On Mon, 2010-11-29 at 16:56 +0100, Hannes Reinecke wrote: > On 11/28/2010 08:28 AM, Nicholas A. Bellinger wrote: > > On Sat, 2010-11-27 at 23:17 -0800, Nicholas A. Bellinger wrote: > >> From: Nicholas Bellinger <nab@xxxxxxxxxxxxxxx> > >> > >> Greetings Hannes, Bo and Co, > >> > >> This series converts megaraid_sas to run in modern host_lock-less mode > >> for >= .37-rc3+ with interrupts disabled internally around megasas_instance->hba_lock. > >> This series is currently living in lio-core-2.6.git/lock_less-LLDs-for-38-v2, > >> and is intended for .38 mainline code. > >> > >> The first patch adds a handful of missing barriers around instance->fw_outstanding > >> usage w/ atomic_add() and atomic_dec(). > >> > >> The second converts instance->issuepend_done to an atomic_t, along with > >> the necessary assignments in order to run w/ Scsi_Host->host_lock, and > >> without instance->hba_lock. > >> > >> The third patch does the actual conversion, and adds a __megasas_get_cmd() > >> usased by megasas_queue_command() w/ instance->hba_lock, along with > >> being held for megasas_build_ldio() and megasas_build_dcdb() in order > >> to locate the proper frame for struct megasas_cmd. This is really the > >> one major change in order to get host_lock-less to function with interrupts > >> disabled around hba_lock. > >> > >> So far this has been tested with Hannes's QEMU 8708EM2 HBA emulation with > >> TCM_Loop backends using SG_IO from KVM host in a paired Host/Guest .37-rc3 > >> environment. This has not been tested on real silicon yet, but I believe > >> this series should be working there as well. > >> > > > > And quick screenshot running in KVM Host TCM_Loop host_lock-less mode > > into KVM Guest megaraid_sas host_lock-less mode w/ small block LTP > > disktest. > > > > http://www.linux-iscsi.org/index.php/File:TCM_loop-megasas-37-rc3.png > > > > Just a FYI, with scsi-generic <-> TCM_Loop everything appears to be > > stable for both small and large block size tests. > > > Oh, good. I did some initial tests with megaraid_sas & host_lock > disabled, but ran into the problems you mentioned. > And have been too lazy to pursue this problem further. > Thanks for your work; I'll give it a spin and see if the throughput > speed has increased. > > Thanks for the patches. > Aside from the extra __megasas_get_cmd() breakage, the conversion was really quite straight-forward.. ;) > Btw, I've found one issue wrt Windows7 booting; now we're not > crashing anymore. But we're not exactly progressing, either; > apparently Win7 fails to notify that we've finished the commands. > Continue debugging. > Ok, then we probably will need to verify that our v0.12.5-megasas in the following qemu-kvm.git branch is still working as expected w/ your Win7 build: windows7-megasas-working dde5ee2 [scsi-bus]: Add MAINTENANCE_IN and \ MAINTENANCE_OUT case for SCSIRequest xfer and xfer_mode setup Just for reference, this is back with the hardcoded defs for fw_sge and fw_cmds: /* Static definitions */ #define MEGASAS_MAX_FRAMES 1000 #define MEGASAS_MAX_SGE 8 #define MEGASAS_MAX_LUNS 128 Perhaps the larger default fw_sge in recent megasas emulation code is causing an issue here..? --nab -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
