X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=doc%2Fguides%2Frel_notes%2Fknown_issues.rst;h=7b620853585a38fbec5e56266b1de755de3270d1;hb=4d4ebca430474180cdc5cb14edfda04ee4189d19;hp=d72dfd8a3f79ebe7f8748bd784a13627acbc4434;hpb=ac8ada004c12df2473f854f72e8d0ca1de3d4e60;p=dpdk.git diff --git a/doc/guides/rel_notes/known_issues.rst b/doc/guides/rel_notes/known_issues.rst index d72dfd8a3f..7b62085358 100644 --- a/doc/guides/rel_notes/known_issues.rst +++ b/doc/guides/rel_notes/known_issues.rst @@ -31,175 +31,62 @@ Known Issues and Limitations ============================ -This section describes known issues with the DPDK software, Release 1.8.0. - -Pause Frame Forwarding does not work properly on igb ----------------------------------------------------- +This section describes known issues with the DPDK software. +Unit Test for Link Bonding may fail at test_tlb_tx_burst() +---------------------------------------------------------- +--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Pause Frame forwarding does not work properly on igb | +| Title | Unit Test for Link Bonding may fail at test_tlb_tx_burst() | | | | +================================+======================================================================================+ -| Reference # | IXA00384637 | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Description | For igb devices rte_eth_flow_ctrl_set is not working as expected. | -| | Pause frames are always forwarded on igb, regardless of the RFCE, MPMCF and DPF | -| | registers. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | Pause frames will never be rejected by the host on 1G NICs and they will always be | -| | forwarded. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution/ Workaround | There is no workaround available. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | All | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | Poll Mode Driver (PMD) | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ - -Running TestPMD with SRIOV in Domain U may cause it to hang when XENVIRT switch is on -------------------------------------------------------------------------------------- - -+--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Running TestPMD with SRIOV in Domain U may cause it to hang when XENVIRT switch is on| -| | | -+================================+======================================================================================+ -| Reference # | IXA00168949 | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Description | When TestPMD is run with only SRIOV port “./testpmd -c f -n 4 -- -i” , the following | -| | error occurs: | -| | | -| | PMD: gntalloc: ioctl error | -| | | -| | EAL: Error - exiting with code: 1 | -| | | -| | Cause: Creation of mbuf pool for socket 0 failed | -| | | -| | Then, alternately run SRIOV port and virtIO with testpmd: | -| | | -| | testpmd -c f -n 4 -- -i | -| | | -| | testpmd -c f -n 4 --use-dev="eth_xenvirt0" -- -i | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | DomU will not be accessible after you repeat this action some times | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution/ Workaround | Run testpmd with a "--total-num-mbufs=N(N<=3500)" | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | Fedora 16, 64 bits + Xen hypervisor 4.2.3 + Domain 0 kernel 3.10.0 | -| | +Domain U kernel 3.6.11 | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | TestPMD Sample Application | +| Reference # | IXA00390304 | | | | +--------------------------------+--------------------------------------------------------------------------------------+ - -Vhost-xen cannot detect Domain U application exit on Xen version 4.0.1 ----------------------------------------------------------------------- - -+--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Vhost-xen cannot detect Domain U application exit on Xen 4.0.1. | -| | | -+================================+======================================================================================+ -| Reference # | IXA00168947 | +| Description | Unit tests will fail at test_tlb_tx_burst function with error for uneven distribution| +| | of packets. | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Description | When using DPDK applications on Xen 4.0.1, e.g. TestPMD Sample Application, | -| | on killing the application (e.g. killall testmd) vhost-switch cannot detect | -| | the domain U exited and does not free the Virtio device. | +| Implication | Unit test link_bonding_autotest will fail | | | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | Virtio device not freed after application is killed when using vhost-switch on Xen | -| | 4.0.1 | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution | | +| Resolution/ Workaround | There is no workaround available. | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | Xen 4.0.1 | +| Affected Environment/ Platform | Fedora 20 | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | Vhost-switch | +| Driver/Module | Link Bonding | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -Virtio incorrect header length used if MSI-X is disabled by kernel driver -------------------------------------------------------------------------- -+--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Virtio incorrect header length used if MSI-X is disabled by kernel driver or | -| | if VIRTIO_NET_F_MAC is not negotiated. | -| | | -+================================+======================================================================================+ -| Reference # | IXA00384256 | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Description | The Virtio header for host-guest communication is of variable length and | -| | is dependent on whether MSI-X has been enabled by the kernel driver for the network | -| | device. | -| | | -| | The base header length of 20 bytes will be extended by 4 bytes to accommodate MSI-X | -| | vectors and the Virtio Network Device header will appear at byte offset 24. | -| | | -| | The Userspace Virtio Poll Mode Driver tests the guest feature bits for the presence | -| | of VIRTIO_PCI_FLAG_MISIX, however this bit field is not part of the Virtio | -| | specification and resolves to the VIRTIO_NET_F_MAC feature instead. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | The DPDK kernel driver will enable MSI-X by default, | -| | however if loaded with “intr_mode=legacy” on a guest with a Virtio Network Device, | -| | a KVM-Qemu guest may crash with the following error: “virtio-net header not in first | -| | element”. | -| | | -| | If VIRTIO_NET_F_MAC feature has not been negotiated, then the Userspace Poll Mode | -| | Driver will assume that MSI-X has been disabled and will prevent the proper | -| | functioning of the driver. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution | Ensure #define VIRTIO_PCI_CONFIG(hw) returns the correct offset (20 or 24 bytes) for | -| | the devices where in rare cases MSI-X is disabled or VIRTIO_NET_F_MAC has not been | -| | negotiated. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | Virtio devices where MSI-X is disabled or VIRTIO_NET_F_MAC feature has not been | -| | negotiated. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | librte_pmd_virtio | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ - -gmake clean may silently fail for some example applications ------------------------------------------------------------ +Pause Frame Forwarding does not work properly on igb +---------------------------------------------------- +--------------------------------+--------------------------------------------------------------------------------------+ -| Title | When using Free BSD* 9.2 gmake clean may silently fail for some sample applications | +| Title | Pause Frame forwarding does not work properly on igb | | | | +================================+======================================================================================+ -| Reference # | IXA00834605 | +| Reference # | IXA00384637 | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Description | gmake clean may silently fail leaving the source object files intact. | +| Description | For igb devices rte_eth_flow_ctrl_set is not working as expected. | +| | Pause frames are always forwarded on igb, regardless of the RFCE, MPMCF and DPF | +| | registers. | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | Application object files are not removed. | +| Implication | Pause frames will never be rejected by the host on 1G NICs and they will always be | +| | forwarded. | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution | Manually remove ./build folders or rebuild application source after editing. | +| Resolution/ Workaround | There is no workaround available. | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | FreeBSD* 9.2 and below | +| Affected Environment/ Platform | All | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | Example Applications | +| Driver/Module | Poll Mode Driver (PMD) | | | | +--------------------------------+--------------------------------------------------------------------------------------+ @@ -338,6 +225,12 @@ Not all variants of supported NIC types have been used in testing | | | | | The NIC device identifiers used during testing: | | | | +| | * Intel® Ethernet Controller XL710 for 40GbE QSFP+ [8086:1584] | +| | | +| | * Intel® Ethernet Controller XL710 for 40GbE QSFP+ [8086:1583] | +| | | +| | * Intel® Ethernet Controller X710 for 10GbE SFP+ [8086:1572] | +| | | | | * Intel® 82576 Gigabit Ethernet Controller [8086:10c9] | | | | | | * Intel® 82576 Quad Copper Gigabit Ethernet Controller [8086:10e8] | @@ -409,35 +302,6 @@ Multi-process sample app requires exact memory mapping | | | +--------------------------------+--------------------------------------------------------------------------------------+ -Unstable system performance across application executions with 2MB pages ------------------------------------------------------------------------- - -+--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Unstable system performance across application executions with 2MB pages | -| | | -+================================+======================================================================================+ -| Reference # | IXA00372346 | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Description | The performance of a DPDK application may vary across executions of an | -| | application due to a varying number of TLB misses depending on the location of | -| | accessed structures in memory. | -| | This situation occurs on rare occasions. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | Occasionally, relatively poor performance of DPDK applications is encountered. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution/ Workaround | Using 1 GB pages results in lower usage of TLB entries, resolving this issue. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | Systems using 2 MB pages | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | All | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ - Packets are not sent by the 1 GbE/10 GbE SR-IOV driver when the source MAC address is not the MAC address assigned to the VF NIC -------------------------------------------------------------------------------------------------------------------------------- @@ -568,7 +432,7 @@ Some hardware off-load functions are not supported by the VF Driver | | | +--------------------------------+--------------------------------------------------------------------------------------+ | Implication | Any configuration for these items in the VF register will be ignored. The behavior | -| | is dependant on the current PF setting. | +| | is dependent on the current PF setting. | | | | +--------------------------------+--------------------------------------------------------------------------------------+ | Resolution/ Workaround | For the PF (Physical Function) status on which the VF driver depends, there is an | @@ -611,33 +475,6 @@ Kernel crash on IGB port unbinding | | | +--------------------------------+--------------------------------------------------------------------------------------+ -Link status change not working with MSI interrupts --------------------------------------------------- - -+--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Link status change not working with MSI interrupts | -| | | -+================================+======================================================================================+ -| Reference # | IXA00378191 | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Description | MSI interrupts are not supported by the PMD. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | Link status change will only work with legacy or MSI-X interrupts. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution/ Workaround | The igb_uio driver can now be loaded with either legacy or MSI-X interrupt support. | -| | However, this configuration is not tested. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | All | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | Poll Mode Driver (PMD) | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ - Twinpond and Ironpond NICs do not report link status correctly -------------------------------------------------------------- @@ -815,34 +652,6 @@ Differences in how different Intel NICs handle maximum packet length for jumbo f | | | +--------------------------------+--------------------------------------------------------------------------------------+ -Link status interrupt not working in VF drivers ------------------------------------------------ - -+--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Link status interrupts not working in the VF drivers | -| | | -+================================+======================================================================================+ -| Reference | IXA00381312 | -| # | | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Description | Due to the driver not setting up interrupts for VF drivers, the NIC does not report | -| | link status change to VF devices. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | Link status interrupts will not work in VM guests. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution/ Workaround | There is no workaround available. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | All | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | VF (SR-IOV) Poll Mode Driver (PMD) | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ - Binding PCI devices to igb_uio fails on Linux* kernel 3.9 when more than one device is used ------------------------------------------------------------------------------------------- @@ -874,8 +683,8 @@ Binding PCI devices to igb_uio fails on Linux* kernel 3.9 when more than one dev | | | +--------------------------------+--------------------------------------------------------------------------------------+ -GCC might generate Intel® AVX instructions forprocessors without Intel® AVX support ------------------------------------------------------------------------------------ +GCC might generate Intel® AVX instructions for processors without Intel® AVX support +------------------------------------------------------------------------------------ +--------------------------------+--------------------------------------------------------------------------------------+ | Title | Gcc might generate Intel® AVX instructions for processors without Intel® AVX support | @@ -940,41 +749,11 @@ Ethertype filter could receive other packets (non-assigned) in Niantic | | | +--------------------------------+--------------------------------------------------------------------------------------+ -Double VLAN does not work on Intel® 40G ethernet controller ------------------------------------------------------------ - -+--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Double VLAN does not work on Intel® 40G ethernet controller | -| | | -+================================+======================================================================================+ -| Reference # | IXA00386480 | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Description | On Intel® 40G Ethernet Controller: | -| | | -| | Double VLAN does not work. This was confirmed a firmware issue which will be fixed | -| | in later versions of firmware. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | After setting double vlan to be enabled on a port, no packets can be transmitted out | -| | on that port. | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Resolution/ Workaround | None | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Affected Environment/ Platform | All | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ -| Driver/Module | Poll Mode Driver (PMD) | -| | | -+--------------------------------+--------------------------------------------------------------------------------------+ - -Cannot set link speed on Intel® 40G ethernet controller +Cannot set link speed on Intel® 40G Ethernet controller ------------------------------------------------------- +--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Cannot set link speed on Intel® 40G ethernet controller | +| Title | Cannot set link speed on Intel® 40G Ethernet controller | | | | +================================+======================================================================================+ | Reference # | IXA00386379 | @@ -985,7 +764,7 @@ Cannot set link speed on Intel® 40G ethernet controller | | It cannot set the link to specific speed. | | | | +--------------------------------+--------------------------------------------------------------------------------------+ -| Implication | The link speed cannot be changed forcedly, though it can be configured by | +| Implication | The link speed cannot be changed forcibly, though it can be configured by | | | application. | | | | +--------------------------------+--------------------------------------------------------------------------------------+ @@ -999,11 +778,11 @@ Cannot set link speed on Intel® 40G ethernet controller | | | +--------------------------------+--------------------------------------------------------------------------------------+ -Stopping the port does not down the link on Intel® 40G ethernet controller +Stopping the port does not down the link on Intel® 40G Ethernet controller -------------------------------------------------------------------------- +--------------------------------+--------------------------------------------------------------------------------------+ -| Title | Stopping the port does not down the link on Intel® 40G ethernet controller | +| Title | Stopping the port does not down the link on Intel® 40G Ethernet controller | | | | +================================+======================================================================================+ | Reference # | IXA00386380 | @@ -1026,3 +805,35 @@ Stopping the port does not down the link on Intel® 40G ethernet controller | Driver/Module | Poll Mode Driver (PMD) | | | | +--------------------------------+--------------------------------------------------------------------------------------+ + +Devices bound to igb_uio with VT-d enabled do not work on Linux* kernel 3.15-3.17 +--------------------------------------------------------------------------------- + ++--------------------------------+--------------------------------------------------------------------------------------+ +| Title | Devices bound to igb_uio with VT-d enabled do not work on Linux* kernel 3.15-3.17 | ++================================+======================================================================================+ +| Description | | When VT-d is enabled (iommu=pt intel_iommu=on), devices are 1:1 mapped. | +| | In the Linux* kernel unbinding devices from drivers removes that mapping which | +| | result in IOMMU errors. | +| | | Introduced in Linux `kernel 3.15 commit `_, | +| | solved in Linux `kernel 3.18 commit `_. | ++--------------------------------+--------------------------------------------------------------------------------------+ +| Implication | | Devices will not be allowed to access memory, resulting in following kernel errors:| +| | | ``dmar: DRHD: handling fault status reg 2`` | +| | | ``dmar: DMAR:[DMA Read] Request device [02:00.0] fault addr a0c58000`` | +| | | ``DMAR:[fault reason 02] Present bit in context entry is clear`` | ++--------------------------------+--------------------------------------------------------------------------------------+ +| Resolution/ Workaround | | Use earlier or later kernel versions, or avoid driver binding on boot by | +| | blacklisting the driver modules. | +| | | ie. in the case of ixgbe, we can pass the kernel command line option: | +| | | ``modprobe.blacklist=ixgbe`` | +| | | This way we do not need to unbind the device to bind it to igb_uio. | ++--------------------------------+--------------------------------------------------------------------------------------+ +| Affected Environment/ Platform | Linux* systems with kernel versions 3.15 to 3.17 | ++--------------------------------+--------------------------------------------------------------------------------------+ +| Driver/Module | igb_uio module | ++--------------------------------+--------------------------------------------------------------------------------------+