From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751649AbcBEH4c (ORCPT ); Fri, 5 Feb 2016 02:56:32 -0500 Received: from mail-wm0-f44.google.com ([74.125.82.44]:37583 "EHLO mail-wm0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751328AbcBEH4a (ORCPT ); Fri, 5 Feb 2016 02:56:30 -0500 MIME-Version: 1.0 In-Reply-To: <56B368C2.8080607@redhat.com> References: <1453963027-19770-1-git-send-email-suganath-prabu.subramani@avagotech.com> <1453963027-19770-8-git-send-email-suganath-prabu.subramani@avagotech.com> <56B368C2.8080607@redhat.com> Date: Fri, 5 Feb 2016 13:26:28 +0530 Message-ID: Subject: Re: [mpt3sas driver 07/10] mpt3sas: Add support for configurable Chain Frame Size From: Sreekanth Reddy To: Tomas Henzl Cc: Suganath prabu Subaramani , "James E.J. Bottomley" , "jejb@kernel.org" , Christoph Hellwig , "Martin K. Petersen" , "linux-scsi@vger.kernel.org" , Sathya Prakash , Kashyap Desai , Krishnaraddi Mankani , "linux-kernel@vger.kernel.org" , Chaitra Basappa , Sreekanth Reddy Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Feb 4, 2016 at 8:35 PM, Tomas Henzl wrote: > On 28.1.2016 07:37, Suganath prabu Subaramani wrote: >> From: Suganath prabu Subramani >> >> Added support for configurable Chain Frame Size. Calculate the >> Chain Message Frame size from the IOCMaxChainSegementSize (iocfacts). >> Applicable only for mpt3sas/SAS3.0 HBA's. >> >> Signed-off-by: Suganath prabu Subramani >> Signed-off-by: Chaitra P B > > Is this done in order to prepare for some future patch, or how does it work? > I noticed only changes in allocation of (probably larger) fields, but > how is it actually used ? Tomas, In this patch set we have added new Intruder and Cutlass (high ports) PCI devices which are based on MPI26 version. In this MPI26 version, Request Frame size is doubled when compared to older MPI version, but their is no change in the Chain buffer size from MPI26 to older versions. In older version Request frame size and chain buffer size were same, so in the drivers previously we used to use Request Frame size reference for Chain buffer size. But we can't use that reference now. So now we are determining the chain buffer size from IOCFact's IOCMaxChainSegementSize field instead of using the Request frame size reference. ~Sreekanth > > --tm > >> --- >> drivers/scsi/mpt3sas/mpt3sas_base.c | 27 ++++++++++++++++++++++----- >> drivers/scsi/mpt3sas/mpt3sas_base.h | 8 +++++++- >> 2 files changed, 29 insertions(+), 6 deletions(-) >> >> diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c >> index a1a3b39..0f623fb 100644 >> --- a/drivers/scsi/mpt3sas/mpt3sas_base.c >> +++ b/drivers/scsi/mpt3sas/mpt3sas_base.c >> @@ -3248,6 +3248,19 @@ _base_allocate_memory_pools(struct MPT3SAS_ADAPTER *ioc, int sleep_flag) >> /* reply frame size */ >> ioc->reply_sz = facts->ReplyFrameSize * 4; >> >> + /* chain segment size */ >> + if (ioc->hba_mpi_version_belonged != MPI2_VERSION) { >> + if (facts->IOCMaxChainSegmentSize) >> + ioc->chain_segment_sz = >> + facts->IOCMaxChainSegmentSize * >> + MAX_CHAIN_ELEMT_SZ; >> + else >> + /* set to 128 bytes size if IOCMaxChainSegmentSize is zero */ >> + ioc->chain_segment_sz = DEFAULT_NUM_FWCHAIN_ELEMTS * >> + MAX_CHAIN_ELEMT_SZ; >> + } else >> + ioc->chain_segment_sz = ioc->request_sz; >> + >> /* calculate the max scatter element size */ >> sge_size = max_t(u16, ioc->sge_size, ioc->sge_size_ieee); >> >> @@ -3259,7 +3272,7 @@ _base_allocate_memory_pools(struct MPT3SAS_ADAPTER *ioc, int sleep_flag) >> ioc->max_sges_in_main_message = max_sge_elements/sge_size; >> >> /* now do the same for a chain buffer */ >> - max_sge_elements = ioc->request_sz - sge_size; >> + max_sge_elements = ioc->chain_segment_sz - sge_size; >> ioc->max_sges_in_chain_message = max_sge_elements/sge_size; >> >> /* >> @@ -3451,7 +3464,7 @@ _base_allocate_memory_pools(struct MPT3SAS_ADAPTER *ioc, int sleep_flag) >> goto out; >> } >> ioc->chain_dma_pool = pci_pool_create("chain pool", ioc->pdev, >> - ioc->request_sz, 16, 0); >> + ioc->chain_segment_sz, 16, 0); >> if (!ioc->chain_dma_pool) { >> pr_err(MPT3SAS_FMT "chain_dma_pool: pci_pool_create failed\n", >> ioc->name); >> @@ -3465,13 +3478,13 @@ _base_allocate_memory_pools(struct MPT3SAS_ADAPTER *ioc, int sleep_flag) >> ioc->chain_depth = i; >> goto chain_done; >> } >> - total_sz += ioc->request_sz; >> + total_sz += ioc->chain_segment_sz; >> } >> chain_done: >> dinitprintk(ioc, pr_info(MPT3SAS_FMT >> "chain pool depth(%d), frame_size(%d), pool_size(%d kB)\n", >> - ioc->name, ioc->chain_depth, ioc->request_sz, >> - ((ioc->chain_depth * ioc->request_sz))/1024)); >> + ioc->name, ioc->chain_depth, ioc->chain_segment_sz, >> + ((ioc->chain_depth * ioc->chain_segment_sz))/1024)); >> >> /* initialize hi-priority queue smid's */ >> ioc->hpr_lookup = kcalloc(ioc->hi_priority_depth, >> @@ -4332,6 +4345,10 @@ _base_get_ioc_facts(struct MPT3SAS_ADAPTER *ioc, int sleep_flag) >> facts->FWVersion.Word = le32_to_cpu(mpi_reply.FWVersion.Word); >> facts->IOCRequestFrameSize = >> le16_to_cpu(mpi_reply.IOCRequestFrameSize); >> + if (ioc->hba_mpi_version_belonged != MPI2_VERSION) { >> + facts->IOCMaxChainSegmentSize = >> + le16_to_cpu(mpi_reply.IOCMaxChainSegmentSize); >> + } >> facts->MaxInitiators = le16_to_cpu(mpi_reply.MaxInitiators); >> facts->MaxTargets = le16_to_cpu(mpi_reply.MaxTargets); >> ioc->shost->max_id = -1; >> diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h >> index 4b52a07..d1fee34 100644 >> --- a/drivers/scsi/mpt3sas/mpt3sas_base.h >> +++ b/drivers/scsi/mpt3sas/mpt3sas_base.h >> @@ -129,6 +129,9 @@ >> >> #define MPT3SAS_INVALID_DEVICE_HANDLE 0xFFFF >> >> +#define MAX_CHAIN_ELEMT_SZ 16 >> +#define DEFAULT_NUM_FWCHAIN_ELEMTS 8 >> + >> /* >> * reset phases >> */ >> @@ -759,7 +762,7 @@ struct mpt3sas_facts { >> u32 IOCCapabilities; >> union mpi3_version_union FWVersion; >> u16 IOCRequestFrameSize; >> - u16 Reserved3; >> + u16 IOCMaxChainSegmentSize; >> u16 MaxInitiators; >> u16 MaxTargets; >> u16 MaxSasExpanders; >> @@ -906,6 +909,8 @@ typedef void (*MPT3SAS_FLUSH_RUNNING_CMDS)(struct MPT3SAS_ADAPTER *ioc); >> * @max_sges_in_chain_message: number sg elements per chain >> * @chains_needed_per_io: max chains per io >> * @chain_depth: total chains allocated >> + * @chain_segment_sz: gives the max number of >> + * SGEs accommodate on single chain buffer >> * @hi_priority_smid: >> * @hi_priority: >> * @hi_priority_dma: >> @@ -1113,6 +1118,7 @@ struct MPT3SAS_ADAPTER { >> u16 max_sges_in_chain_message; >> u16 chains_needed_per_io; >> u32 chain_depth; >> + u16 chain_segment_sz; >> >> /* hi-priority queue */ >> u16 hi_priority_smid; >