[dpdk-dev] [PATCH] ethdev: fix a regression due to cache alignment issue
Prior to "ethdev: add port ownership" change, the fast
path structure struct rte_eth_dev->data was cache aligned
due to the fact that eth_dev_data was allocated from
rte_malloc with cache aligned attribute.
"ethdev: add port ownership" change set introduced a
rte_eth_dev_shared_data container for port ownership change,
This resulted in rte_eth_dev->data memory as cache unaligned.
Added a compiler alignment attribute to make sure
rte_eth_dev->data always cache aligned and hence compiler
can load/store the elements in struct rte_eth_dev_data
as naturally aligned.
Some platform like thunderx + l3fwd showed 1% regression in
the performance with the offending changeset.
Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
Cc: Matan Azrad
Cc: Thomas Monjalon
Cc: Konstantin Ananyev
Signed-off-by: Jerin Jacob
Signed-off-by: Pavan Nikhilesh
---
lib/librte_ether/rte_ethdev_core.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/librte_ether/rte_ethdev_core.h
b/lib/librte_ether/rte_ethdev_core.h
index 315b31723..e5681e466 100644
--- a/lib/librte_ether/rte_ethdev_core.h
+++ b/lib/librte_ether/rte_ethdev_core.h
@@ -601,7 +601,7 @@ struct rte_eth_dev_data {
struct rte_vlan_filter_conf vlan_filter_conf;
/**< VLAN filter configuration. */
struct rte_eth_dev_owner owner; /**< The port owner. */
-};
+} __rte_cache_aligned;
/**
* @internal
--
2.16.1
Re: [dpdk-dev] IXGBE, IOMMU DMAR DRHD handling fault issue
On 09-Feb-18 5:41 PM, Ravi Kerur wrote: On Thu, Feb 8, 2018 at 3:20 AM, Burakov, Anatoly mailto:anatoly.bura...@intel.com>> wrote: On 06-Feb-18 5:55 PM, Ravi Kerur wrote: Hi Anatoly, I am actually confused with the state of vIOMMU + DPDK. Can you please help me clarify? I tested following DPDK versions (1) DPDK 17.11, exhibits the issue (IOMMU width as reported by RedHat and solution is to prevent using the patch) (2) DPDK 17.05.02 (stable release) using 'testpmd' I was able to bind a device in VM with VFIO driver and no DMAR error message on host (3) DPDK 16.11.02 (stable release) using 'testpmd' I was able to bind a device in VM with VFIO driver and no DMAR error message on host Clearly issue seen in 17.11 without the patch you mentioned is a regression or the issue was masked in earlier DPDK version? I did not test traffic with any DPDK version because I wanted to first get DMAR errors on host gone. Our application 'v1' is integrated with DPDK 16.11 and 'v2' is integrated with DPDK 17.05.01. In both 'v1' and 'v2' cases I don't see IOMMU width error messages on VM, however, DMAR error messages are seen host. I am not able to relate what causes DMAR error messages on host? Thanks. Hi Ravi, vIOMMU support is out of our hands, really - we can only make use of hardware (or emulation of it) that is available. 39-bit wide address *can* work, you just have to be lucky and get PA addresses that would fit into 39 bits (so under 512G limit), because we set up IOVA addresses to be 1:1 to physical addresses. We could, in principle, set up IOVA addresses to go from zero instead of them being a 1:1 mapping to physical addresses, but that would introduce need to translate addresses between IOVA and physical in some cases (e.g. KNI). I'm not aware of any changes between 16.11 and 17.11 (and indeed 18.02) that would make or break support for 39-bit wide PA addresses for IOMMU. It is possible that VF/PF drivers do something differently which results in DMAR errors showing up sooner rather than later, but as far as VFIO support itself is concerned, there were no major changes in those releases. Hi Anatoly, Hi Ravi, Thank you for your explanation. I would like to ask one more thing as I need to get v-iommu+ dpdk working in VM. Can you please tell me what determines 'Host Address Width", I know my question has nothing to do with dpdk and this is a dpdk list, but if you have any information please share it? I googled and found couple of ways to influence 'Host Address Width = 46' in guest as well (since dpdk + iommu works fine on host and DMAR on host reports address width as 46). As far as i'm aware, address width supported by IOMMU is taken from hardware (real or emulated). If e.g. QEMU's vIOMMU emulates an IOMMU that only supports 39-bit addresses, that's what we'll get. Not sure what "host address width" means exactly, but if i had to guess, usually in VM-speak, "host" refers to, well, host - that is, the physical machine the VM is running on. So, under that assumption, "host address width" would be telling QEMU maximum width of IOMMU addresses on the host (i.e. what you have in hardware, on the host - for example, you might be running on a platform that only supports a 39-bit wide physical address). Naturally, changing it around wouldn't change much - it would presumably affect only the way vIOMMU remaps requests from guest to host, but not the emulated hardware itself. However, i'm not knowledgeable enough in that area to answer this question definitively. (1) Qemu has CPU param 'host-phys-bits' boolean, when set to true copies it from host (2) Qemu has 'phys-bits' integer, when set to '46' should influence guest Using above options when instantiating a VM doesn't help, Guest VM still ends up with 'Host address width = 39'. (3) There is another Qemu option 'x-aw-bits' which is for VT-d which can be set to '39' or '48'. This doesn't help either. Thanks. -- Thanks, Anatoly -- Thanks, Anatoly
Re: [dpdk-dev] IXGBE, IOMMU DMAR DRHD handling fault issue
On 29-Jan-18 10:35 PM, Ravi Kerur wrote: Hi Burakov, When using vfio-pci on host both VF and PF interfaces works fine with dpdk i.e. I don't see DMAR fault messages anymore. However, when I attach a VF interface to a VM and start DPDK with vfio-pci inside VM I still see DMAR fault messages on host. Both host and VM are booted with 'intel-iommu=on' on GRUB. Ping from VM with DPDK/vfio-pci doesn't work (I think it's expected because of DMAR faults), however, when VF interface uses ixgbevf driver ping works. Following are some details /*On VM***/ dpdk-devbind -s Network devices using DPDK-compatible driver :00:07.0 '82599 Ethernet Controller Virtual Function' drv=vfio-pci unused=ixgbevf Network devices using kernel driver === :03:00.0 'Device 1041' if=eth0 drv=virtio-pci unused=vfio-pci *Active* :04:00.0 'Device 1041' if=eth1 drv=virtio-pci unused=vfio-pci :05:00.0 'Device 1041' if=eth2 drv=virtio-pci unused=vfio-pci Other network devices = Crypto devices using DPDK-compatible driver === Crypto devices using kernel driver == Other crypto devices 00:07.0 Ethernet controller: Intel Corporation 82599 Ethernet Controller Virtual Function (rev 01) Subsystem: Intel Corporation 82599 Ethernet Controller Virtual Function Control: I/O- Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+ Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- SERR- Region 0: Memory at fda0 (64-bit, prefetchable) [size=16K] Region 3: Memory at fda04000 (64-bit, prefetchable) [size=16K] Capabilities: [70] MSI-X: Enable+ Count=3 Masked- Vector table: BAR=3 offset= PBA: BAR=3 offset=2000 Capabilities: [a0] Express (v1) Root Complex Integrated Endpoint, MSI 00 DevCap: MaxPayload 128 bytes, PhantFunc 0 ExtTag- RBE- DevCtl: Report errors: Correctable- Non-Fatal- Fatal- Unsupported- RlxdOrd- ExtTag- PhantFunc- AuxPwr- NoSnoop- MaxPayload 128 bytes, MaxReadReq 128 bytes DevSta: CorrErr- UncorrErr- FatalErr- UnsuppReq- AuxPwr- TransPend- Capabilities: [100 v1] Advanced Error Reporting UESta: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol- UEMsk: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol- UESvrt: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol- CESta: RxErr- BadTLP- BadDLLP- Rollover- Timeout- NonFatalErr- CEMsk: RxErr- BadTLP- BadDLLP- Rollover- Timeout- NonFatalErr- AERCap: First Error Pointer: 00, GenCap- CGenEn- ChkCap- ChkEn- Kernel driver in use: vfio-pci Kernel modules: ixgbevf /***on Host*/ dmesg | grep DMAR ... [ 978.268143] DMAR: DRHD: handling fault status reg 2 [ 978.268147] DMAR: [DMA Read] *Request device [04:10.0]* fault addr 33a128000 [fault reason 06] PTE Read access is not set [ 1286.677726] DMAR: DRHD: handling fault status reg 102 [ 1286.677730] DMAR: [DMA Read] Request device [04:10.0] fault addr fb663000 [fault reason 06] PTE Read access is not set [ 1676.436145] DMAR: DRHD: handling fault status reg 202 [ 1676.436149] DMAR: [DMA Read] Request device [04:10.0] fault addr 33a128000 [fault reason 06] PTE Read access is not set [ 1734.433649] DMAR: DRHD: handling fault status reg 302 [ 1734.433652] DMAR: [DMA Read] Request device [04:10.0] fault addr 33a128000 [fault reason 06] PTE Read access is not set [ 2324.428938] DMAR: DRHD: handling fault status reg 402 [ 2324.428942] DMAR: [DMA Read] Request device [04:10.0] fault addr 7770c000 [fault reason 06] PTE Read access is not set [ 2388.553640] DMAR: DRHD: handling fault status reg 502 [ 2388.553643] DMAR: [DMA Read] *Request device [04:10.0]* fault addr 33a128000 [fault reason 06] PTE Read access is not set Going back to this, i would like to suggest run a few tests to ensure that we have all information that we can gather. First of all, i'm assuming that you're using native ixgbe Linux driver on the host, and that you're only passing through the VF device to the VM using VFIO. Is my understanding correct here? Now, let's forget about the iommu=pt and igb_uio for a moment. Boot both your host and your VM with iommu=on and intel_iommu=on (or whatever command-line enables full IOMMU support on both host and guest) and do the same tests you've done before. Do you still see your issues? It would also be very useful t
[dpdk-dev] [PATCH] net/dpaa2: remove unused global variable
Due to the missing __packed definition, compiler treating
it as as global variable.
#grep "Allocating common symbols" -A 100 build/app/testacl.map
__packed 0x8 /build/lib/librte_pmd_dpaa2.a(dpaa2_rxtx.o)
Fixes: a5fc38d422a7 ("net/dpaa2: support Rx packet parsing")
Cc: hemant.agra...@nxp.com
Signed-off-by: Jerin Jacob
---
drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h
b/drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h
index 41931e858..779cdf2b2 100644
--- a/drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h
+++ b/drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h
@@ -37,7 +37,7 @@ struct dpaa2_fas {
uint8_t ppid;
__le16 ifpid;
__le32 status;
-} __packed;
+} __attribute__((__packed__));
/**
* HW Packet Annotation Register structures
--
2.16.1
Re: [dpdk-dev] [PATCH] net/dpaa2: remove unused global variable
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Jerin Jacob
> Sent: Saturday, February 10, 2018 5:28 PM
> To: dev@dpdk.org
> Cc: Jerin Jacob ; Hemant Agrawal
>
> Subject: [dpdk-dev] [PATCH] net/dpaa2: remove unused global variable
> Importance: High
>
> Due to the missing __packed definition, compiler treating it as as global
> variable.
>
> #grep "Allocating common symbols" -A 100 build/app/testacl.map
> __packed 0x8 /build/lib/librte_pmd_dpaa2.a(dpaa2_rxtx.o)
[Hemant] Which compiler, you mean you faced issue with testacl app?
>
> Fixes: a5fc38d422a7 ("net/dpaa2: support Rx packet parsing")
>
> Cc: hemant.agra...@nxp.com
> Signed-off-by: Jerin Jacob
> ---
> drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
Acked-by: Hemant Agrawal
Re: [dpdk-dev] [PATCH] doc: fix ethdev API port_id parameter size
-Original Message- > Date: Fri, 9 Feb 2018 15:01:28 + > From: Hemant Agrawal > To: "Nicolau, Radu" , "Yigit, Ferruh" > , Neil Horman , "Mcnamara, > John" , "Kovacevic, Marko" > > CC: "dev@dpdk.org" , Thomas Monjalon , > Boris Pismenny , Aviad Yehezkel > , "Doherty, Declan" > Subject: Re: [dpdk-dev] [PATCH] doc: fix ethdev API port_id parameter size > > > > > > > Fix rte_eth_dev_get_sec_ctx() parameter port_id storage size, form > > > uint8_t to uint16_t > > > > > > Signed-off-by: Ferruh Yigit > > > --- > > Acked-by: Radu Nicolau > > Acked-by: Hemant Agrawal Acked-by: Jerin Jacob
Re: [dpdk-dev] [PATCH v2] doc: update ethdev APIs to return named opaque type
-Original Message- > Date: Fri, 9 Feb 2018 13:22:06 + > From: "Ananyev, Konstantin" > To: "Yigit, Ferruh" , Neil Horman > , "Mcnamara, John" , > "Kovacevic, Marko" > CC: "dev@dpdk.org" , Stephen Hemminger > , "Richardson, Bruce" > , Thomas Monjalon > Subject: Re: [dpdk-dev] [PATCH v2] doc: update ethdev APIs to return named > opaque type > > > > > -Original Message- > > From: Yigit, Ferruh > > Sent: Friday, February 9, 2018 10:18 AM > > To: Neil Horman ; Mcnamara, John > > ; Kovacevic, Marko > > > > Cc: dev@dpdk.org; Yigit, Ferruh ; Ananyev, > > Konstantin ; Stephen Hemminger > > ; Richardson, Bruce > > ; Thomas Monjalon > > Subject: [PATCH v2] doc: update ethdev APIs to return named opaque type > > > > Ethdev APIs to add callback return the callback object as "void *", > > update return type to actual object type > > "struct rte_eth_rxtx_callback *" > > > > Signed-off-by: Ferruh Yigit > > --- > > Cc: Konstantin Ananyev > > Cc: Stephen Hemminger > > Cc: Bruce Richardson > > Cc: Thomas Monjalon > > --- > > doc/guides/rel_notes/deprecation.rst | 7 +++ > > 1 file changed, 7 insertions(+) > > > > diff --git a/doc/guides/rel_notes/deprecation.rst > > b/doc/guides/rel_notes/deprecation.rst > > index bbd9456a7..5cb5a00d2 100644 > > --- a/doc/guides/rel_notes/deprecation.rst > > +++ b/doc/guides/rel_notes/deprecation.rst > > @@ -49,6 +49,13 @@ Deprecation Notices > >rte_eth_dev_get_sec_ctx() is using uint8_t for port_id, which should be > >uint16_t. > > > > +* ethdev: functions add rx/tx callback will return named opaque type > > + rte_eth_add_rx_callback(), rte_eth_add_first_rx_callback() and > > + rte_eth_add_tx_callback() functions currently return callback object as > > + "void \*" but APIs to delete callbacks get "struct rte_eth_rxtx_callback > > \*" > > + as parameter. For consistency functions adding callback will return > > + "struct rte_eth_rxtx_callback \*" instead of "void * ". > > + > > * i40e: The default flexible payload configuration which extracts the > > first 16 > >bytes of the payload for RSS will be deprecated starting from 18.02. If > >required the previous behavior can be configured using existing flow > > -- > > Acked-by: Konstantin Ananyev Acked-by: Jerin Jacob > > > 2.14.3 >
Re: [dpdk-dev] [PATCH] net/dpaa2: remove unused global variable
-Original Message-
> Date: Sat, 10 Feb 2018 12:02:24 +
> From: Hemant Agrawal
> To: Jerin Jacob , "dev@dpdk.org"
>
> Subject: RE: [dpdk-dev] [PATCH] net/dpaa2: remove unused global variable
>
> > -Original Message-
> > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Jerin Jacob
> > Sent: Saturday, February 10, 2018 5:28 PM
> > To: dev@dpdk.org
> > Cc: Jerin Jacob ; Hemant Agrawal
> >
> > Subject: [dpdk-dev] [PATCH] net/dpaa2: remove unused global variable
> > Importance: High
> >
> > Due to the missing __packed definition, compiler treating it as as global
> > variable.
>
> >
> > #grep "Allocating common symbols" -A 100 build/app/testacl.map
> > __packed 0x8 /build/lib/librte_pmd_dpaa2.a(dpaa2_rxtx.o)
>
> [Hemant] Which compiler, you mean you faced issue with testacl app?
arm64 gcc. Not specific to testacl. We were getting minor performance
regression (<1%) with l3fwd after the following change set.
http://dpdk.org/dev/patchwork/patch/34709/
I don't think, the above change set directly has anything due to issue.
I was chasing that issue and found that global variables has something
due to with change.Most probably because it has indirect effect of changing
the complete address generated in the binary and thus showing up some cached
alignment related artifacts.
>
> >
> > Fixes: a5fc38d422a7 ("net/dpaa2: support Rx packet parsing")
> >
> > Cc: hemant.agra...@nxp.com
> > Signed-off-by: Jerin Jacob
> > ---
> > drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
>
> Acked-by: Hemant Agrawal
>
>
Re: [dpdk-dev] [PATCH] doc: add change notice for mbuf sched field
-Original Message- > Date: Fri, 9 Feb 2018 16:45:16 + > From: Cristian Dumitrescu > To: dev@dpdk.org > CC: olivier.m...@6wind.com, tho...@monjalon.net, jasvinder.si...@intel.com, > roy.fan.zh...@intel.com, kevin.la...@intel.com > Subject: [dpdk-dev] [PATCH] doc: add change notice for mbuf sched field > X-Mailer: git-send-email 2.7.4 > > Signed-off-by: Cristian Dumitrescu > Acked-by: Jasvinder Singh > Acked-by: Roy Fan Zhang > Acked-by: Kevin Laatz Acked-by: Jerin Jacob > --- > doc/guides/rel_notes/deprecation.rst | 5 + > 1 file changed, 5 insertions(+) > > diff --git a/doc/guides/rel_notes/deprecation.rst > b/doc/guides/rel_notes/deprecation.rst > index d59ad59..db4fea3 100644 > --- a/doc/guides/rel_notes/deprecation.rst > +++ b/doc/guides/rel_notes/deprecation.rst > @@ -59,3 +59,8 @@ Deprecation Notices >be added between the producer and consumer structures. The size of the >structure and the offset of the fields will remain the same on >platforms with 64B cache line, but will change on other platforms. > + > +* mbuf: The opaque mbuf->hash.sched field will be updated to support generic > + definition in line with the ethdev TM and MTR APIs. Currently, this field > + is defined in librte_sched in a non-generic way. The new generic format > + will contain: queue ID, traffic class, color. Field size will not change. > -- > 2.7.4 >
[dpdk-dev] [PATCH] eal: fix rte_errno values for IPC API
rte_errno values should not be negative.
Fixes: bacaa2754017 ("eal: add channel for multi-process communication")
Fixes: 783b6e54971d ("eal: add synchronous multi-process communication")
Cc: jianfeng@intel.com
Signed-off-by: Anatoly Burakov
---
Notes:
This is in DPDK coding style guide:
http://dpdk.org/doc/guides/contributing/coding_style.html
However, i should note that situation with this is a mess
in DPDK, and in a lot of cases, rte_errno is set to -errno.
lib/librte_eal/common/eal_common_proc.c | 24
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/lib/librte_eal/common/eal_common_proc.c
b/lib/librte_eal/common/eal_common_proc.c
index b974837..caa8774 100644
--- a/lib/librte_eal/common/eal_common_proc.c
+++ b/lib/librte_eal/common/eal_common_proc.c
@@ -131,16 +131,16 @@ validate_action_name(const char *name)
{
if (name == NULL) {
RTE_LOG(ERR, EAL, "Action name cannot be NULL\n");
- rte_errno = -EINVAL;
+ rte_errno = EINVAL;
return -1;
}
if (strnlen(name, RTE_MP_MAX_NAME_LEN) == 0) {
RTE_LOG(ERR, EAL, "Length of action name is zero\n");
- rte_errno = -EINVAL;
+ rte_errno = EINVAL;
return -1;
}
if (strnlen(name, RTE_MP_MAX_NAME_LEN) == RTE_MP_MAX_NAME_LEN) {
- rte_errno = -E2BIG;
+ rte_errno = E2BIG;
return -1;
}
return 0;
@@ -156,7 +156,7 @@ rte_mp_action_register(const char *name, rte_mp_t action)
entry = malloc(sizeof(struct action_entry));
if (entry == NULL) {
- rte_errno = -ENOMEM;
+ rte_errno = ENOMEM;
return -1;
}
strcpy(entry->action_name, name);
@@ -165,7 +165,7 @@ rte_mp_action_register(const char *name, rte_mp_t action)
pthread_mutex_lock(&mp_mutex_action);
if (find_action_entry_by_name(name) != NULL) {
pthread_mutex_unlock(&mp_mutex_action);
- rte_errno = -EEXIST;
+ rte_errno = EEXIST;
free(entry);
return -1;
}
@@ -505,7 +505,7 @@ check_input(const struct rte_mp_msg *msg)
{
if (msg == NULL) {
RTE_LOG(ERR, EAL, "Msg cannot be NULL\n");
- rte_errno = -EINVAL;
+ rte_errno = EINVAL;
return false;
}
@@ -514,14 +514,14 @@ check_input(const struct rte_mp_msg *msg)
if (msg->len_param > RTE_MP_MAX_PARAM_LEN) {
RTE_LOG(ERR, EAL, "Message data is too long\n");
- rte_errno = -E2BIG;
+ rte_errno = E2BIG;
return false;
}
if (msg->num_fds > RTE_MP_MAX_FD_NUM) {
RTE_LOG(ERR, EAL, "Cannot send more than %d FDs\n",
RTE_MP_MAX_FD_NUM);
- rte_errno = -E2BIG;
+ rte_errno = E2BIG;
return false;
}
@@ -560,7 +560,7 @@ mp_request_one(const char *dst, struct rte_mp_msg *req,
pthread_mutex_unlock(&sync_requests.lock);
if (exist) {
RTE_LOG(ERR, EAL, "A pending request %s:%s\n", dst, req->name);
- rte_errno = -EEXIST;
+ rte_errno = EEXIST;
return -1;
}
@@ -596,7 +596,7 @@ mp_request_one(const char *dst, struct rte_mp_msg *req,
if (sync_req.reply_received == 0) {
RTE_LOG(ERR, EAL, "Fail to recv reply for request %s:%s\n",
dst, req->name);
- rte_errno = -ETIMEDOUT;
+ rte_errno = ETIMEDOUT;
return -1;
}
@@ -604,7 +604,7 @@ mp_request_one(const char *dst, struct rte_mp_msg *req,
if (!tmp) {
RTE_LOG(ERR, EAL, "Fail to alloc reply for request %s:%s\n",
dst, req->name);
- rte_errno = -ENOMEM;
+ rte_errno = ENOMEM;
return -1;
}
memcpy(&tmp[reply->nb_received], &msg, sizeof(msg));
@@ -676,7 +676,7 @@ rte_mp_reply(struct rte_mp_msg *msg, const char *peer)
if (peer == NULL) {
RTE_LOG(ERR, EAL, "peer is not specified\n");
- rte_errno = -EINVAL;
+ rte_errno = EINVAL;
return -1;
}
--
2.7.4
Re: [dpdk-dev] IXGBE, IOMMU DMAR DRHD handling fault issue
On Sat, Feb 10, 2018 at 2:58 AM, Burakov, Anatoly wrote: > On 29-Jan-18 10:35 PM, Ravi Kerur wrote: > >> Hi Burakov, >> >> When using vfio-pci on host both VF and PF interfaces works fine with >> dpdk i.e. I don't see DMAR fault messages anymore. However, when I attach a >> VF interface to a VM and start DPDK with vfio-pci inside VM I still see >> DMAR fault messages on host. Both host and VM are booted with >> 'intel-iommu=on' on GRUB. Ping from VM with DPDK/vfio-pci doesn't work (I >> think it's expected because of DMAR faults), however, when VF interface >> uses ixgbevf driver ping works. >> >> Following are some details >> >> /*On VM***/ >> dpdk-devbind -s >> >> Network devices using DPDK-compatible driver >> >> :00:07.0 '82599 Ethernet Controller Virtual Function' drv=vfio-pci >> unused=ixgbevf >> >> Network devices using kernel driver >> === >> :03:00.0 'Device 1041' if=eth0 drv=virtio-pci unused=vfio-pci *Active* >> :04:00.0 'Device 1041' if=eth1 drv=virtio-pci unused=vfio-pci >> :05:00.0 'Device 1041' if=eth2 drv=virtio-pci unused=vfio-pci >> >> Other network devices >> = >> >> >> Crypto devices using DPDK-compatible driver >> === >> >> >> Crypto devices using kernel driver >> == >> >> >> Other crypto devices >> >> >> >> >> 00:07.0 Ethernet controller: Intel Corporation 82599 Ethernet Controller >> Virtual Function (rev 01) >> Subsystem: Intel Corporation 82599 Ethernet Controller Virtual >> Function >> Control: I/O- Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- >> ParErr- Stepping- SERR- FastB2B- DisINTx+ >> Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- >> SERR- > Region 0: Memory at fda0 (64-bit, prefetchable) [size=16K] >> Region 3: Memory at fda04000 (64-bit, prefetchable) [size=16K] >> Capabilities: [70] MSI-X: Enable+ Count=3 Masked- >> Vector table: BAR=3 offset= >> PBA: BAR=3 offset=2000 >> Capabilities: [a0] Express (v1) Root Complex Integrated >> Endpoint, MSI 00 >> DevCap: MaxPayload 128 bytes, PhantFunc 0 >> ExtTag- RBE- >> DevCtl: Report errors: Correctable- Non-Fatal- Fatal- >> Unsupported- >> RlxdOrd- ExtTag- PhantFunc- AuxPwr- NoSnoop- >> MaxPayload 128 bytes, MaxReadReq 128 bytes >> DevSta: CorrErr- UncorrErr- FatalErr- UnsuppReq- AuxPwr- >> TransPend- >> Capabilities: [100 v1] Advanced Error Reporting >> UESta: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- >> UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol- >> UEMsk: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- >> UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol- >> UESvrt: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- >> UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol- >> CESta: RxErr- BadTLP- BadDLLP- Rollover- Timeout- >> NonFatalErr- >> CEMsk: RxErr- BadTLP- BadDLLP- Rollover- Timeout- >> NonFatalErr- >> AERCap: First Error Pointer: 00, GenCap- CGenEn- ChkCap- >> ChkEn- >> Kernel driver in use: vfio-pci >> Kernel modules: ixgbevf >> >> /***on Host*/ >> dmesg | grep DMAR >> ... >> [ 978.268143] DMAR: DRHD: handling fault status reg 2 >> [ 978.268147] DMAR: [DMA Read] *Request device [04:10.0]* fault addr >> 33a128000 [fault reason 06] PTE Read access is not set >> [ 1286.677726] DMAR: DRHD: handling fault status reg 102 >> [ 1286.677730] DMAR: [DMA Read] Request device [04:10.0] fault addr >> fb663000 [fault reason 06] PTE Read access is not set >> [ 1676.436145] DMAR: DRHD: handling fault status reg 202 >> [ 1676.436149] DMAR: [DMA Read] Request device [04:10.0] fault addr >> 33a128000 [fault reason 06] PTE Read access is not set >> [ 1734.433649] DMAR: DRHD: handling fault status reg 302 >> [ 1734.433652] DMAR: [DMA Read] Request device [04:10.0] fault addr >> 33a128000 [fault reason 06] PTE Read access is not set >> [ 2324.428938] DMAR: DRHD: handling fault status reg 402 >> [ 2324.428942] DMAR: [DMA Read] Request device [04:10.0] fault addr >> 7770c000 [fault reason 06] PTE Read access is not set >> [ 2388.553640] DMAR: DRHD: handling fault status reg 502 >> [ 2388.553643] DMAR: [DMA Read] *Request device [04:10.0]* fault addr >> 33a128000 [fault reason 06] PTE Read access is not set >> >> >> > Going back to this, i would like to suggest run a few tests to ensure that > we have all information that we can gather. > > First of all, i'm assuming that you're using native ixgbe Linux driver on > the host, and that you're only passing through the VF device to the VM > using VFIO. Is my understandin
Re: [dpdk-dev] [PATCH] ethdev: fix a regression due to cache alignment issue
Hi Jerin
From: Jerin Jacob,Sent: Saturday, February 10, 2018 11:42 AM
> Prior to "ethdev: add port ownership" change, the fast path structure struct
> rte_eth_dev->data was cache aligned due to the fact that eth_dev_data was
> allocated from rte_malloc with cache aligned attribute.
> "ethdev: add port ownership" change set introduced a
> rte_eth_dev_shared_data container for port ownership change, This
> resulted in rte_eth_dev->data memory as cache unaligned.
> Added a compiler alignment attribute to make sure rte_eth_dev->data
> always cache aligned and hence compiler can load/store the elements in
> struct rte_eth_dev_data as naturally aligned.
>
> Some platform like thunderx + l3fwd showed 1% regression in the
> performance with the offending changeset.
>
> Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
>
Are you sure the previous port data allocation(before port ownership) was cache
aligned for all the ports for every system (cache line size 16\32\64\128...) or
maybe it was only aligned to the first port (port_id=0)?
I think that if the answer is no, this fix fixes early patch, or just
improvement.
> Cc: Matan Azrad
> Cc: Thomas Monjalon
> Cc: Konstantin Ananyev
>
> Signed-off-by: Jerin Jacob
> Signed-off-by: Pavan Nikhilesh
> ---
> lib/librte_ether/rte_ethdev_core.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib/librte_ether/rte_ethdev_core.h
> b/lib/librte_ether/rte_ethdev_core.h
> index 315b31723..e5681e466 100644
> --- a/lib/librte_ether/rte_ethdev_core.h
> +++ b/lib/librte_ether/rte_ethdev_core.h
> @@ -601,7 +601,7 @@ struct rte_eth_dev_data {
> struct rte_vlan_filter_conf vlan_filter_conf;
> /**< VLAN filter configuration. */
> struct rte_eth_dev_owner owner; /**< The port owner. */ -};
> +} __rte_cache_aligned;
>
> /**
> * @internal
> --
> 2.16.1
[dpdk-dev] [PATCH] maintainers: update vhost lib and pmd
Signed-off-by: Jianfeng Tan --- MAINTAINERS | 2 ++ 1 file changed, 2 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index 4f1f33b..38e5fb8 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -559,6 +559,7 @@ F: doc/guides/nics/features/vmxnet3.ini Vhost-user M: Yuanhan Liu M: Maxime Coquelin +M: Jianfeng Tan T: git://dpdk.org/next/dpdk-next-virtio F: lib/librte_vhost/ F: doc/guides/prog_guide/vhost_lib.rst @@ -571,6 +572,7 @@ Vhost PMD M: Tetsuya Mukawa M: Yuanhan Liu M: Maxime Coquelin +M: Jianfeng Tan T: git://dpdk.org/next/dpdk-next-virtio F: drivers/net/vhost/ F: doc/guides/nics/features/vhost.ini -- 2.7.4
Re: [dpdk-dev] [PATCH v2] vhost: fix check if cmsg is NULL
> -Original Message-
> From: Kulasek, TomaszX
> Sent: Saturday, February 10, 2018 1:06 AM
> To: y...@fridaylinux.org
> Cc: dev@dpdk.org; Tan, Jianfeng; sta...@dpdk.org; Wodkowski, PawelX
> Subject: [PATCH v2] vhost: fix check if cmsg is NULL
>
> Fixes: 8f972312b8f4 ("vhost: support vhost-user")
> Cc: jianfeng@intel.com
> Cc: sta...@dpdk.org
>
> Signed-off-by: Pawel Wodkowski
> Signed-off-by: Tomasz Kulasek
Reviewed-by: Jianfeng Tan
Thanks for the fix.
Re: [dpdk-dev] [PATCH] eal: fix rte_errno values for IPC API
> -Original Message-
> From: Burakov, Anatoly
> Sent: Saturday, February 10, 2018 9:15 PM
> To: dev@dpdk.org
> Cc: Tan, Jianfeng
> Subject: [PATCH] eal: fix rte_errno values for IPC API
>
> rte_errno values should not be negative.
>
> Fixes: bacaa2754017 ("eal: add channel for multi-process communication")
> Fixes: 783b6e54971d ("eal: add synchronous multi-process communication")
> Cc: jianfeng@intel.com
> Signed-off-by: Anatoly Burakov
Reviewed-by: Jianfeng Tan
Thanks for fixing this.
[dpdk-dev] i354 autoneg-off link issue
Hi, I am observing a strange issue with i354 when auto-negotiation is turned off. The device I have has 6 ethernet ports. Port 1 and 2 are based on i210 and port 3 to 6 are based on i354 (connected thru an external PHY). The CPU is Atom Rangeley (4 core). I have connected ports 3 and 4 back to back with an external cable. When the link is forced to 100M/Full duplex with auto-negotiation off on both the ports, the link does not come up. Whereas the link comes up fine when eth3 (i354) and eth1 (i210) are connected back to back and forced to 100M/Full duplex with auto-negotiation off. This issue is seen with DPDK (16.04) and also linux kernel (ubuntu 14.04). Is this a known issue ? Following are logs when running the test using ethtool from linux admin@Ananda-Desk:~$ lspci | grep Ethernet 00:14.0 Ethernet controller: Intel Corporation Ethernet Connection I354 (rev 03) - eth2 00:14.1 Ethernet controller: Intel Corporation Ethernet Connection I354 (rev 03) - eth3 00:14.2 Ethernet controller: Intel Corporation Ethernet Connection I354 (rev 03) - eth4 00:14.3 Ethernet controller: Intel Corporation Ethernet Connection I354 (rev 03) - eth5 01:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03) - eth0 02:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03) - eth1 admin@Ananda-Desk:~$ ethtool -i eth3 driver: igb version: 5.2.13-k firmware-version: 0.0.0 bus-info: :00:14.1 supports-statistics: yes supports-test: yes supports-eeprom-access: yes supports-register-dump: yes supports-priv-flags: no [Force 100M/Full on eth3] admin@Ananda-Desk:~$ sudo ethtool -s eth3 autoneg off duplex full speed 100 admin@Ananda-Desk:~$ ethtool eth3 Settings for eth3: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supported pause frame use: Symmetric Supports auto-negotiation: Yes Advertised link modes: Not reported Advertised pause frame use: Symmetric Advertised auto-negotiation: No Speed: 100Mb/s Duplex: Full Port: Twisted Pair PHYAD: 1 Transceiver: internal Auto-negotiation: off MDI-X: off (auto) Cannot get wake-on-lan settings: Operation not permitted Current message level: 0x0007 (7) drv probe link Link detected: yes admin@Ananda-Desk:~$ ethtool eth4 Settings for eth4: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supported pause frame use: Symmetric Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Advertised pause frame use: Symmetric Advertised auto-negotiation: Yes Speed: 100Mb/s Duplex: Half Port: Twisted Pair PHYAD: 2 Transceiver: internal Auto-negotiation: on MDI-X: on (auto) Cannot get wake-on-lan settings: Operation not permitted Current message level: 0x0007 (7) drv probe link Link detected: yes [Force 100M/Full on eth4] admin@Ananda-Desk:~$ sudo ethtool -s eth4 autoneg off duplex full speed 100 admin@Ananda-Desk:~$ admin@Ananda-Desk:~$ ethtool eth4 Settings for eth4: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supported pause frame use: Symmetric Supports auto-negotiation: Yes Advertised link modes: Not reported Advertised pause frame use: Symmetric Advertised auto-negotiation: No Speed: Unknown! Duplex: Unknown! (255) Port: Twisted Pair PHYAD: 2 Transceiver: internal Auto-negotiation: off MDI-X: on (auto) Cannot get wake-on-lan settings: Operation not permitted Current message level: 0x0007 (7) drv probe link Link detected: no << admin@Ananda-Desk:~$ ethtool eth3 Settings for eth3: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supported pause frame use: Symmetric Supports auto-negotiation: Yes Advertised link modes: Not reported Advertised pause frame use: Symmetric Advertised auto-negotiation: No Speed: Unknown! Duplex: Unknown! (255) Port: Twisted Pair PHYAD: 1
Re: [dpdk-dev] [PATCH] maintainers: update vhost lib and pmd
> -Original Message- > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Jianfeng Tan > Sent: Sunday, February 11, 2018 9:05 AM > To: y...@fridaylinux.org; maxime.coque...@redhat.com; > mtetsu...@gmail.com; dev@dpdk.org > Cc: tho...@monjalon.net; Tan, Jianfeng > Subject: [dpdk-dev] [PATCH] maintainers: update vhost lib and pmd > > Signed-off-by: Jianfeng Tan Acked-by: Zhiyong Yang
[dpdk-dev] PF vlan filter can not work in x710
hi, when I use X710 PF in dpdk mode, i found that vlan filter function does not work. when I only add vlan 102, but I can receive packets with vlan tag100,200,300 etc. I can reproduce this phenomenon following the step: 1. set iommu=pt intel_iommu=on in file /etc/default/grub 2. dpdk-devbind -b vfio-pci eth0 driver: i40e version: 2.3.6 firmware-version: 5.04 0x8000250b 17.5.10 expansion-rom-version: bus-info: :05:00.0 supports-statistics: yes supports-test: yes supports-eeprom-access: yes supports-register-dump: yes supports-priv-flags: yes 3.use dpdk version 17.08 and ovs version 2.8.0 4.in ovs code, add vlan filter in function dpdk_eth_dev_init(): a. set conf.rxmode.hw_vlan_filter =1 before use function rte_eth_dev_configure() b. call function rte_eth_dev_vlan_filter(dev->port_id, 102,1) c. call function rte_eth_dev_start
Re: [dpdk-dev] [PATCH] vhost: fix close callfd on get vring base
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Tomasz Kulasek
> Sent: Saturday, February 10, 2018 1:11 AM
> To: y...@fridaylinux.org
> Cc: dev@dpdk.org; Xie, Huawei; sta...@dpdk.org; Stojaczyk, DariuszX
> Subject: [dpdk-dev] [PATCH] vhost: fix close callfd on get vring base
>
> This prevents from destroying & recreating user device in "incomplete"
> vring state. virtio_is_ready() was returning true for devices with
> vrings which did not have valid callfd (their VHOST_USER_SET_VRING_CALL
> hasn't arrived yet)
>
> Fixes: 8f972312b8f4 ("vhost: support vhost-user")
> Cc: huawei@intel.com
> Cc: sta...@dpdk.org
>
> Signed-off-by: Dariusz Stojaczyk
> Signed-off-by: Tomasz Kulasek
Reviewed-by: Jianfeng Tan
Thanks for fixing this.
Re: [dpdk-dev] [PATCH] vhost: fix double free on shutdown
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Tomasz Kulasek
> Sent: Saturday, February 10, 2018 1:15 AM
> To: y...@fridaylinux.org
> Cc: dev@dpdk.org; yuanhan@linux.intel.com; sta...@dpdk.org; Stojaczyk,
> DariuszX
> Subject: [dpdk-dev] [PATCH] vhost: fix double free on shutdown
>
> The vhost connection can be closed concurrently from 2 places:
> * the connection thread itself
> * rte_vhost_driver_unregister
>
> The connection thread will terminate the connection if any recv error
> occurred. The unregister function will terminate the connection together
> with the thread. However, there is no sychronization between those two.
> The connection thread runs in the background without any mutex.
Isn't it already protected by vsocket->conn_mutex?
Thanks,
Jianfeng
>
> The rte_vhost_driver_unregister now signals the connection thread
> to terminate itself and waits until it's killed.
>
> Fixes: 65388b43f592 ("vhost: fix fd leaks for vhost-user server mode")
> Cc: yuanhan@linux.intel.com
> Cc: sta...@dpdk.org
>
> Signed-off-by: Dariusz Stojaczyk
> Signed-off-by: Tomasz Kulasek
> ---
> lib/librte_vhost/socket.c | 21 -
> 1 file changed, 8 insertions(+), 13 deletions(-)
>
> diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c
> index 83befdced..46ac88efd 100644
> --- a/lib/librte_vhost/socket.c
> +++ b/lib/librte_vhost/socket.c
> @@ -735,7 +735,7 @@ rte_vhost_driver_unregister(const char *path)
> {
> int i;
> int count;
> - struct vhost_user_connection *conn, *next;
> + struct vhost_user_connection *conn;
>
> pthread_mutex_lock(&vhost_user.mutex);
>
> @@ -752,22 +752,17 @@ rte_vhost_driver_unregister(const char *path)
> }
>
> pthread_mutex_lock(&vsocket->conn_mutex);
> - for (conn = TAILQ_FIRST(&vsocket->conn_list);
> - conn != NULL;
> - conn = next) {
> - next = TAILQ_NEXT(conn, next);
> -
> - fdset_del(&vhost_user.fdset, conn->connfd);
> - RTE_LOG(INFO, VHOST_CONFIG,
> - "free connfd = %d for device '%s'\n",
> - conn->connfd, path);
> + TAILQ_FOREACH(conn, &vsocket->conn_list, next) {
> close(conn->connfd);
> - vhost_destroy_device(conn->vid);
> - TAILQ_REMOVE(&vsocket->conn_list, conn,
> next);
> - free(conn);
> }
> pthread_mutex_unlock(&vsocket->conn_mutex);
>
> + do {
> + pthread_mutex_lock(&vsocket-
> >conn_mutex);
> + conn = TAILQ_FIRST(&vsocket->conn_list);
> + pthread_mutex_unlock(&vsocket-
> >conn_mutex);
> + } while (conn != NULL);
> +
> pthread_mutex_destroy(&vsocket->conn_mutex);
> free(vsocket->path);
> free(vsocket);
> --
> 2.14.1
Re: [dpdk-dev] [PATCH] vhost: fix realloc failure
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Tomasz Kulasek
> Sent: Saturday, February 10, 2018 1:19 AM
> To: y...@fridaylinux.org
> Cc: dev@dpdk.org; yuanhan@linux.intel.com; sta...@dpdk.org; Yang,
> Ziye
> Subject: [dpdk-dev] [PATCH] vhost: fix realloc failure
>
> When reallocation of guest pages fails, vhost_user_set_mem_table() also
> should fail.
>
> Fixes: e246896178e6 ("vhost: get guest/host physical address mappings")
> Cc: yuanhan@linux.intel.com
> Cc: sta...@dpdk.org
>
> Signed-off-by: Ziye Yang
> Signed-off-by: Tomasz Kulasek
Reviewed-by: Jianfeng Tan
Re: [dpdk-dev] [PATCH] maintainers: update vhost lib and pmd
On Sun, Feb 11, 2018 at 01:04:51AM +, Jianfeng Tan wrote: > Signed-off-by: Jianfeng Tan Acked-by: Yuanhan Liu And welcome! BTW, the verb "update" doesn't look good to me ;) --yliu > --- > MAINTAINERS | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/MAINTAINERS b/MAINTAINERS > index 4f1f33b..38e5fb8 100644 > --- a/MAINTAINERS > +++ b/MAINTAINERS > @@ -559,6 +559,7 @@ F: doc/guides/nics/features/vmxnet3.ini > Vhost-user > M: Yuanhan Liu > M: Maxime Coquelin > +M: Jianfeng Tan > T: git://dpdk.org/next/dpdk-next-virtio > F: lib/librte_vhost/ > F: doc/guides/prog_guide/vhost_lib.rst > @@ -571,6 +572,7 @@ Vhost PMD > M: Tetsuya Mukawa > M: Yuanhan Liu > M: Maxime Coquelin > +M: Jianfeng Tan > T: git://dpdk.org/next/dpdk-next-virtio > F: drivers/net/vhost/ > F: doc/guides/nics/features/vhost.ini > -- > 2.7.4
Re: [dpdk-dev] [PATCH] vhost: fix remove macro name conflict
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Tomasz Kulasek
> Sent: Saturday, February 10, 2018 1:24 AM
> To: y...@fridaylinux.org
> Cc: dev@dpdk.org; Xie, Huawei; sta...@dpdk.org; Walker, Benjamin
> Subject: [dpdk-dev] [PATCH] vhost: fix remove macro name conflict
>
> LOG_DEBUG is a symbol defined by POSIX, so if sys/log.h is
> included the symbols conflict.
>
> This patch changes LOG_DEBUG to VHOST_LOG_DEBUG.
>
> Fixes: 1c01d52392d5 ("vhost: add debug print")
> Cc: huawei@intel.com
> Cc: sta...@dpdk.org
>
> Signed-off-by: Ben Walker
> Signed-off-by: Tomasz Kulasek
Reviewed-by: Jianfeng Tan
> ---
> lib/librte_vhost/vhost.h | 13 +++--
> lib/librte_vhost/vhost_user.c | 10 +-
> lib/librte_vhost/virtio_net.c | 16
> 3 files changed, 20 insertions(+), 19 deletions(-)
>
> diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
> index d947bc9e3..319cc6620 100644
> --- a/lib/librte_vhost/vhost.h
> +++ b/lib/librte_vhost/vhost.h
> @@ -296,8 +296,9 @@ vhost_log_used_vring(struct virtio_net *dev, struct
> vhost_virtqueue *vq,
>
> #ifdef RTE_LIBRTE_VHOST_DEBUG
> #define VHOST_MAX_PRINT_BUFF 6072
> -#define LOG_LEVEL RTE_LOG_DEBUG
> -#define LOG_DEBUG(log_type, fmt, args...) RTE_LOG(DEBUG, log_type, fmt,
> ##args)
> +#define VHOST_LOG_LEVEL RTE_LOG_DEBUG
Seems that we can remove LOG_LEVEL definition. But it's another issue.
Thanks,
Jianfeng
Re: [dpdk-dev] [PATCH] vhost: fix return avail ring position in get vring base
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Tomasz Kulasek
> Sent: Saturday, February 10, 2018 1:28 AM
> To: y...@fridaylinux.org
> Cc: dev@dpdk.org; Xie, Huawei; sta...@dpdk.org; Wodkowski, PawelX
> Subject: [dpdk-dev] [PATCH] vhost: fix return avail ring position in get vring
> base
>
> According to the "Vhost-user Protocol" document,
> VHOST_USER_GET_VRING_BASE should get the available vring base offset.
>
> Fixes: 8f972312b8f4 ("vhost: support vhost-user")
> Cc: huawei@intel.com
> Cc: sta...@dpdk.org
>
> Signed-off-by: Pawel Wodkowski
> Signed-off-by: Tomasz Kulasek
Reviewed-by: Jianfeng Tan
Nice catch!
> ---
> lib/librte_vhost/vhost_user.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
> index 65ee33919..04eee3a3a 100644
> --- a/lib/librte_vhost/vhost_user.c
> +++ b/lib/librte_vhost/vhost_user.c
> @@ -881,8 +881,8 @@ vhost_user_get_vring_base(struct virtio_net *dev,
>
> dev->flags &= ~VIRTIO_DEV_READY;
>
> - /* Here we are safe to get the last used index */
> - msg->payload.state.num = vq->last_used_idx;
> + /* Here we are safe to get the last avail index */
> + msg->payload.state.num = vq->last_avail_idx;
>
> RTE_LOG(INFO, VHOST_CONFIG,
> "vring base idx:%d file:%d\n", msg->payload.state.index,
> --
> 2.14.1
Re: [dpdk-dev] [PATCH] vhost: fix wait for valid descriptor
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Tomasz Kulasek
> Sent: Saturday, February 10, 2018 1:35 AM
> To: y...@fridaylinux.org
> Cc: dev@dpdk.org; yuanhan@linux.intel.com; sta...@dpdk.org; Liu,
> Changpeng
> Subject: [dpdk-dev] [PATCH] vhost: fix wait for valid descriptor
>
> For each virt queue's kickfd and callfd, there are 2 invalid
> status: VIRTIO_UNINITIALIZED_EVENTFD and VIRTIO_INVALID_EVENTFD.
> Don't set the virt queue to ready status until got the valid
> descriptor.
Cannot remember why we introduce two invalid status.
If we treat "VIRTIO_INVALID_EVENTFD" as an indicator to judge if device is
ready, then defining one status (invalid) is enough.
Thanks,
Jianfeng
>
> This is safe for polling mode drivers in Guest OS, the backend
> vhost process will not post notification to interrupt vector for
> PMD mode in guest, but the interrupt vector still valid.
>
> Fixes: e049ca6d10e0 ("vhost-user: prepare multiple queue setup")
> Cc: yuanhan@linux.intel.com
> Cc: sta...@dpdk.org
>
> Signed-off-by: Changpeng Liu
> Signed-off-by: Tomasz Kulasek
> ---
> lib/librte_vhost/vhost_user.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
> index 65ee33919..4508f697b 100644
> --- a/lib/librte_vhost/vhost_user.c
> +++ b/lib/librte_vhost/vhost_user.c
> @@ -766,7 +766,9 @@ vq_is_ready(struct vhost_virtqueue *vq)
> {
> return vq && vq->desc && vq->avail && vq->used &&
> vq->kickfd != VIRTIO_UNINITIALIZED_EVENTFD &&
> -vq->callfd != VIRTIO_UNINITIALIZED_EVENTFD;
> +vq->callfd != VIRTIO_UNINITIALIZED_EVENTFD &&
> +vq->kickfd != VIRTIO_INVALID_EVENTFD &&
> +vq->callfd != VIRTIO_INVALID_EVENTFD;
> }
>
> static int
> --
> 2.14.1
Re: [dpdk-dev] [PATCH] vhost: reduce size of coredump file
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Tomasz Kulasek
> Sent: Saturday, February 10, 2018 1:40 AM
> To: y...@fridaylinux.org
> Cc: dev@dpdk.org; Basierski, SebastianX
> Subject: [dpdk-dev] [PATCH] vhost: reduce size of coredump file
>
> If application coredumps with vhost-user devices connected to it,
> the generated coredump file size is huge.
>
> To limit its size, this patch adds call to madvise() with MADV_DONTDUMP
> on memory regions mapped from the VM.
As virtqueue is on guest memory, by this patch, the vq information will not be
available in the coredump.
Just wonder if adjusting /proc/pid/coredump_filter can help your case?
Thanks,
Jianfeng
>
> Signed-off-by: Sebastian Basierski
> Signed-off-by: Tomasz Kulasek
> ---
> lib/librte_vhost/vhost_user.c | 5 +
> 1 file changed, 5 insertions(+)
>
> diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
> index 65ee33919..fc1f1a948 100644
> --- a/lib/librte_vhost/vhost_user.c
> +++ b/lib/librte_vhost/vhost_user.c
> @@ -723,6 +723,11 @@ vhost_user_set_mem_table(struct virtio_net *dev,
> struct VhostUserMsg *pmsg)
> goto err_mmap;
> }
>
> + if (madvise(mmap_addr, mmap_size,
> MADV_DONTDUMP) != 0) {
> + RTE_LOG(INFO, VHOST_CONFIG,
> + "MADV_DONTDUMP advice setting
> failed.\n");
> + }
> +
> reg->mmap_addr = mmap_addr;
> reg->mmap_size = mmap_size;
> reg->host_user_addr = (uint64_t)(uintptr_t)mmap_addr +
> --
> 2.14.1
[dpdk-dev] [PATCH] net/i40e: add log when setting input set
This patch adds log when setting input set since global
configuration is changed.
Signed-off-by: Beilei Xing
---
drivers/net/i40e/rte_pmd_i40e.c | 30 +-
1 file changed, 21 insertions(+), 9 deletions(-)
diff --git a/drivers/net/i40e/rte_pmd_i40e.c b/drivers/net/i40e/rte_pmd_i40e.c
index dae59e6..1288c51 100644
--- a/drivers/net/i40e/rte_pmd_i40e.c
+++ b/drivers/net/i40e/rte_pmd_i40e.c
@@ -3071,6 +3071,7 @@ rte_pmd_i40e_inset_set(uint16_t port, uint8_t pctype,
{
struct rte_eth_dev *dev;
struct i40e_hw *hw;
+ struct i40e_pf *pf;
uint64_t inset_reg;
uint32_t mask_reg[2];
int i;
@@ -3086,6 +3087,12 @@ rte_pmd_i40e_inset_set(uint16_t port, uint8_t pctype,
return -EINVAL;
hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+
+ if (pf->support_multi_driver) {
+ PMD_DRV_LOG(ERR, "Input set configuration is not supported.");
+ return -ENOTSUP;
+ }
/* Clear mask first */
for (i = 0; i < 2; i++)
@@ -3098,14 +3105,17 @@ rte_pmd_i40e_inset_set(uint16_t port, uint8_t pctype,
switch (inset_type) {
case INSET_HASH:
- i40e_check_write_reg(hw, I40E_GLQF_HASH_INSET(0, pctype),
-(uint32_t)(inset_reg & UINT32_MAX));
- i40e_check_write_reg(hw, I40E_GLQF_HASH_INSET(1, pctype),
-(uint32_t)((inset_reg >>
- I40E_32_BIT_WIDTH) & UINT32_MAX));
+ i40e_check_write_global_reg(hw, I40E_GLQF_HASH_INSET(0, pctype),
+ (uint32_t)(inset_reg & UINT32_MAX));
+ i40e_check_write_global_reg(hw, I40E_GLQF_HASH_INSET(1, pctype),
+ (uint32_t)((inset_reg >>
+I40E_32_BIT_WIDTH) & UINT32_MAX));
for (i = 0; i < 2; i++)
- i40e_check_write_reg(hw, I40E_GLQF_HASH_MSK(i, pctype),
-mask_reg[i]);
+ i40e_check_write_global_reg(hw,
+ I40E_GLQF_HASH_MSK(i, pctype),
+ mask_reg[i]);
+ i40e_global_cfg_warning(I40E_WARNING_HASH_INSET);
+ i40e_global_cfg_warning(I40E_WARNING_HASH_MSK);
break;
case INSET_FDIR:
i40e_check_write_reg(hw, I40E_PRTQF_FD_INSET(pctype, 0),
@@ -3114,8 +3124,10 @@ rte_pmd_i40e_inset_set(uint16_t port, uint8_t pctype,
(uint32_t)((inset_reg >>
I40E_32_BIT_WIDTH) & UINT32_MAX));
for (i = 0; i < 2; i++)
- i40e_check_write_reg(hw, I40E_GLQF_FD_MSK(i, pctype),
-mask_reg[i]);
+ i40e_check_write_global_reg(hw,
+ I40E_GLQF_FD_MSK(i, pctype),
+ mask_reg[i]);
+ i40e_global_cfg_warning(I40E_WARNING_FD_MSK);
break;
case INSET_FDIR_FLX:
i40e_check_write_reg(hw, I40E_PRTQF_FD_FLXINSET(pctype),
--
2.5.5
[dpdk-dev] [PATCH] net/i40e: remove unnecessary FDIR mask configuration
Remove unnecessary FDIR mask configuration, otherwise flow directory mask will be changed when configuring hash input set. Signed-off-by: Beilei Xing --- drivers/net/i40e/rte_pmd_i40e.c | 4 1 file changed, 4 deletions(-) diff --git a/drivers/net/i40e/rte_pmd_i40e.c b/drivers/net/i40e/rte_pmd_i40e.c index 1288c51..b684249 100644 --- a/drivers/net/i40e/rte_pmd_i40e.c +++ b/drivers/net/i40e/rte_pmd_i40e.c @@ -3094,10 +3094,6 @@ rte_pmd_i40e_inset_set(uint16_t port, uint8_t pctype, return -ENOTSUP; } - /* Clear mask first */ - for (i = 0; i < 2; i++) - i40e_check_write_reg(hw, I40E_GLQF_FD_MSK(i, pctype), 0); - inset_reg = inset->inset; for (i = 0; i < 2; i++) mask_reg[i] = (inset->mask[i].field_idx << 16) | -- 2.5.5
