Re: [ewg] RE: OFED libraries - on the web at last
Todd Rimmer wrote: Here is the change needed to mvapich as included in OFED 1.4.1 However end users will need some notice to rebuild their apps. Especially ISVs. I'm not sure what the OFED policy is for ABI changes, but at least 1-2 year notice seems the minimum (in QLogic's past efforts we try to maintain ABIs much longer than that as we still have some customers running 2-3 year old applications). Todd, We do not remove libibcommon in 1.4.1 So - if having this library is important for 1.5 can someone from Qlogic take care for it? Tziporet ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ewg] Re: [PATCH 2/9] ib_core: kernel API for GID -- MAC translations
On Wed, Jun 17, 2009 at 11:19:04AM -0700, Roland Dreier wrote: diff --git a/include/rdma/ib_user_verbs.h b/include/rdma/ib_user_verbs.h index a17f771..5bfa2e6 100644 --- a/include/rdma/ib_user_verbs.h +++ b/include/rdma/ib_user_verbs.h @@ -81,7 +81,8 @@ enum { IB_USER_VERBS_CMD_MODIFY_SRQ, IB_USER_VERBS_CMD_QUERY_SRQ, IB_USER_VERBS_CMD_DESTROY_SRQ, - IB_USER_VERBS_CMD_POST_SRQ_RECV + IB_USER_VERBS_CMD_POST_SRQ_RECV, + IB_USER_VERBS_CMD_GET_MAC }; This breaks the ABI defined for XRC. So I guess these patches are not really intended to be applied? I definitely want to finish up XRC before starting to look at this series in detail anyway. Roland, on top of which trees would like me to rebase the patches on and update the ABI version? ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] OFED 1.5 Problem on ia64 with your kernel_addons/backport/2.6.22/bitops.h
Jon, The commit below added file kernel_addons/backport/2.6.22/bitops.h. On ia64, we get the following error in the OFED 1.5 build (from e-mail: [ofa-general] ofa_1_5_kernel 20090618-0200 daily build status): Build failed on ia64 with linux-2.6.22 Log: from /home/vlad/tmp/ofa_1_5_kernel-20090618-0200_linux-2.6.22_ia64_check/drivers/infiniband/core/addr.c:36: /home/vlad/tmp/ofa_1_5_kernel-20090618-0200_linux-2.6.22_ia64_check/kernel_addons/backport/2.6.22/include/asm/bitops.h: In function 'clear_bit_unlock': /home/vlad/tmp/ofa_1_5_kernel-20090618-0200_linux-2.6.22_ia64_check/kernel_addons/backport/2.6.22/include/asm/bitops.h:8: error: implicit declaration of function 'smp_mb' make[4]: *** [/home/vlad/tmp/ofa_1_5_kernel-20090618-0200_linux-2.6.22_ia64_check/drivers/infiniband/core/addr.o] Error 1 Problem is here (see ***==): #ifndef BACKPORT_ASM_BITOPS_H #define BACKPORT_ASM_BITOPS_H #include_next asm/bitops.h static inline void clear_bit_unlock(unsigned long nr, volatile unsigned long *addr) { ***== smp_mb__before_clear_bit(); ===** clear_bit(nr, addr); } #endif -Jack == commit efd0ec53f7ee8104d6648b1df494af8413e7db55 Author: Jon Mason j...@opengridcomputing.com Date: Tue Jun 2 18:41:13 2009 -0500 NFSRDMA: NFS backport for 2.6.22 This patch provides the NFS backport for 2.6.22. Since the RDMA infrastructure does not work for this kernel release, only TCP support has been tested. It passes Connectathon as a client and server. Signed-Off-By: Jon Mason j...@opengridcomputing.com ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ewg] [GIT PULL - ofed-1.5] cxgb3 RHEL5.2 backports
Steve Wise wrote: Hey Vlad, Please pull from: ssh://v...@sofa.openfabrics.org/~swise/scm/ofed_kernel.git ofed_1_5 This adds RHEL5.2 backports for cxgb3. Thanks, Steve. Done, Regards, Vladimir ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Re: OFED 1.5 Problem on ia64 with your kernel_addons/backport/2.6.22/bitops.h
I should have the backports for 2.6.25 and 2.6.22 done today. I'll roll the fox for this into those changes. Thanks, Jon On Thu, Jun 18, 2009 at 02:35:55PM +0300, Jack Morgenstein wrote: Jon, The commit below added file kernel_addons/backport/2.6.22/bitops.h. On ia64, we get the following error in the OFED 1.5 build (from e-mail: [ofa-general] ofa_1_5_kernel 20090618-0200 daily build status): Build failed on ia64 with linux-2.6.22 Log: from /home/vlad/tmp/ofa_1_5_kernel-20090618-0200_linux-2.6.22_ia64_check/drivers/infiniband/core/addr.c:36: /home/vlad/tmp/ofa_1_5_kernel-20090618-0200_linux-2.6.22_ia64_check/kernel_addons/backport/2.6.22/include/asm/bitops.h: In function 'clear_bit_unlock': /home/vlad/tmp/ofa_1_5_kernel-20090618-0200_linux-2.6.22_ia64_check/kernel_addons/backport/2.6.22/include/asm/bitops.h:8: error: implicit declaration of function 'smp_mb' make[4]: *** [/home/vlad/tmp/ofa_1_5_kernel-20090618-0200_linux-2.6.22_ia64_check/drivers/infiniband/core/addr.o] Error 1 Problem is here (see ***==): #ifndef BACKPORT_ASM_BITOPS_H #define BACKPORT_ASM_BITOPS_H #include_next asm/bitops.h static inline void clear_bit_unlock(unsigned long nr, volatile unsigned long *addr) { ***== smp_mb__before_clear_bit(); ===** clear_bit(nr, addr); } #endif -Jack == commit efd0ec53f7ee8104d6648b1df494af8413e7db55 Author: Jon Mason j...@opengridcomputing.com Date: Tue Jun 2 18:41:13 2009 -0500 NFSRDMA: NFS backport for 2.6.22 This patch provides the NFS backport for 2.6.22. Since the RDMA infrastructure does not work for this kernel release, only TCP support has been tested. It passes Connectathon as a client and server. Signed-Off-By: Jon Mason j...@opengridcomputing.com ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] RE: [ofa-general] [PATCH 0/9] RDMAoE - RDMA over Ethernet
Let's just say that at this point I completely disagree with where these patches try to abstract the differences, which are many. RDMA apps that want to use this and IB without going through an abstraction will need different code -- just like they would for iWarp, which also provides RDMA over Ethernet, and is a standard. IB mad and SA query modules are not appropriate places for abstracting the differences between IB, iWarp, and whatever name we give this. This could change depending on whether this is really trying to be IB with a different L2, or is just another RDMA protocol that runs on Ethernet. - Sean Sean, These are indeed real concerns; I know that the cma is the natural place for abstracting transport differences, but I am worried about non-cma Infiniband ULPs which can work just as well with RDMAoE (perhaps we can specifically expose RDMAoE path queries as a simple library function). We will rethink our approach to SA queries and post new patches shortly. Note that without SA query emulation, the RDMAoE patches really amount to just a few cosmetic changes to ib_core...:) Thanks for the feedback. --Liran ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ewg] Re: [PATCH 2/9] ib_core: kernel API for GID -- MAC translations
on top of which trees would like me to rebase the patches on and update the ABI version? I guess you could base it on my xrc branch but as I said I'm not planning on merging this until I'm done with the XRC stuff. ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] [GIT PULL ofed-1.5] cxgb3 backports for RHEL 5.3
Vlad, Please pull from: ssh://v...@sofa.openfabrics.org/~swise/scm/ofed_kernel.git ofed_1_5 This adds RH5.3 backports for cxgb3. Thanks, Steve. --- commit 69248ffde5b0a59ec738073368aeee2aa545668a Author: Steve Wise sw...@opengridcomputing.com Date: Thu Jun 18 11:06:57 2009 -0500 cxgb3: rhel5.3 backports for ofed-1.5. Compiles ok. Signed-off-by: Steve Wise sw...@opengridcomputing.com .../backport/2.6.18-EL5.3/include/linux/skbuff.h | 102 .../cxgb3_0001_backout_multq_netdeviceops.patch| 228 .../2.6.18-EL5.3/cxgb3_0002_undo_250.patch | 164 ++ .../2.6.18-EL5.3/cxgb3_0004_undo_240.patch | 86 +++ ...xgb3_0008_pci_dma_mapping_error_to_2_6_26.patch | 17 + .../backport/2.6.18-EL5.3/cxgb3_0010_napi.patch| 592 .../backport/2.6.18-EL5.3/cxgb3_0020_sysfs.patch | 202 +++ .../backport/2.6.18-EL5.3/cxgb3_0030_sset.patch| 34 ++ .../2.6.18-EL5.3/cxgb3_0100_remove_lro.patch | 391 + .../2.6.18-EL5.3/cxgb3_0110_provider_sysfs.patch | 120 .../2.6.18-EL5.3/cxgb3_0120_fixwarnings.patch | 39 ++ 11 files changed, 1975 insertions(+), 0 deletions(-) ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] [PATCH OFED-1.5] install.pl: modify kernels supported by NFSRDMA
Currently, the only vanilla kernels that are supported are 2.6.22, 2.6.25, and 2.6.30. This patch modifies the install.pl to only work for those. With the removal of RHEL5.1 support, the kernel check for RHEL should not install for it. Signed-Off-By: Jon Mason j...@opengridcomputing.com diff --git a/install.pl b/install.pl index 206aca5..bbbccf4 100755 --- a/install.pl +++ b/install.pl @@ -1721,7 +1721,7 @@ sub set_availability } # NFSRDMA -if ($kernel =~ m/2.6.22|2.6.2[6-9]|2.6.30/) { +if ($kernel =~ m/2.6.2[25]|2.6.30/) { $kernel_modules_info{'nfsrdma'}{'available'} = 1; $packages_info{'rnfs-utils'}{'available'} = 1; $packages_info{'rnfs-utils-debuginfo'}{'available'} = 1; @@ -1734,7 +1734,7 @@ sub set_availability elsif ($kernel =~ m/el5/) { my $minor = (split '-', $kernel)[1]; $minor =~ s/(\.el5).*//; -if ($minor = 53) { +if ($minor 53) { $kernel_modules_info{'nfsrdma'}{'available'} = 1; $packages_info{'rnfs-utils'}{'available'} = 1; $packages_info{'rnfs-utils-debuginfo'}{'available'} = 1; ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg