, errno))
+ return;
+
+ efd = bpf_raw_tracepoint_open("sys_enter", prog_fd);
+ if (CHECK(efd < 0, "raw_tp_open", "err %d errno %d\n", efd, errno))
+ goto close_prog;
+
+ /* query (getpid(), efd) */
+ len = sizeof(buf);
+ err = bpf_task_fd_query(getpid(), efd, 0, buf, &len, &prog_id,
+ &fd_type, &probe_offset, &probe_addr);
+ if (CHECK(err < 0, "bpf_task_fd_query", "err %d errno %d\n", err,
+ errno))
+ goto close_prog;
+
+ err = fd_type == BPF_FD_TYPE_RAW_TRACEPOINT &&
+ strcmp(buf, "sys_enter") == 0;
+ if (CHECK(!err, "check_results", "fd_type %d tp_name %s\n",
+ fd_type, buf))
+ goto close_prog;
+
+ /* test zero len */
+ len = 0;
+ err = bpf_task_fd_query(getpid(), efd, 0, buf, &len, &prog_id,
+ &fd_type, &probe_offset, &probe_addr);
+ if (CHECK(err < 0, "bpf_task_fd_query (len = 0)", "err %d errno %d\n",
+ err, errno))
+ goto close_prog;
+ err = fd_type == BPF_FD_TYPE_RAW_TRACEPOINT &&
+ len == strlen("sys_enter");
+ if (CHECK(!err, "check_results", "fd_type %d len %u\n", fd_type, len))
+ goto close_prog;
+
+ /* test empty buffer */
+ len = sizeof(buf);
+ err = bpf_task_fd_query(getpid(), efd, 0, 0, &len, &prog_id,
+ &fd_type, &probe_offset, &probe_addr);
+ if (CHECK(err < 0, "bpf_task_fd_query (buf = 0)", "err %d errno %d\n",
+ err, errno))
+ goto close_prog;
+ err = fd_type == BPF_FD_TYPE_RAW_TRACEPOINT &&
+ len == strlen("sys_enter");
+ if (CHECK(!err, "check_results", "fd_type %d len %u\n", fd_type, len))
+ goto close_prog;
+
+ /* test smaller buffer */
+ len = 3;
+ err = bpf_task_fd_query(getpid(), efd, 0, buf, &len, &prog_id,
+ &fd_type, &probe_offset, &probe_addr);
+ if (CHECK(err >= 0 || errno != ENOSPC, "bpf_task_fd_query (len = 3)",
+ "err %d errno %d\n", err, errno))
+ goto close_prog;
+ err = fd_type == BPF_FD_TYPE_RAW_TRACEPOINT &&
+ len == strlen("sys_enter") &&
+ strcmp(buf, "sy") == 0;
+ if (CHECK(!err, "check_results", "fd_type %d len %u\n", fd_type, len))
+ goto close_prog;
+
+ goto close_prog_noerr;
+close_prog:
+ error_cnt++;
+close_prog_noerr:
+ bpf_object__close(obj);
+}
+
+static void test_task_fd_query_tp_core(const char *probe_name,
+ const char *tp_name)
+{
+ const char *file = "./test_tracepoint.o";
+ int err, bytes, efd, prog_fd, pmu_fd;
+ struct perf_event_attr attr = {};
+ __u64 probe_offset, probe_addr;
+ __u32 len, prog_id, fd_type;
+ struct bpf_object *obj;
+ __u32 duration = 0;
+ char buf[256];
+
+ err = bpf_prog_load(file, BPF_PROG_TYPE_TRACEPOINT, &obj, &prog_fd);
+ if (CHECK(err, "bpf_prog_load", "err %d errno %d\n", err, errno))
+ goto close_prog;
+
+ snprintf(buf, sizeof(buf),
+ "/sys/kernel/debug/tracing/events/%s/id", probe_name);
+ efd = open(buf, O_RDONLY, 0);
+ if (CHECK(efd < 0, "open", "err %d errno %d\n", efd, errno))
+ goto close_prog;
+ bytes = read(efd, buf, sizeof(buf));
+ close(efd);
+ if (CHECK(bytes <= 0 || bytes >= sizeof(buf), "read",
+ "bytes %d errno %d\n", bytes, errno))
+ goto close_prog;
+
+ attr.config = strtol(buf, NULL, 0);
+ attr.type = PERF_TYPE_TRACEPOINT;
+ attr.sample_type = PERF_SAMPLE_RAW;
+ attr.sample_period = 1;
+ attr.wakeup_events = 1;
+ pmu_fd = syscall(__NR_perf_event_open, &attr, -1 /* pid */,
+ 0 /* cpu 0 */, -1 /* group id */,
+ 0 /* flags */);
+ if (CHECK(err, "perf_event_open", "err %d errno %d\n", err, errno))
+ goto close_pmu;
+
+ err = ioctl(pmu_fd, PERF_EVENT_IOC_ENABLE, 0);
+ if (CHECK(err, "perf_event_ioc_enable", "err %d errno %d\n", err,
+ errno))
+ goto close_pmu;
+
+ err = ioctl(pmu_fd, PERF_EVENT_IOC_SET_BPF, prog_fd);
+ if (CHECK(err, "perf_event_ioc_set_bpf", "err %d errno %d\n", err,
+ errno))
+ goto close_pmu;
+
+ /* query (getpid(), pmu_fd) */
+ len = sizeof(buf);
+ err = bpf_task_fd_query(getpid(), pmu_fd, 0, buf, &len, &prog_id,
+ &fd_type, &probe_offset, &probe_addr);
+ if (CHECK(err < 0, "bpf_task_fd_query", "err %d errno %d\n", err,
+ errno))
+ goto close_pmu;
+
+ err = (fd_type == BPF_FD_TYPE_TRACEPOINT) && !strcmp(buf, tp_name);
+ if (CHECK(!err, "check_results", "fd_type %d tp_name %s\n",
+ fd_type, buf))
+ goto close_pmu;
+
+ close(pmu_fd);
+ goto close_prog_noerr;
+
+close_pmu:
+ close(pmu_fd);
+close_prog:
+ error_cnt++;
+close_prog_noerr:
+ bpf_object__close(obj);
+}
+
+static void test_task_fd_query_tp(void)
+{
+ test_task_fd_query_tp_core("sched/sched_switch",
+ "sched_switch");
+ test_task_fd_query_tp_core("syscalls/sys_enter_read",
+ "sys_enter_read");
+}
+
int main(void)
{
jit_enabled = is_jit_enabled();
@@ -1561,6 +1717,8 @@ int main(void)
test_stacktrace_build_id_nmi();
test_stacktrace_map_raw_tp();
test_get_stack_raw_tp();
+ test_task_fd_query_rawtp();
+ test_task_fd_query_tp();
printf("Summary: %d PASSED, %d FAILED\n", pass_cnt, error_cnt);
return error_cnt ? EXIT_FAILURE : EXIT_SUCCESS;
--
2.9.5
^ permalink raw reply related
* [PATCH bpf-next v5 5/7] samples/bpf: add a samples/bpf test for BPF_TASK_FD_QUERY
From: Yonghong Song @ 2018-05-24 18:21 UTC (permalink / raw)
To: peterz, ast, daniel, netdev; +Cc: kernel-team
This is mostly to test kprobe/uprobe which needs kernel headers.
Signed-off-by: Yonghong Song <yhs@fb.com>
---
samples/bpf/Makefile | 4 +
samples/bpf/task_fd_query_kern.c | 19 ++
samples/bpf/task_fd_query_user.c | 382 +++++++++++++++++++++++++++++++++++++++
3 files changed, 405 insertions(+)
create mode 100644 samples/bpf/task_fd_query_kern.c
create mode 100644 samples/bpf/task_fd_query_user.c
diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index 62d1aa1..7dc85ed 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -51,6 +51,7 @@ hostprogs-y += cpustat
hostprogs-y += xdp_adjust_tail
hostprogs-y += xdpsock
hostprogs-y += xdp_fwd
+hostprogs-y += task_fd_query
# Libbpf dependencies
LIBBPF = $(TOOLS_PATH)/lib/bpf/libbpf.a
@@ -105,6 +106,7 @@ cpustat-objs := bpf_load.o cpustat_user.o
xdp_adjust_tail-objs := xdp_adjust_tail_user.o
xdpsock-objs := bpf_load.o xdpsock_user.o
xdp_fwd-objs := bpf_load.o xdp_fwd_user.o
+task_fd_query-objs := bpf_load.o task_fd_query_user.o $(TRACE_HELPERS)
# Tell kbuild to always build the programs
always := $(hostprogs-y)
@@ -160,6 +162,7 @@ always += cpustat_kern.o
always += xdp_adjust_tail_kern.o
always += xdpsock_kern.o
always += xdp_fwd_kern.o
+always += task_fd_query_kern.o
HOSTCFLAGS += -I$(objtree)/usr/include
HOSTCFLAGS += -I$(srctree)/tools/lib/
@@ -175,6 +178,7 @@ HOSTCFLAGS_offwaketime_user.o += -I$(srctree)/tools/lib/bpf/
HOSTCFLAGS_spintest_user.o += -I$(srctree)/tools/lib/bpf/
HOSTCFLAGS_trace_event_user.o += -I$(srctree)/tools/lib/bpf/
HOSTCFLAGS_sampleip_user.o += -I$(srctree)/tools/lib/bpf/
+HOSTCFLAGS_task_fd_query_user.o += -I$(srctree)/tools/lib/bpf/
HOST_LOADLIBES += $(LIBBPF) -lelf
HOSTLOADLIBES_tracex4 += -lrt
diff --git a/samples/bpf/task_fd_query_kern.c b/samples/bpf/task_fd_query_kern.c
new file mode 100644
index 0000000..f4b0a9e
--- /dev/null
+++ b/samples/bpf/task_fd_query_kern.c
@@ -0,0 +1,19 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/version.h>
+#include <linux/ptrace.h>
+#include <uapi/linux/bpf.h>
+#include "bpf_helpers.h"
+
+SEC("kprobe/blk_start_request")
+int bpf_prog1(struct pt_regs *ctx)
+{
+ return 0;
+}
+
+SEC("kretprobe/blk_account_io_completion")
+int bpf_prog2(struct pt_regs *ctx)
+{
+ return 0;
+}
+char _license[] SEC("license") = "GPL";
+u32 _version SEC("version") = LINUX_VERSION_CODE;
diff --git a/samples/bpf/task_fd_query_user.c b/samples/bpf/task_fd_query_user.c
new file mode 100644
index 0000000..8381d79
--- /dev/null
+++ b/samples/bpf/task_fd_query_user.c
@@ -0,0 +1,382 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <unistd.h>
+#include <stdbool.h>
+#include <string.h>
+#include <stdint.h>
+#include <fcntl.h>
+#include <linux/bpf.h>
+#include <sys/ioctl.h>
+#include <sys/resource.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+
+#include "libbpf.h"
+#include "bpf_load.h"
+#include "bpf_util.h"
+#include "perf-sys.h"
+#include "trace_helpers.h"
+
+#define CHECK_PERROR_RET(condition) ({ \
+ int __ret = !!(condition); \
+ if (__ret) { \
+ printf("FAIL: %s:\n", __func__); \
+ perror(" "); \
+ return -1; \
+ } \
+})
+
+#define CHECK_AND_RET(condition) ({ \
+ int __ret = !!(condition); \
+ if (__ret) \
+ return -1; \
+})
+
+static __u64 ptr_to_u64(void *ptr)
+{
+ return (__u64) (unsigned long) ptr;
+}
+
+#define PMU_TYPE_FILE "/sys/bus/event_source/devices/%s/type"
+static int bpf_find_probe_type(const char *event_type)
+{
+ char buf[256];
+ int fd, ret;
+
+ ret = snprintf(buf, sizeof(buf), PMU_TYPE_FILE, event_type);
+ CHECK_PERROR_RET(ret < 0 || ret >= sizeof(buf));
+
+ fd = open(buf, O_RDONLY);
+ CHECK_PERROR_RET(fd < 0);
+
+ ret = read(fd, buf, sizeof(buf));
+ close(fd);
+ CHECK_PERROR_RET(ret < 0 || ret >= sizeof(buf));
+
+ errno = 0;
+ ret = (int)strtol(buf, NULL, 10);
+ CHECK_PERROR_RET(errno);
+ return ret;
+}
+
+#define PMU_RETPROBE_FILE "/sys/bus/event_source/devices/%s/format/retprobe"
+static int bpf_get_retprobe_bit(const char *event_type)
+{
+ char buf[256];
+ int fd, ret;
+
+ ret = snprintf(buf, sizeof(buf), PMU_RETPROBE_FILE, event_type);
+ CHECK_PERROR_RET(ret < 0 || ret >= sizeof(buf));
+
+ fd = open(buf, O_RDONLY);
+ CHECK_PERROR_RET(fd < 0);
+
+ ret = read(fd, buf, sizeof(buf));
+ close(fd);
+ CHECK_PERROR_RET(ret < 0 || ret >= sizeof(buf));
+ CHECK_PERROR_RET(strlen(buf) < strlen("config:"));
+
+ errno = 0;
+ ret = (int)strtol(buf + strlen("config:"), NULL, 10);
+ CHECK_PERROR_RET(errno);
+ return ret;
+}
+
+static int test_debug_fs_kprobe(int prog_fd_idx, const char *fn_name,
+ __u32 expected_fd_type)
+{
+ __u64 probe_offset, probe_addr;
+ __u32 len, prog_id, fd_type;
+ char buf[256];
+ int err;
+
+ len = sizeof(buf);
+ err = bpf_task_fd_query(getpid(), event_fd[prog_fd_idx], 0, buf, &len,
+ &prog_id, &fd_type, &probe_offset,
+ &probe_addr);
+ if (err < 0) {
+ printf("FAIL: %s, for event_fd idx %d, fn_name %s\n",
+ __func__, prog_fd_idx, fn_name);
+ perror(" :");
+ return -1;
+ }
+ if (strcmp(buf, fn_name) != 0 ||
+ fd_type != expected_fd_type ||
+ probe_offset != 0x0 || probe_addr != 0x0) {
+ printf("FAIL: bpf_trace_event_query(event_fd[%d]):\n",
+ prog_fd_idx);
+ printf("buf: %s, fd_type: %u, probe_offset: 0x%llx,"
+ " probe_addr: 0x%llx\n",
+ buf, fd_type, probe_offset, probe_addr);
+ return -1;
+ }
+ return 0;
+}
+
+static int test_nondebug_fs_kuprobe_common(const char *event_type,
+ const char *name, __u64 offset, __u64 addr, bool is_return,
+ char *buf, __u32 *buf_len, __u32 *prog_id, __u32 *fd_type,
+ __u64 *probe_offset, __u64 *probe_addr)
+{
+ int is_return_bit = bpf_get_retprobe_bit(event_type);
+ int type = bpf_find_probe_type(event_type);
+ struct perf_event_attr attr = {};
+ int fd;
+
+ if (type < 0 || is_return_bit < 0) {
+ printf("FAIL: %s incorrect type (%d) or is_return_bit (%d)\n",
+ __func__, type, is_return_bit);
+ return -1;
+ }
+
+ attr.sample_period = 1;
+ attr.wakeup_events = 1;
+ if (is_return)
+ attr.config |= 1 << is_return_bit;
+
+ if (name) {
+ attr.config1 = ptr_to_u64((void *)name);
+ attr.config2 = offset;
+ } else {
+ attr.config1 = 0;
+ attr.config2 = addr;
+ }
+ attr.size = sizeof(attr);
+ attr.type = type;
+
+ fd = sys_perf_event_open(&attr, -1, 0, -1, 0);
+ CHECK_PERROR_RET(fd < 0);
+
+ CHECK_PERROR_RET(ioctl(fd, PERF_EVENT_IOC_ENABLE, 0) < 0);
+ CHECK_PERROR_RET(ioctl(fd, PERF_EVENT_IOC_SET_BPF, prog_fd[0]) < 0);
+ CHECK_PERROR_RET(bpf_task_fd_query(getpid(), fd, 0, buf, buf_len,
+ prog_id, fd_type, probe_offset, probe_addr) < 0);
+
+ return 0;
+}
+
+static int test_nondebug_fs_probe(const char *event_type, const char *name,
+ __u64 offset, __u64 addr, bool is_return,
+ __u32 expected_fd_type,
+ __u32 expected_ret_fd_type,
+ char *buf, __u32 buf_len)
+{
+ __u64 probe_offset, probe_addr;
+ __u32 prog_id, fd_type;
+ int err;
+
+ err = test_nondebug_fs_kuprobe_common(event_type, name,
+ offset, addr, is_return,
+ buf, &buf_len, &prog_id,
+ &fd_type, &probe_offset,
+ &probe_addr);
+ if (err < 0) {
+ printf("FAIL: %s, "
+ "for name %s, offset 0x%llx, addr 0x%llx, is_return %d\n",
+ __func__, name ? name : "", offset, addr, is_return);
+ perror(" :");
+ return -1;
+ }
+ if ((is_return && fd_type != expected_ret_fd_type) ||
+ (!is_return && fd_type != expected_fd_type)) {
+ printf("FAIL: %s, incorrect fd_type %u\n",
+ __func__, fd_type);
+ return -1;
+ }
+ if (name) {
+ if (strcmp(name, buf) != 0) {
+ printf("FAIL: %s, incorrect buf %s\n", __func__, buf);
+ return -1;
+ }
+ if (probe_offset != offset) {
+ printf("FAIL: %s, incorrect probe_offset 0x%llx\n",
+ __func__, probe_offset);
+ return -1;
+ }
+ } else {
+ if (buf_len != 0) {
+ printf("FAIL: %s, incorrect buf %p\n",
+ __func__, buf);
+ return -1;
+ }
+
+ if (probe_addr != addr) {
+ printf("FAIL: %s, incorrect probe_addr 0x%llx\n",
+ __func__, probe_addr);
+ return -1;
+ }
+ }
+ return 0;
+}
+
+static int test_debug_fs_uprobe(char *binary_path, long offset, bool is_return)
+{
+ const char *event_type = "uprobe";
+ struct perf_event_attr attr = {};
+ char buf[256], event_alias[256];
+ __u64 probe_offset, probe_addr;
+ __u32 len, prog_id, fd_type;
+ int err, res, kfd, efd;
+ ssize_t bytes;
+
+ snprintf(buf, sizeof(buf), "/sys/kernel/debug/tracing/%s_events",
+ event_type);
+ kfd = open(buf, O_WRONLY | O_APPEND, 0);
+ CHECK_PERROR_RET(kfd < 0);
+
+ res = snprintf(event_alias, sizeof(event_alias), "test_%d", getpid());
+ CHECK_PERROR_RET(res < 0 || res >= sizeof(event_alias));
+
+ res = snprintf(buf, sizeof(buf), "%c:%ss/%s %s:0x%lx",
+ is_return ? 'r' : 'p', event_type, event_alias,
+ binary_path, offset);
+ CHECK_PERROR_RET(res < 0 || res >= sizeof(buf));
+ CHECK_PERROR_RET(write(kfd, buf, strlen(buf)) < 0);
+
+ close(kfd);
+ kfd = -1;
+
+ snprintf(buf, sizeof(buf), "/sys/kernel/debug/tracing/events/%ss/%s/id",
+ event_type, event_alias);
+ efd = open(buf, O_RDONLY, 0);
+ CHECK_PERROR_RET(efd < 0);
+
+ bytes = read(efd, buf, sizeof(buf));
+ CHECK_PERROR_RET(bytes <= 0 || bytes >= sizeof(buf));
+ close(efd);
+ buf[bytes] = '\0';
+
+ attr.config = strtol(buf, NULL, 0);
+ attr.type = PERF_TYPE_TRACEPOINT;
+ attr.sample_period = 1;
+ attr.wakeup_events = 1;
+ kfd = sys_perf_event_open(&attr, -1, 0, -1, PERF_FLAG_FD_CLOEXEC);
+ CHECK_PERROR_RET(kfd < 0);
+ CHECK_PERROR_RET(ioctl(kfd, PERF_EVENT_IOC_SET_BPF, prog_fd[0]) < 0);
+ CHECK_PERROR_RET(ioctl(kfd, PERF_EVENT_IOC_ENABLE, 0) < 0);
+
+ len = sizeof(buf);
+ err = bpf_task_fd_query(getpid(), kfd, 0, buf, &len,
+ &prog_id, &fd_type, &probe_offset,
+ &probe_addr);
+ if (err < 0) {
+ printf("FAIL: %s, binary_path %s\n", __func__, binary_path);
+ perror(" :");
+ return -1;
+ }
+ if ((is_return && fd_type != BPF_FD_TYPE_URETPROBE) ||
+ (!is_return && fd_type != BPF_FD_TYPE_UPROBE)) {
+ printf("FAIL: %s, incorrect fd_type %u\n", __func__,
+ fd_type);
+ return -1;
+ }
+ if (strcmp(binary_path, buf) != 0) {
+ printf("FAIL: %s, incorrect buf %s\n", __func__, buf);
+ return -1;
+ }
+ if (probe_offset != offset) {
+ printf("FAIL: %s, incorrect probe_offset 0x%llx\n", __func__,
+ probe_offset);
+ return -1;
+ }
+
+ close(kfd);
+ return 0;
+}
+
+int main(int argc, char **argv)
+{
+ struct rlimit r = {1024*1024, RLIM_INFINITY};
+ extern char __executable_start;
+ char filename[256], buf[256];
+ __u64 uprobe_file_offset;
+
+ snprintf(filename, sizeof(filename), "%s_kern.o", argv[0]);
+ if (setrlimit(RLIMIT_MEMLOCK, &r)) {
+ perror("setrlimit(RLIMIT_MEMLOCK)");
+ return 1;
+ }
+
+ if (load_kallsyms()) {
+ printf("failed to process /proc/kallsyms\n");
+ return 1;
+ }
+
+ if (load_bpf_file(filename)) {
+ printf("%s", bpf_log_buf);
+ return 1;
+ }
+
+ /* test two functions in the corresponding *_kern.c file */
+ CHECK_AND_RET(test_debug_fs_kprobe(0, "blk_start_request",
+ BPF_FD_TYPE_KPROBE));
+ CHECK_AND_RET(test_debug_fs_kprobe(1, "blk_account_io_completion",
+ BPF_FD_TYPE_KRETPROBE));
+
+ /* test nondebug fs kprobe */
+ CHECK_AND_RET(test_nondebug_fs_probe("kprobe", "bpf_check", 0x0, 0x0,
+ false, BPF_FD_TYPE_KPROBE,
+ BPF_FD_TYPE_KRETPROBE,
+ buf, sizeof(buf)));
+#ifdef __x86_64__
+ /* set a kprobe on "bpf_check + 0x5", which is x64 specific */
+ CHECK_AND_RET(test_nondebug_fs_probe("kprobe", "bpf_check", 0x5, 0x0,
+ false, BPF_FD_TYPE_KPROBE,
+ BPF_FD_TYPE_KRETPROBE,
+ buf, sizeof(buf)));
+#endif
+ CHECK_AND_RET(test_nondebug_fs_probe("kprobe", "bpf_check", 0x0, 0x0,
+ true, BPF_FD_TYPE_KPROBE,
+ BPF_FD_TYPE_KRETPROBE,
+ buf, sizeof(buf)));
+ CHECK_AND_RET(test_nondebug_fs_probe("kprobe", NULL, 0x0,
+ ksym_get_addr("bpf_check"), false,
+ BPF_FD_TYPE_KPROBE,
+ BPF_FD_TYPE_KRETPROBE,
+ buf, sizeof(buf)));
+ CHECK_AND_RET(test_nondebug_fs_probe("kprobe", NULL, 0x0,
+ ksym_get_addr("bpf_check"), false,
+ BPF_FD_TYPE_KPROBE,
+ BPF_FD_TYPE_KRETPROBE,
+ NULL, 0));
+ CHECK_AND_RET(test_nondebug_fs_probe("kprobe", NULL, 0x0,
+ ksym_get_addr("bpf_check"), true,
+ BPF_FD_TYPE_KPROBE,
+ BPF_FD_TYPE_KRETPROBE,
+ buf, sizeof(buf)));
+ CHECK_AND_RET(test_nondebug_fs_probe("kprobe", NULL, 0x0,
+ ksym_get_addr("bpf_check"), true,
+ BPF_FD_TYPE_KPROBE,
+ BPF_FD_TYPE_KRETPROBE,
+ 0, 0));
+
+ /* test nondebug fs uprobe */
+ /* the calculation of uprobe file offset is based on gcc 7.3.1 on x64
+ * and the default linker script, which defines __executable_start as
+ * the start of the .text section. The calculation could be different
+ * on different systems with different compilers. The right way is
+ * to parse the ELF file. We took a shortcut here.
+ */
+ uprobe_file_offset = (__u64)main - (__u64)&__executable_start;
+ CHECK_AND_RET(test_nondebug_fs_probe("uprobe", (char *)argv[0],
+ uprobe_file_offset, 0x0, false,
+ BPF_FD_TYPE_UPROBE,
+ BPF_FD_TYPE_URETPROBE,
+ buf, sizeof(buf)));
+ CHECK_AND_RET(test_nondebug_fs_probe("uprobe", (char *)argv[0],
+ uprobe_file_offset, 0x0, true,
+ BPF_FD_TYPE_UPROBE,
+ BPF_FD_TYPE_URETPROBE,
+ buf, sizeof(buf)));
+
+ /* test debug fs uprobe */
+ CHECK_AND_RET(test_debug_fs_uprobe((char *)argv[0], uprobe_file_offset,
+ false));
+ CHECK_AND_RET(test_debug_fs_uprobe((char *)argv[0], uprobe_file_offset,
+ true));
+
+ return 0;
+}
--
2.9.5
^ permalink raw reply related
* [PATCH bpf-next v5 3/7] tools/bpf: sync kernel header bpf.h and add bpf_task_fd_query in libbpf
From: Yonghong Song @ 2018-05-24 18:21 UTC (permalink / raw)
To: peterz, ast, daniel, netdev; +Cc: kernel-team
In-Reply-To: <20180524182111.454612-1-yhs@fb.com>
Sync kernel header bpf.h to tools/include/uapi/linux/bpf.h and
implement bpf_task_fd_query() in libbpf. The test programs
in samples/bpf and tools/testing/selftests/bpf, and later bpftool
will use this libbpf function to query kernel.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
---
tools/include/uapi/linux/bpf.h | 26 ++++++++++++++++++++++++++
tools/lib/bpf/bpf.c | 23 +++++++++++++++++++++++
tools/lib/bpf/bpf.h | 3 +++
3 files changed, 52 insertions(+)
diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h
index e95fec9..9b8c6e3 100644
--- a/tools/include/uapi/linux/bpf.h
+++ b/tools/include/uapi/linux/bpf.h
@@ -97,6 +97,7 @@ enum bpf_cmd {
BPF_RAW_TRACEPOINT_OPEN,
BPF_BTF_LOAD,
BPF_BTF_GET_FD_BY_ID,
+ BPF_TASK_FD_QUERY,
};
enum bpf_map_type {
@@ -380,6 +381,22 @@ union bpf_attr {
__u32 btf_log_size;
__u32 btf_log_level;
};
+
+ struct {
+ __u32 pid; /* input: pid */
+ __u32 fd; /* input: fd */
+ __u32 flags; /* input: flags */
+ __u32 buf_len; /* input/output: buf len */
+ __aligned_u64 buf; /* input/output:
+ * tp_name for tracepoint
+ * symbol for kprobe
+ * filename for uprobe
+ */
+ __u32 prog_id; /* output: prod_id */
+ __u32 fd_type; /* output: BPF_FD_TYPE_* */
+ __u64 probe_offset; /* output: probe_offset */
+ __u64 probe_addr; /* output: probe_addr */
+ } task_fd_query;
} __attribute__((aligned(8)));
/* The description below is an attempt at providing documentation to eBPF
@@ -2557,4 +2574,13 @@ struct bpf_fib_lookup {
__u8 dmac[6]; /* ETH_ALEN */
};
+enum bpf_task_fd_type {
+ BPF_FD_TYPE_RAW_TRACEPOINT, /* tp name */
+ BPF_FD_TYPE_TRACEPOINT, /* tp name */
+ BPF_FD_TYPE_KPROBE, /* (symbol + offset) or addr */
+ BPF_FD_TYPE_KRETPROBE, /* (symbol + offset) or addr */
+ BPF_FD_TYPE_UPROBE, /* filename + offset */
+ BPF_FD_TYPE_URETPROBE, /* filename + offset */
+};
+
#endif /* _UAPI__LINUX_BPF_H__ */
diff --git a/tools/lib/bpf/bpf.c b/tools/lib/bpf/bpf.c
index 442b4cd..9ddc89d 100644
--- a/tools/lib/bpf/bpf.c
+++ b/tools/lib/bpf/bpf.c
@@ -643,3 +643,26 @@ int bpf_load_btf(void *btf, __u32 btf_size, char *log_buf, __u32 log_buf_size,
return fd;
}
+
+int bpf_task_fd_query(int pid, int fd, __u32 flags, char *buf, __u32 *buf_len,
+ __u32 *prog_id, __u32 *fd_type, __u64 *probe_offset,
+ __u64 *probe_addr)
+{
+ union bpf_attr attr = {};
+ int err;
+
+ attr.task_fd_query.pid = pid;
+ attr.task_fd_query.fd = fd;
+ attr.task_fd_query.flags = flags;
+ attr.task_fd_query.buf = ptr_to_u64(buf);
+ attr.task_fd_query.buf_len = *buf_len;
+
+ err = sys_bpf(BPF_TASK_FD_QUERY, &attr, sizeof(attr));
+ *buf_len = attr.task_fd_query.buf_len;
+ *prog_id = attr.task_fd_query.prog_id;
+ *fd_type = attr.task_fd_query.fd_type;
+ *probe_offset = attr.task_fd_query.probe_offset;
+ *probe_addr = attr.task_fd_query.probe_addr;
+
+ return err;
+}
diff --git a/tools/lib/bpf/bpf.h b/tools/lib/bpf/bpf.h
index d12344f..0639a30 100644
--- a/tools/lib/bpf/bpf.h
+++ b/tools/lib/bpf/bpf.h
@@ -107,4 +107,7 @@ int bpf_prog_query(int target_fd, enum bpf_attach_type type, __u32 query_flags,
int bpf_raw_tracepoint_open(const char *name, int prog_fd);
int bpf_load_btf(void *btf, __u32 btf_size, char *log_buf, __u32 log_buf_size,
bool do_log);
+int bpf_task_fd_query(int pid, int fd, __u32 flags, char *buf, __u32 *buf_len,
+ __u32 *prog_id, __u32 *fd_type, __u64 *probe_offset,
+ __u64 *probe_addr);
#endif
--
2.9.5
^ permalink raw reply related
* [PATCH bpf-next v5 4/7] tools/bpf: add ksym_get_addr() in trace_helpers
From: Yonghong Song @ 2018-05-24 18:21 UTC (permalink / raw)
To: peterz, ast, daniel, netdev; +Cc: kernel-team
In-Reply-To: <20180524182111.454612-1-yhs@fb.com>
Given a kernel function name, ksym_get_addr() will return the kernel
address for this function, or 0 if it cannot find this function name
in /proc/kallsyms. This function will be used later when a kernel
address is used to initiate a kprobe perf event.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
---
tools/testing/selftests/bpf/trace_helpers.c | 12 ++++++++++++
tools/testing/selftests/bpf/trace_helpers.h | 1 +
2 files changed, 13 insertions(+)
diff --git a/tools/testing/selftests/bpf/trace_helpers.c b/tools/testing/selftests/bpf/trace_helpers.c
index 8fb4fe8..3868dcb 100644
--- a/tools/testing/selftests/bpf/trace_helpers.c
+++ b/tools/testing/selftests/bpf/trace_helpers.c
@@ -72,6 +72,18 @@ struct ksym *ksym_search(long key)
return &syms[0];
}
+long ksym_get_addr(const char *name)
+{
+ int i;
+
+ for (i = 0; i < sym_cnt; i++) {
+ if (strcmp(syms[i].name, name) == 0)
+ return syms[i].addr;
+ }
+
+ return 0;
+}
+
static int page_size;
static int page_cnt = 8;
static struct perf_event_mmap_page *header;
diff --git a/tools/testing/selftests/bpf/trace_helpers.h b/tools/testing/selftests/bpf/trace_helpers.h
index 36d90e3..3b4bcf7 100644
--- a/tools/testing/selftests/bpf/trace_helpers.h
+++ b/tools/testing/selftests/bpf/trace_helpers.h
@@ -11,6 +11,7 @@ struct ksym {
int load_kallsyms(void);
struct ksym *ksym_search(long key);
+long ksym_get_addr(const char *name);
typedef enum bpf_perf_event_ret (*perf_event_print_fn)(void *data, int size);
--
2.9.5
^ permalink raw reply related
* [PATCH bpf-next v5 0/7] bpf: implement BPF_TASK_FD_QUERY
From: Yonghong Song @ 2018-05-24 18:21 UTC (permalink / raw)
To: peterz, ast, daniel, netdev; +Cc: kernel-team
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, this command will return bpf related information
to user space. Right now it only supports tracepoint/kprobe/uprobe
perf event fd's. For such a fd, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Patch #1 adds function perf_get_event() in kernel/events/core.c.
Patch #2 implements the bpf subcommand BPF_TASK_FD_QUERY.
Patch #3 syncs tools bpf.h header and also add bpf_task_fd_query()
in the libbpf library for samples/selftests/bpftool to use.
Patch #4 adds ksym_get_addr() utility function.
Patch #5 add a test in samples/bpf for querying k[ret]probes and
u[ret]probes.
Patch #6 add a test in tools/testing/selftests/bpf for querying
raw_tracepoint and tracepoint.
Patch #7 add a new subcommand "perf" to bpftool.
Changelogs:
v4 -> v5:
. return strlen(buf) instead of strlen(buf) + 1
in the attr.buf_len. As long as user provides
non-empty buffer, it will be filed with empty
string, truncated string, or full string
based on the buffer size and the length of
to-be-copied string.
v3 -> v4:
. made attr buf_len input/output. The length of
actual buffter is written to buf_len so user space knows
what is actually needed. If user provides a buffer
with length >= 1 but less than required, do partial
copy and return -ENOSPC.
. code simplification with put_user.
. changed query result attach_info to fd_type.
. add tests at selftests/bpf to test zero len, null buf and
insufficient buf.
v2 -> v3:
. made perf_get_event() return perf_event pointer const.
this was to ensure that event fields are not meddled.
. detect whether newly BPF_TASK_FD_QUERY is supported or
not in "bpftool perf" and warn users if it is not.
v1 -> v2:
. changed bpf subcommand name from BPF_PERF_EVENT_QUERY
to BPF_TASK_FD_QUERY.
. fixed various "bpftool perf" issues and added documentation
and auto-completion.
Yonghong Song (7):
perf/core: add perf_get_event() to return perf_event given a struct
file
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
tools/bpf: sync kernel header bpf.h and add bpf_task_fd_query in
libbpf
tools/bpf: add ksym_get_addr() in trace_helpers
samples/bpf: add a samples/bpf test for BPF_TASK_FD_QUERY
tools/bpf: add two BPF_TASK_FD_QUERY tests in test_progs
tools/bpftool: add perf subcommand
include/linux/perf_event.h | 5 +
include/linux/trace_events.h | 17 +
include/uapi/linux/bpf.h | 26 ++
kernel/bpf/syscall.c | 131 ++++++++
kernel/events/core.c | 8 +
kernel/trace/bpf_trace.c | 48 +++
kernel/trace/trace_kprobe.c | 29 ++
kernel/trace/trace_uprobe.c | 22 ++
samples/bpf/Makefile | 4 +
samples/bpf/task_fd_query_kern.c | 19 ++
samples/bpf/task_fd_query_user.c | 382 +++++++++++++++++++++++
tools/bpf/bpftool/Documentation/bpftool-perf.rst | 81 +++++
tools/bpf/bpftool/Documentation/bpftool.rst | 5 +-
tools/bpf/bpftool/bash-completion/bpftool | 9 +
tools/bpf/bpftool/main.c | 3 +-
tools/bpf/bpftool/main.h | 1 +
tools/bpf/bpftool/perf.c | 246 +++++++++++++++
tools/include/uapi/linux/bpf.h | 26 ++
tools/lib/bpf/bpf.c | 23 ++
tools/lib/bpf/bpf.h | 3 +
tools/testing/selftests/bpf/test_progs.c | 158 ++++++++++
tools/testing/selftests/bpf/trace_helpers.c | 12 +
tools/testing/selftests/bpf/trace_helpers.h | 1 +
23 files changed, 1257 insertions(+), 2 deletions(-)
create mode 100644 samples/bpf/task_fd_query_kern.c
create mode 100644 samples/bpf/task_fd_query_user.c
create mode 100644 tools/bpf/bpftool/Documentation/bpftool-perf.rst
create mode 100644 tools/bpf/bpftool/perf.c
--
2.9.5
^ permalink raw reply
* [PATCH bpf-next v5 2/7] bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
From: Yonghong Song @ 2018-05-24 18:21 UTC (permalink / raw)
To: peterz, ast, daniel, netdev; +Cc: kernel-team
In-Reply-To: <20180524182111.454612-1-yhs@fb.com>
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
---
include/linux/trace_events.h | 17 ++++++
include/uapi/linux/bpf.h | 26 +++++++++
kernel/bpf/syscall.c | 131 +++++++++++++++++++++++++++++++++++++++++++
kernel/trace/bpf_trace.c | 48 ++++++++++++++++
kernel/trace/trace_kprobe.c | 29 ++++++++++
kernel/trace/trace_uprobe.c | 22 ++++++++
6 files changed, 273 insertions(+)
diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
index 2bde3ef..d34144a 100644
--- a/include/linux/trace_events.h
+++ b/include/linux/trace_events.h
@@ -473,6 +473,9 @@ int perf_event_query_prog_array(struct perf_event *event, void __user *info);
int bpf_probe_register(struct bpf_raw_event_map *btp, struct bpf_prog *prog);
int bpf_probe_unregister(struct bpf_raw_event_map *btp, struct bpf_prog *prog);
struct bpf_raw_event_map *bpf_find_raw_tracepoint(const char *name);
+int bpf_get_perf_event_info(const struct perf_event *event, u32 *prog_id,
+ u32 *fd_type, const char **buf,
+ u64 *probe_offset, u64 *probe_addr);
#else
static inline unsigned int trace_call_bpf(struct trace_event_call *call, void *ctx)
{
@@ -504,6 +507,13 @@ static inline struct bpf_raw_event_map *bpf_find_raw_tracepoint(const char *name
{
return NULL;
}
+static inline int bpf_get_perf_event_info(const struct perf_event *event,
+ u32 *prog_id, u32 *fd_type,
+ const char **buf, u64 *probe_offset,
+ u64 *probe_addr)
+{
+ return -EOPNOTSUPP;
+}
#endif
enum {
@@ -560,10 +570,17 @@ extern void perf_trace_del(struct perf_event *event, int flags);
#ifdef CONFIG_KPROBE_EVENTS
extern int perf_kprobe_init(struct perf_event *event, bool is_retprobe);
extern void perf_kprobe_destroy(struct perf_event *event);
+extern int bpf_get_kprobe_info(const struct perf_event *event,
+ u32 *fd_type, const char **symbol,
+ u64 *probe_offset, u64 *probe_addr,
+ bool perf_type_tracepoint);
#endif
#ifdef CONFIG_UPROBE_EVENTS
extern int perf_uprobe_init(struct perf_event *event, bool is_retprobe);
extern void perf_uprobe_destroy(struct perf_event *event);
+extern int bpf_get_uprobe_info(const struct perf_event *event,
+ u32 *fd_type, const char **filename,
+ u64 *probe_offset, bool perf_type_tracepoint);
#endif
extern int ftrace_profile_set_filter(struct perf_event *event, int event_id,
char *filter_str);
diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
index e95fec9..9b8c6e3 100644
--- a/include/uapi/linux/bpf.h
+++ b/include/uapi/linux/bpf.h
@@ -97,6 +97,7 @@ enum bpf_cmd {
BPF_RAW_TRACEPOINT_OPEN,
BPF_BTF_LOAD,
BPF_BTF_GET_FD_BY_ID,
+ BPF_TASK_FD_QUERY,
};
enum bpf_map_type {
@@ -380,6 +381,22 @@ union bpf_attr {
__u32 btf_log_size;
__u32 btf_log_level;
};
+
+ struct {
+ __u32 pid; /* input: pid */
+ __u32 fd; /* input: fd */
+ __u32 flags; /* input: flags */
+ __u32 buf_len; /* input/output: buf len */
+ __aligned_u64 buf; /* input/output:
+ * tp_name for tracepoint
+ * symbol for kprobe
+ * filename for uprobe
+ */
+ __u32 prog_id; /* output: prod_id */
+ __u32 fd_type; /* output: BPF_FD_TYPE_* */
+ __u64 probe_offset; /* output: probe_offset */
+ __u64 probe_addr; /* output: probe_addr */
+ } task_fd_query;
} __attribute__((aligned(8)));
/* The description below is an attempt at providing documentation to eBPF
@@ -2557,4 +2574,13 @@ struct bpf_fib_lookup {
__u8 dmac[6]; /* ETH_ALEN */
};
+enum bpf_task_fd_type {
+ BPF_FD_TYPE_RAW_TRACEPOINT, /* tp name */
+ BPF_FD_TYPE_TRACEPOINT, /* tp name */
+ BPF_FD_TYPE_KPROBE, /* (symbol + offset) or addr */
+ BPF_FD_TYPE_KRETPROBE, /* (symbol + offset) or addr */
+ BPF_FD_TYPE_UPROBE, /* filename + offset */
+ BPF_FD_TYPE_URETPROBE, /* filename + offset */
+};
+
#endif /* _UAPI__LINUX_BPF_H__ */
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 788456c..388d4fe 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -18,7 +18,9 @@
#include <linux/vmalloc.h>
#include <linux/mmzone.h>
#include <linux/anon_inodes.h>
+#include <linux/fdtable.h>
#include <linux/file.h>
+#include <linux/fs.h>
#include <linux/license.h>
#include <linux/filter.h>
#include <linux/version.h>
@@ -2178,6 +2180,132 @@ static int bpf_btf_get_fd_by_id(const union bpf_attr *attr)
return btf_get_fd_by_id(attr->btf_id);
}
+static int bpf_task_fd_query_copy(const union bpf_attr *attr,
+ union bpf_attr __user *uattr,
+ u32 prog_id, u32 fd_type,
+ const char *buf, u64 probe_offset,
+ u64 probe_addr)
+{
+ char __user *ubuf = u64_to_user_ptr(attr->task_fd_query.buf);
+ u32 len = buf ? strlen(buf) : 0, input_len;
+ int err = 0;
+
+ if (put_user(len, &uattr->task_fd_query.buf_len))
+ return -EFAULT;
+ input_len = attr->task_fd_query.buf_len;
+ if (input_len && ubuf) {
+ if (!len) {
+ /* nothing to copy, just make ubuf NULL terminated */
+ char zero = '\0';
+
+ if (put_user(zero, ubuf))
+ return -EFAULT;
+ } else if (input_len >= len + 1) {
+ /* ubuf can hold the string with NULL terminator */
+ if (copy_to_user(ubuf, buf, len + 1))
+ return -EFAULT;
+ } else {
+ /* ubuf cannot hold the string with NULL terminator,
+ * do a partial copy with NULL terminator.
+ */
+ char zero = '\0';
+
+ err = -ENOSPC;
+ if (copy_to_user(ubuf, buf, input_len - 1))
+ return -EFAULT;
+ if (put_user(zero, ubuf + input_len - 1))
+ return -EFAULT;
+ }
+ }
+
+ if (put_user(prog_id, &uattr->task_fd_query.prog_id) ||
+ put_user(fd_type, &uattr->task_fd_query.fd_type) ||
+ put_user(probe_offset, &uattr->task_fd_query.probe_offset) ||
+ put_user(probe_addr, &uattr->task_fd_query.probe_addr))
+ return -EFAULT;
+
+ return err;
+}
+
+#define BPF_TASK_FD_QUERY_LAST_FIELD task_fd_query.probe_addr
+
+static int bpf_task_fd_query(const union bpf_attr *attr,
+ union bpf_attr __user *uattr)
+{
+ pid_t pid = attr->task_fd_query.pid;
+ u32 fd = attr->task_fd_query.fd;
+ const struct perf_event *event;
+ struct files_struct *files;
+ struct task_struct *task;
+ struct file *file;
+ int err;
+
+ if (CHECK_ATTR(BPF_TASK_FD_QUERY))
+ return -EINVAL;
+
+ if (!capable(CAP_SYS_ADMIN))
+ return -EPERM;
+
+ if (attr->task_fd_query.flags != 0)
+ return -EINVAL;
+
+ task = get_pid_task(find_vpid(pid), PIDTYPE_PID);
+ if (!task)
+ return -ENOENT;
+
+ files = get_files_struct(task);
+ put_task_struct(task);
+ if (!files)
+ return -ENOENT;
+
+ err = 0;
+ spin_lock(&files->file_lock);
+ file = fcheck_files(files, fd);
+ if (!file)
+ err = -EBADF;
+ else
+ get_file(file);
+ spin_unlock(&files->file_lock);
+ put_files_struct(files);
+
+ if (err)
+ goto out;
+
+ if (file->f_op == &bpf_raw_tp_fops) {
+ struct bpf_raw_tracepoint *raw_tp = file->private_data;
+ struct bpf_raw_event_map *btp = raw_tp->btp;
+
+ err = bpf_task_fd_query_copy(attr, uattr,
+ raw_tp->prog->aux->id,
+ BPF_FD_TYPE_RAW_TRACEPOINT,
+ btp->tp->name, 0, 0);
+ goto put_file;
+ }
+
+ event = perf_get_event(file);
+ if (!IS_ERR(event)) {
+ u64 probe_offset, probe_addr;
+ u32 prog_id, fd_type;
+ const char *buf;
+
+ err = bpf_get_perf_event_info(event, &prog_id, &fd_type,
+ &buf, &probe_offset,
+ &probe_addr);
+ if (!err)
+ err = bpf_task_fd_query_copy(attr, uattr, prog_id,
+ fd_type, buf,
+ probe_offset,
+ probe_addr);
+ goto put_file;
+ }
+
+ err = -ENOTSUPP;
+put_file:
+ fput(file);
+out:
+ return err;
+}
+
SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, size)
{
union bpf_attr attr = {};
@@ -2264,6 +2392,9 @@ SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, siz
case BPF_BTF_GET_FD_BY_ID:
err = bpf_btf_get_fd_by_id(&attr);
break;
+ case BPF_TASK_FD_QUERY:
+ err = bpf_task_fd_query(&attr, uattr);
+ break;
default:
err = -EINVAL;
break;
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index ce2cbbf..81fdf2f 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -14,6 +14,7 @@
#include <linux/uaccess.h>
#include <linux/ctype.h>
#include <linux/kprobes.h>
+#include <linux/syscalls.h>
#include <linux/error-injection.h>
#include "trace_probe.h"
@@ -1163,3 +1164,50 @@ int bpf_probe_unregister(struct bpf_raw_event_map *btp, struct bpf_prog *prog)
mutex_unlock(&bpf_event_mutex);
return err;
}
+
+int bpf_get_perf_event_info(const struct perf_event *event, u32 *prog_id,
+ u32 *fd_type, const char **buf,
+ u64 *probe_offset, u64 *probe_addr)
+{
+ bool is_tracepoint, is_syscall_tp;
+ struct bpf_prog *prog;
+ int flags, err = 0;
+
+ prog = event->prog;
+ if (!prog)
+ return -ENOENT;
+
+ /* not supporting BPF_PROG_TYPE_PERF_EVENT yet */
+ if (prog->type == BPF_PROG_TYPE_PERF_EVENT)
+ return -EOPNOTSUPP;
+
+ *prog_id = prog->aux->id;
+ flags = event->tp_event->flags;
+ is_tracepoint = flags & TRACE_EVENT_FL_TRACEPOINT;
+ is_syscall_tp = is_syscall_trace_event(event->tp_event);
+
+ if (is_tracepoint || is_syscall_tp) {
+ *buf = is_tracepoint ? event->tp_event->tp->name
+ : event->tp_event->name;
+ *fd_type = BPF_FD_TYPE_TRACEPOINT;
+ *probe_offset = 0x0;
+ *probe_addr = 0x0;
+ } else {
+ /* kprobe/uprobe */
+ err = -EOPNOTSUPP;
+#ifdef CONFIG_KPROBE_EVENTS
+ if (flags & TRACE_EVENT_FL_KPROBE)
+ err = bpf_get_kprobe_info(event, fd_type, buf,
+ probe_offset, probe_addr,
+ event->attr.type == PERF_TYPE_TRACEPOINT);
+#endif
+#ifdef CONFIG_UPROBE_EVENTS
+ if (flags & TRACE_EVENT_FL_UPROBE)
+ err = bpf_get_uprobe_info(event, fd_type, buf,
+ probe_offset,
+ event->attr.type == PERF_TYPE_TRACEPOINT);
+#endif
+ }
+
+ return err;
+}
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 02aed76..daa8157 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -1287,6 +1287,35 @@ kretprobe_perf_func(struct trace_kprobe *tk, struct kretprobe_instance *ri,
head, NULL);
}
NOKPROBE_SYMBOL(kretprobe_perf_func);
+
+int bpf_get_kprobe_info(const struct perf_event *event, u32 *fd_type,
+ const char **symbol, u64 *probe_offset,
+ u64 *probe_addr, bool perf_type_tracepoint)
+{
+ const char *pevent = trace_event_name(event->tp_event);
+ const char *group = event->tp_event->class->system;
+ struct trace_kprobe *tk;
+
+ if (perf_type_tracepoint)
+ tk = find_trace_kprobe(pevent, group);
+ else
+ tk = event->tp_event->data;
+ if (!tk)
+ return -EINVAL;
+
+ *fd_type = trace_kprobe_is_return(tk) ? BPF_FD_TYPE_KRETPROBE
+ : BPF_FD_TYPE_KPROBE;
+ if (tk->symbol) {
+ *symbol = tk->symbol;
+ *probe_offset = tk->rp.kp.offset;
+ *probe_addr = 0;
+ } else {
+ *symbol = NULL;
+ *probe_offset = 0;
+ *probe_addr = (unsigned long)tk->rp.kp.addr;
+ }
+ return 0;
+}
#endif /* CONFIG_PERF_EVENTS */
/*
diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c
index ac89287..bf89a51 100644
--- a/kernel/trace/trace_uprobe.c
+++ b/kernel/trace/trace_uprobe.c
@@ -1161,6 +1161,28 @@ static void uretprobe_perf_func(struct trace_uprobe *tu, unsigned long func,
{
__uprobe_perf_func(tu, func, regs, ucb, dsize);
}
+
+int bpf_get_uprobe_info(const struct perf_event *event, u32 *fd_type,
+ const char **filename, u64 *probe_offset,
+ bool perf_type_tracepoint)
+{
+ const char *pevent = trace_event_name(event->tp_event);
+ const char *group = event->tp_event->class->system;
+ struct trace_uprobe *tu;
+
+ if (perf_type_tracepoint)
+ tu = find_probe_event(pevent, group);
+ else
+ tu = event->tp_event->data;
+ if (!tu)
+ return -EINVAL;
+
+ *fd_type = is_ret_probe(tu) ? BPF_FD_TYPE_URETPROBE
+ : BPF_FD_TYPE_UPROBE;
+ *filename = tu->filename;
+ *probe_offset = tu->offset;
+ return 0;
+}
#endif /* CONFIG_PERF_EVENTS */
static int
--
2.9.5
^ permalink raw reply related
* [PATCH bpf-next v5 1/7] perf/core: add perf_get_event() to return perf_event given a struct file
From: Yonghong Song @ 2018-05-24 18:21 UTC (permalink / raw)
To: peterz, ast, daniel, netdev; +Cc: kernel-team
In-Reply-To: <20180524182111.454612-1-yhs@fb.com>
A new extern function, perf_get_event(), is added to return a perf event
given a struct file. This function will be used in later patches.
Signed-off-by: Yonghong Song <yhs@fb.com>
---
include/linux/perf_event.h | 5 +++++
kernel/events/core.c | 8 ++++++++
2 files changed, 13 insertions(+)
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index e71e99e..eec302b 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -868,6 +868,7 @@ extern void perf_event_exit_task(struct task_struct *child);
extern void perf_event_free_task(struct task_struct *task);
extern void perf_event_delayed_put(struct task_struct *task);
extern struct file *perf_event_get(unsigned int fd);
+extern const struct perf_event *perf_get_event(struct file *file);
extern const struct perf_event_attr *perf_event_attrs(struct perf_event *event);
extern void perf_event_print_debug(void);
extern void perf_pmu_disable(struct pmu *pmu);
@@ -1289,6 +1290,10 @@ static inline void perf_event_exit_task(struct task_struct *child) { }
static inline void perf_event_free_task(struct task_struct *task) { }
static inline void perf_event_delayed_put(struct task_struct *task) { }
static inline struct file *perf_event_get(unsigned int fd) { return ERR_PTR(-EINVAL); }
+static inline const struct perf_event *perf_get_event(struct file *file)
+{
+ return ERR_PTR(-EINVAL);
+}
static inline const struct perf_event_attr *perf_event_attrs(struct perf_event *event)
{
return ERR_PTR(-EINVAL);
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 67612ce..6eeab86 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -11212,6 +11212,14 @@ struct file *perf_event_get(unsigned int fd)
return file;
}
+const struct perf_event *perf_get_event(struct file *file)
+{
+ if (file->f_op != &perf_fops)
+ return ERR_PTR(-EINVAL);
+
+ return file->private_data;
+}
+
const struct perf_event_attr *perf_event_attrs(struct perf_event *event)
{
if (!event)
--
2.9.5
^ permalink raw reply related
* Re: [PATCH bpf-next v3 01/15] net: initial AF_XDP skeleton
From: Alexei Starovoitov @ 2018-05-24 17:57 UTC (permalink / raw)
To: Stephen Hemminger
Cc: Björn Töpel, magnus.karlsson, alexander.h.duyck,
alexander.duyck, john.fastabend, ast, brouer,
willemdebruijn.kernel, daniel, mst, netdev, Björn Töpel,
michael.lundkvist, jesse.brandeburg, anjali.singhai, qi.z.zhang
In-Reply-To: <20180523155047.6c136279@xeon-e3>
On Wed, May 23, 2018 at 03:50:47PM -0700, Stephen Hemminger wrote:
> Most distributions will want it to be a module so that it is not loaded
> unless used, and AF_XDP could be also be disabled by blacklisting the module.
I think the opposite will be the case. Anyone who cares about performance
would want AF_XDP code to be builtin, since builtin vs module gives additional
performance. All our NIC drivers are builtin, since we see noticeable
perf gains on production workloads.
Hence I'd rather see us spending time on improving AF_XDP instead
of making it a module and forever struggling with maintaining it as a module.
More so I think it's time to get rid of IPV6=m for good. The kernel
is full of ugly hacks and performance degradation due to indirect calls
just because IPV6=m is still supported.
Folks that care about vmlinux size should be using kconfig to compile it out.
^ permalink raw reply
* Re: [PATCH net-next] tcp: use data length instead of skb->len in tcp_probe
From: Song Liu @ 2018-05-24 17:44 UTC (permalink / raw)
To: Yafang Shao
Cc: davem@davemloft.net, netdev@vger.kernel.org,
linux-kernel@vger.kernel.org
In-Reply-To: <1527166097-10908-1-git-send-email-laoar.shao@gmail.com>
> On May 24, 2018, at 5:48 AM, Yafang Shao <laoar.shao@gmail.com> wrote:
>
> skb->len is meaningless to user.
> data length could be more helpful, with which we can easily filter out
> the packet without payload.
>
> Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
> ---
> include/trace/events/tcp.h | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/include/trace/events/tcp.h b/include/trace/events/tcp.h
> index c1a5284..259b991 100644
> --- a/include/trace/events/tcp.h
> +++ b/include/trace/events/tcp.h
> @@ -261,7 +261,7 @@
> __entry->dport = ntohs(inet->inet_dport);
> __entry->mark = skb->mark;
>
> - __entry->length = skb->len;
> + __entry->length = skb->len - tcp_hdrlen(skb);
We should also rename __entry->length to __entry->data_len, so that whoever
using this field will notice the change.
Thanks,
Song
> __entry->snd_nxt = tp->snd_nxt;
> __entry->snd_una = tp->snd_una;
> __entry->snd_cwnd = tp->snd_cwnd;
> @@ -272,7 +272,7 @@
> __entry->sock_cookie = sock_gen_cookie(sk);
> ),
>
> - TP_printk("src=%pISpc dest=%pISpc mark=%#x length=%d snd_nxt=%#x snd_una=%#x snd_cwnd=%u ssthresh=%u snd_wnd=%u srtt=%u rcv_wnd=%u sock_cookie=%llx",
> + TP_printk("src=%pISpc dest=%pISpc mark=%#x data_len=%d snd_nxt=%#x snd_una=%#x snd_cwnd=%u ssthresh=%u snd_wnd=%u srtt=%u rcv_wnd=%u sock_cookie=%llx",
> __entry->saddr, __entry->daddr, __entry->mark,
> __entry->length, __entry->snd_nxt, __entry->snd_una,
> __entry->snd_cwnd, __entry->ssthresh, __entry->snd_wnd,
> --
> 1.8.3.1
>
^ permalink raw reply
* Re: [PATCH net-next 8/8] nfp: flower: compute link aggregation action
From: John Hurley @ 2018-05-24 17:36 UTC (permalink / raw)
To: Or Gerlitz; +Cc: Jakub Kicinski, David Miller, Linux Netdev List, oss-drivers
In-Reply-To: <CAJ3xEMiNn+k8b0ixYgL2M6X=7WLMM=i1gEoLcdEYkztjsphi=g@mail.gmail.com>
On Thu, May 24, 2018 at 6:09 PM, Or Gerlitz <gerlitz.or@gmail.com> wrote:
> On Thu, May 24, 2018 at 5:22 AM, Jakub Kicinski
> <jakub.kicinski@netronome.com> wrote:
>> From: John Hurley <john.hurley@netronome.com>
>>
>> If the egress device of an offloaded rule is a LAG port, then encode the
>> output port to the NFP with a LAG identifier and the offloaded group ID.
>>
>> A prelag action is also offloaded which must be the first action of the
>> series (although may appear after other pre-actions - e.g. tunnels). This
>> causes the FW to check that it has the necessary information to output to
>> the requested LAG port. If it does not, the packet is sent to the kernel
>> before any other actions are applied to it.
>
> Offload decision typically also looks if both devices have the same
> switchdev ID.
>
> In your case, do both reprs gets the same switchdev ID automatically when being
> put into the same team/bond instance? I wasn't sure to see here
> changes for that matter
Hi Or,
Yes, you are correct.
We essentially substituted the switchdev ID check for a repr app check.
So an app runs per card and spawns the reprs for that card - each repr
has a backpointer to its creating app.
When deciding if we can offload a bond, we ensure that all bond ports
are reprs and belong to the same app and so same card/switchdev_id.
^ permalink raw reply
* Re: [PATCH 0/6] ravb/sh_eth: fix sleep in atomic by reusing shared ethtool handlers
From: Sergei Shtylyov @ 2018-05-24 17:24 UTC (permalink / raw)
To: Vladimir Zapolskiy, David S. Miller; +Cc: netdev, linux-renesas-soc
In-Reply-To: <6ffa0c08-2210-379c-50f0-d5bf43ab569e@cogentembedded.com>
On 05/24/2018 07:40 PM, Sergei Shtylyov wrote:
>> For ages trivial changes to RAVB and SuperH ethernet links by means of
>> standard 'ethtool' trigger a 'sleeping function called from invalid
>> context' bug, to visualize it on r8a7795 ULCB:
>>
>> % ethtool -r eth0
>> BUG: sleeping function called from invalid context at kernel/locking/mutex.c:747
>> in_atomic(): 1, irqs_disabled(): 128, pid: 554, name: ethtool
>> INFO: lockdep is turned off.
>> irq event stamp: 0
>> hardirqs last enabled at (0): [<0000000000000000>] (null)
>> hardirqs last disabled at (0): [<ffff0000080e1d3c>] copy_process.isra.7.part.8+0x2cc/0x1918
>> softirqs last enabled at (0): [<ffff0000080e1d3c>] copy_process.isra.7.part.8+0x2cc/0x1918
>> softirqs last disabled at (0): [<0000000000000000>] (null)
>> CPU: 5 PID: 554 Comm: ethtool Not tainted 4.17.0-rc4-arm64-renesas+ #33
>> Hardware name: Renesas H3ULCB board based on r8a7795 ES2.0+ (DT)
>> Call trace:
>> dump_backtrace+0x0/0x198
>> show_stack+0x24/0x30
>> dump_stack+0xb8/0xf4
>> ___might_sleep+0x1c8/0x1f8
>> __might_sleep+0x58/0x90
>> __mutex_lock+0x50/0x890
>> mutex_lock_nested+0x3c/0x50
>> phy_start_aneg_priv+0x38/0x180
>> phy_start_aneg+0x24/0x30
>> ravb_nway_reset+0x3c/0x68
>> dev_ethtool+0x3dc/0x2338
>> dev_ioctl+0x19c/0x490
>> sock_do_ioctl+0xe0/0x238
>> sock_ioctl+0x254/0x460
>> do_vfs_ioctl+0xb0/0x918
>> ksys_ioctl+0x50/0x80
>> sys_ioctl+0x34/0x48
>> __sys_trace_return+0x0/0x4
>>
>> The root cause is that an attempt to modify ECMR and GECMR registers
>> only when RX/TX function is disabled was too overcomplicated in its
>> original implementation, also processing of an optional Link Change
>> interrupt added even more complexity, as a result the implementation
>> was error prone.
>>
>> The new locking scheme is confirmed to be correct by dumping driver
>> specific and generic PHY framework function calls with aid of ftrace
>> while running more or less advanced tests.
>>
>> Please note that sh_eth patches from the series were built-tested only.
>>
>> On purpose I do not add Fixes tags, the reused PHY handlers were added
>> way later than the fixed problems were firstly found in the drivers.
>
> I think you went one step too far with these fixes. On the first glance,
> the real fixes are to remove grabbing/releasing the spinlock for the duration
> of the phylib calls. Am I right? If so, making use of the new phylib APIs
> would be a further enhancement, it's not needed for fixing the splats per se...
Note that I hadn't looked at the patches #3/#6 at the time of writing this;
those seem to be more complicated than the rest.
MBR, Sergei
^ permalink raw reply
* Re: [PATCH 1/6] ravb: remove custom .nway_reset from ethtool ops
From: Andrew Lunn @ 2018-05-24 17:23 UTC (permalink / raw)
To: Sergei Shtylyov
Cc: Vladimir Zapolskiy, Vladimir Zapolskiy, David S. Miller, netdev,
linux-renesas-soc
In-Reply-To: <f6e0b8c5-7fb9-babe-0114-350e7d6b2186@cogentembedded.com>
> > For it to be unsafe, i think that would mean phylib would need to call
> > back into the MAC driver? The only way that could happen is via the
> > adjust_link call. And that will deadlock, since it takes the same
> > lock.
> >
> > Or am i/we missing something?
>
> It doesn't take any locks currently, only patches #3/#6 makes it do so...
Ah, yes.
You should not be holding any spinlocks when calling into phylib.
It does its own locking, which is mutex based.
The code in this patch is not touching the MAC, so looks safe to me.
Andrew
^ permalink raw reply
* Re: [net-next 1/6] net/dcb: Add dcbnl buffer attribute
From: Ido Schimmel @ 2018-05-24 17:13 UTC (permalink / raw)
To: Jakub Kicinski
Cc: Huy Nguyen, Saeed Mahameed, David S. Miller, netdev, Jiri Pirko,
Or Gerlitz, Parav Pandit, Ido Schimmel
In-Reply-To: <20180523022314.783e47fa@cakuba>
Hi Jakub,
On Wed, May 23, 2018 at 02:23:14AM -0700, Jakub Kicinski wrote:
> Are you referring to XOFF/XON thresholds? I don't think the "threshold
> type" in devlink API implies we are setting XON/XOFF thresholds
> directly :S If PFC is enabled we may be setting them indirectly,
> obviously.
>
> My understanding is that for static threshold type the size parameter
> specifies the max amount of memory given pool can consume.
Correct.
> Yes, we must have a different definitions of "shared buffer" :) That
> link, however, didn't clarify much for me... In mlx5 you seem to have a
> buffer which is shared between priorities, even if it's not what would
> be referred to as shared buffer in switch context.
The following link is my attempt at explaining the above concepts:
https://github.com/Mellanox/mlxsw/wiki/Quality-of-Service
Please let me know if something is not clear.
Basically, we use devlink-sb and dcbl to configure two different
buffers:
* devlink-sb is used to configure the switch's shared buffer which is
shared between all the ports and thus can't take a netdev as an handle
* dcbnl is used to configure per-port buffers (also called headroom
buffers) where received packets are stored while going through the
switch's pipeline before being admitted to the shared buffer and
awaiting transmission
Note that in Huy's case the buffers are of the second type (per-port)
and thus using dcbnl instead of devlink-sb makes sense.
> DCBNL seems to carry standard-based information, which this is not.
> mlxsw supports DCBNL, will it also support this buffer configuration
> mechanism?
I believe so, it's just a matter of doing the work. The hardware
supports this and the interface is identical to the NIC (same
registers).
> > > How does one query the total size of the buffer to be carved?
> > [HQN] This is not necessary. If the total size is too big, error will
> > be return via DCB netlink interface.
>
> Right, I'm not saying it's a bug :) It's just nice when user can be
> told the total size without having to probe for it :)
+1
^ permalink raw reply
* Re: [PATCH net-next 0/7] net: bridge: Notify about bridge VLANs
From: Florian Fainelli @ 2018-05-24 17:20 UTC (permalink / raw)
To: Petr Machata, netdev, devel, bridge
Cc: andrew, nikolay, gregkh, vivien.didelot, idosch, jiri,
razvan.stefanescu, davem
In-Reply-To: <cover.1527173527.git.petrm@mellanox.com>
Hi Petr,
On 05/24/2018 08:09 AM, Petr Machata wrote:
> In commit 946a11e7408e ("mlxsw: spectrum_span: Allow bridge for gretap
> mirror"), mlxsw got support for offloading mirror-to-gretap such that
> the underlay packet path involves a bridge. In that case, the offload is
> also influenced by PVID setting of said bridge. However, changes to VLAN
> configuration of the bridge itself do not generate switchdev
> notifications, so there's no mechanism to prod mlxsw to update the
> offload when these settings change.
>
> In this patchset, the problem is resolved by distributing the switchdev
> notification SWITCHDEV_OBJ_ID_PORT_VLAN also for configuration changes
> on bridge VLANs. Since stacked devices distribute the notification to
> lower devices, such event eventually reaches the driver, which can
> determine whether it's a bridge or port VLAN by inspecting orig_dev.
>
> To keep things consistent, the newly-distributed notifications observe
> the same protocol as the existing ones: dual prepare/commit, with
> -EOPNOTSUPP indicating lack of support, even though there's currently
> nothing to prepare for and nothing to support. Correspondingly, all
> switchdev drivers have been updated to return -EOPNOTSUPP for bridge
> VLAN notifications.
You seem to have approached the bridge changes a little differently from
this series:
https://lists.linux-foundation.org/pipermail/bridge/2016-November/010112.html
Both have the same intent that by targeting the bridge device itself,
you can propagate that through switchdev to the switch drivers, and in
turn create configurations where for instance, you have:
- CPU/management port present in specific VLANs that is a subset or
superset of the VLANs configured on front-panel ports
- CPU/management port tagged/untagged in specific VLANs which can be a
different setting from the front-panel ports
One problem we have in DSA at the moment is that we always add the CPU
port to the VLANs configured to the front-panel port but we do this with
the same attributes as the front panel ports! For instance, if you add
Port 0 to VLAN1 untagged, the the CPU port also gets added to that
VLAN1, also untagged. As long as there is just one VLAN untagged, this
is not much of a problem. Now do this with another VLAN or another port,
and the CPU can no longer differentiate the traffic from which VLAN it
is coming from, no bueno.
I had specifically changed b53 to always add the CPU port as tagged,
because that would always allow for differentiating traffic, but I would
rather have the capability to configure that at the bridge layer, which
you series seem to allow.
For the record, here is what the first commit in the series intended to
let an user do:
The following happens now (assuming bridge master device is already
created):
bridge vlan add vid 2 dev port0 pvid untagged
-> port0 (e.g: switch port 0) gets programmed
-> CPU port gets programmed
bridge vlan add vid 2 dev br0 self
-> CPU port gets programmed
bridge vlan add vid 2 dev port0
-> port0 (switch port 0) gets programmed
Are these use cases possible with your series? It seems to me like it is
if we drop the netif_is_bridge_master() checks and resolve orig_dev as
being a hint for the CPU/management port.
Thanks for reading me :)
>
> In patch #1, the code to send notifications for adding and deleting is
> factored out into two named functions.
>
> In patches #2-#5, respectively for mlxsw, rocker, DSA and DPAA2 ethsw,
> the new notifications (which are not enabled yet) are ignored to
> maintain the current behavior.
>
> In patch #6, the notification is actually enabled.
>
> In patch #7, mlxsw is changed to update offloads of mirror-to-gre also
> for bridge-related notifications.
>
> Petr Machata (7):
> net: bridge: Extract boilerplate around switchdev_port_obj_*()
> mlxsw: spectrum_switchdev: Ignore bridge VLAN events
> rocker: rocker_main: Ignore bridge VLAN events
> dsa: port: Ignore bridge VLAN events
> staging: fsl-dpaa2: ethsw: Ignore bridge VLAN events
> net: bridge: Notify about bridge VLANs
> mlxsw: spectrum_switchdev: Schedule respin during trans prepare
>
> .../ethernet/mellanox/mlxsw/spectrum_switchdev.c | 8 ++-
> drivers/net/ethernet/rocker/rocker_main.c | 6 +++
> drivers/staging/fsl-dpaa2/ethsw/ethsw.c | 6 +++
> net/bridge/br_vlan.c | 58 ++++++++++++++--------
> net/dsa/port.c | 6 +++
> 5 files changed, 62 insertions(+), 22 deletions(-)
>
--
Florian
^ permalink raw reply
* Re: [PATCH net-next 8/8] nfp: flower: compute link aggregation action
From: Or Gerlitz @ 2018-05-24 17:09 UTC (permalink / raw)
To: Jakub Kicinski; +Cc: David Miller, Linux Netdev List, oss-drivers, John Hurley
In-Reply-To: <20180524022255.18548-9-jakub.kicinski@netronome.com>
On Thu, May 24, 2018 at 5:22 AM, Jakub Kicinski
<jakub.kicinski@netronome.com> wrote:
> From: John Hurley <john.hurley@netronome.com>
>
> If the egress device of an offloaded rule is a LAG port, then encode the
> output port to the NFP with a LAG identifier and the offloaded group ID.
>
> A prelag action is also offloaded which must be the first action of the
> series (although may appear after other pre-actions - e.g. tunnels). This
> causes the FW to check that it has the necessary information to output to
> the requested LAG port. If it does not, the packet is sent to the kernel
> before any other actions are applied to it.
Offload decision typically also looks if both devices have the same
switchdev ID.
In your case, do both reprs gets the same switchdev ID automatically when being
put into the same team/bond instance? I wasn't sure to see here
changes for that matter
^ 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