/* $OpenBSD: isp_sbus.c,v 1.20 2002/03/14 01:26:43 millert Exp $ */ /* * SBus specific probe and attach routines for Qlogic ISP SCSI adapters. * * Copyright (c) 1997, 2000 by Matthew Jacob * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice immediately at the beginning of the file, without modification, * this list of conditions, and the following disclaimer. * documentation and/or other materials provided with the distribution. * 2. The name of the author may not be used to endorse or promote products * derived from this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * */ #include #include #include #include #include #include #include #include #include #include #include #include #if defined(ISP_COMPILE_FW) || defined(ISP_COMPILE_1000_FW) #include #endif #define ISP_SBUSIFY_ISPHDR(isp, hdrp) \ ISP_SWAP8((hdrp)->rqs_entry_count, (hdrp)->rqs_entry_type); \ ISP_SWAP8((hdrp)->rqs_flags, (hdrp)->rqs_seqno); #define ISP_SWIZZLE_REQUEST(a, b) \ ISP_SBUSIFY_ISPHDR(a, &(b)->req_header); \ ISP_SWAP8((b)->req_target, (b)->req_lun_trn) static int isp_sbus_rd_isr(struct ispsoftc *, u_int16_t *, u_int16_t *, u_int16_t *); static u_int16_t isp_sbus_rd_reg(struct ispsoftc *, int); static void isp_sbus_wr_reg(struct ispsoftc *, int, u_int16_t); static int isp_sbus_mbxdma(struct ispsoftc *); static int isp_sbus_dmasetup(struct ispsoftc *, struct scsi_xfer *, ispreq_t *, u_int16_t *, u_int16_t); static void isp_sbus_dmateardown(struct ispsoftc *, struct scsi_xfer *, u_int16_t); static int isp_sbus_intr(void *); #ifndef ISP_1000_RISC_CODE #define ISP_1000_RISC_CODE NULL #endif static struct ispmdvec mdvec = { isp_sbus_rd_isr, isp_sbus_rd_reg, isp_sbus_wr_reg, isp_sbus_mbxdma, isp_sbus_dmasetup, isp_sbus_dmateardown, NULL, NULL, NULL, ISP_1000_RISC_CODE, BIU_BURST_ENABLE|BIU_SBUS_CONF1_FIFO_32 }; struct isp_sbussoftc { struct ispsoftc sbus_isp; sdparam sbus_dev; struct intrhand sbus_ih; volatile u_int16_t *sbus_reg; int sbus_node; int sbus_pri; struct ispmdvec sbus_mdvec; int16_t sbus_poff[_NREG_BLKS]; vaddr_t *sbus_kdma_allocs; }; static int isp_match(struct device *, void *, void *); static void isp_sbus_attach(struct device *, struct device *, void *); struct cfattach isp_sbus_ca = { sizeof (struct isp_sbussoftc), isp_match, isp_sbus_attach }; static int isp_match(struct device *parent, void *cfarg, void *aux) { int rv; struct cfdata *cf = cfarg; #ifdef DEBUG static int oneshot = 1; #endif struct confargs *ca = aux; struct romaux *ra = &ca->ca_ra; rv = (strcmp(cf->cf_driver->cd_name, ra->ra_name) == 0 || strcmp("PTI,ptisp", ra->ra_name) == 0 || strcmp("ptisp", ra->ra_name) == 0 || strcmp("SUNW,isp", ra->ra_name) == 0 || strcmp("QLGC,isp", ra->ra_name) == 0); if (rv == 0) return (rv); #ifdef DEBUG if (rv && oneshot) { oneshot = 0; printf("Qlogic ISP Driver, OpenBSD (sbus) Platform Version " "%d.%d Core Version %d.%d\n", ISP_PLATFORM_VERSION_MAJOR, ISP_PLATFORM_VERSION_MINOR, ISP_CORE_VERSION_MAJOR, ISP_CORE_VERSION_MINOR); } #endif if (ca->ca_bustype == BUS_SBUS) return (1); ra->ra_len = NBPG; return (probeget(ra->ra_vaddr, 1) != -1); } static void isp_sbus_attach(struct device *parent, struct device *self, void *aux) { int freq, storebp = 0; struct confargs *ca = aux; struct bootpath *bp; struct isp_sbussoftc *sbc = (struct isp_sbussoftc *) self; struct ispsoftc *isp = &sbc->sbus_isp; if (ca->ca_ra.ra_nintr != 1) { printf(": expected 1 interrupt, got %d\n", ca->ca_ra.ra_nintr); return; } printf("\n"); sbc->sbus_pri = ca->ca_ra.ra_intr[0].int_pri; sbc->sbus_mdvec = mdvec; if (ca->ca_ra.ra_vaddr) { sbc->sbus_reg = (volatile u_int16_t *) ca->ca_ra.ra_vaddr; } else { sbc->sbus_reg = (volatile u_int16_t *) mapiodev(ca->ca_ra.ra_reg, 0, ca->ca_ra.ra_len); } sbc->sbus_node = ca->ca_ra.ra_node; freq = getpropint(ca->ca_ra.ra_node, "clock-frequency", 0); if (freq) { /* * Convert from HZ to MHz, rounding up. */ freq = (freq + 500000)/1000000; #if 0 printf("%s: %d MHz\n", self->dv_xname, freq); #endif } sbc->sbus_mdvec.dv_clock = freq; if ((bp = ca->ca_ra.ra_bp) != NULL) { if (bp->val[0] == ca->ca_slot && bp->val[1] == ca->ca_offset) { if (strcmp("isp", bp->name) == 0 || strcmp("QLGC,isp", bp->name) == 0 || strcmp("PTI,isp", bp->name) == 0 || strcmp("ptisp", bp->name) == 0) { storebp = 1; } } } /* * XXX: Now figure out what the proper burst sizes, etc., to use. */ sbc->sbus_mdvec.dv_conf1 |= BIU_SBUS_CONF1_FIFO_8; /* * Some early versions of the PTI SBus adapter * would fail in trying to download (via poking) * FW. We give up on them. */ if (strcmp("PTI,ptisp", ca->ca_ra.ra_name) == 0 || strcmp("ptisp", ca->ca_ra.ra_name) == 0) { sbc->sbus_mdvec.dv_ispfw = NULL; } isp->isp_mdvec = &sbc->sbus_mdvec; isp->isp_bustype = ISP_BT_SBUS; isp->isp_type = ISP_HA_SCSI_UNKNOWN; isp->isp_param = &sbc->sbus_dev; bzero(isp->isp_param, sizeof (sdparam)); sbc->sbus_poff[BIU_BLOCK >> _BLK_REG_SHFT] = BIU_REGS_OFF; sbc->sbus_poff[MBOX_BLOCK >> _BLK_REG_SHFT] = SBUS_MBOX_REGS_OFF; sbc->sbus_poff[SXP_BLOCK >> _BLK_REG_SHFT] = SBUS_SXP_REGS_OFF; sbc->sbus_poff[RISC_BLOCK >> _BLK_REG_SHFT] = SBUS_RISC_REGS_OFF; sbc->sbus_poff[DMA_BLOCK >> _BLK_REG_SHFT] = DMA_REGS_OFF; /* Establish interrupt channel */ sbc->sbus_ih.ih_fun = (void *) isp_sbus_intr; sbc->sbus_ih.ih_arg = sbc; intr_establish(sbc->sbus_pri, &sbc->sbus_ih); /* * Set up logging levels. */ #ifdef ISP_LOGDEFAULT isp->isp_dblev = ISP_LOGDEFAULT; #else isp->isp_dblev = ISP_LOGWARN|ISP_LOGERR; #ifdef SCSIDEBUG isp->isp_dblev |= ISP_LOGDEBUG1|ISP_LOGDEBUG2; #endif #ifdef DEBUG isp->isp_dblev |= ISP_LOGDEBUG0|ISP_LOGCONFIG|ISP_LOGINFO; #endif #endif isp->isp_confopts = self->dv_cfdata->cf_flags; isp->isp_role = ISP_DEFAULT_ROLES; ISP_LOCK(isp); isp->isp_osinfo.no_mbox_ints = 1; isp_reset(isp); if (isp->isp_state != ISP_RESETSTATE) { ISP_UNLOCK(isp); return; } ENABLE_INTS(isp); isp_init(isp); if (isp->isp_state != ISP_INITSTATE) { isp_uninit(isp); ISP_UNLOCK(isp); return; } /* * do generic attach. */ if (storebp) { /* * We're the booting HBA. * * Override the bootpath name with our driver name * so we will do the correct matching and and store * the next component's boot path entry, also so a * successful match will occur. */ bcopy("isp", bp->name, 4); bp++; bootpath_store(1, bp); } isp_attach(isp); if (isp->isp_state != ISP_RUNSTATE) { isp_uninit(isp); } if (storebp) { bootpath_store(1, NULL); } ISP_UNLOCK(isp); } #define IspVirt2Off(a, x) \ (((struct isp_sbussoftc *)a)->sbus_poff[((x) & _BLK_REG_MASK) >> \ _BLK_REG_SHFT] + ((x) & 0xff)) #define BXR2(pcs, off) (sbc->sbus_reg[off >> 1]) static int isp_sbus_rd_isr(struct ispsoftc *isp, u_int16_t *isrp, u_int16_t *semap, u_int16_t *mbp) { struct isp_sbussoftc *sbc = (struct isp_sbussoftc *) isp; volatile u_int16_t isr, sema; isr = BXR2(pcs, IspVirt2Off(isp, BIU_ISR)); sema = BXR2(pcs, IspVirt2Off(isp, BIU_SEMA)); isp_prt(isp, ISP_LOGDEBUG3, "ISR 0x%x SEMA 0x%x", isr, sema); isr &= INT_PENDING_MASK(isp); sema &= BIU_SEMA_LOCK; if (isr == 0 && sema == 0) { return (0); } *isrp = isr; if ((*semap = sema) != 0) { *mbp = BXR2(pcs, IspVirt2Off(isp, OUTMAILBOX0)); } return (1); } static u_int16_t isp_sbus_rd_reg(struct ispsoftc *isp, int regoff) { struct isp_sbussoftc *sbc = (struct isp_sbussoftc *) isp; int offset = sbc->sbus_poff[(regoff & _BLK_REG_MASK) >> _BLK_REG_SHFT]; offset += (regoff & 0xff); return ((u_int16_t) sbc->sbus_reg[offset >> 1]); } static void isp_sbus_wr_reg(struct ispsoftc *isp, int regoff, u_int16_t val) { struct isp_sbussoftc *sbc = (struct isp_sbussoftc *) isp; int offset = sbc->sbus_poff[(regoff & _BLK_REG_MASK) >> _BLK_REG_SHFT]; offset += (regoff & 0xff); sbc->sbus_reg[offset >> 1] = val; } static int isp_sbus_mbxdma(struct ispsoftc *isp) { struct isp_sbussoftc *sbc = (struct isp_sbussoftc *) isp; size_t len; if (isp->isp_rquest_dma) /* been here before? */ return (0); /* * NOTE: Since most Sun machines aren't I/O coherent, * map the mailboxes through kdvma space to force them * to be uncached. */ len = isp->isp_maxcmds * sizeof (XS_T); isp->isp_xflist = (XS_T **) malloc(len, M_DEVBUF, M_WAITOK); bzero(isp->isp_xflist, len); len = isp->isp_maxcmds * sizeof (vaddr_t); sbc->sbus_kdma_allocs = (vaddr_t *) malloc(len, M_DEVBUF, M_WAITOK); bzero(sbc->sbus_kdma_allocs, len); /* * Allocate and map the request queue. */ len = ISP_QUEUE_SIZE(RQUEST_QUEUE_LEN(isp)); isp->isp_rquest = (volatile caddr_t)malloc(len, M_DEVBUF, M_NOWAIT); if (isp->isp_rquest == 0) { printf("%s: cannot allocate request queue\n", isp->isp_name); return (1); } isp->isp_rquest_dma = (u_int32_t) kdvma_mapin((caddr_t)isp->isp_rquest, len, 0); if (isp->isp_rquest_dma == 0) { printf("%s: could not mapin request queue\n", isp->isp_name); return (1); } /* * Allocate and map the result queue. */ len = ISP_QUEUE_SIZE(RESULT_QUEUE_LEN(isp)); isp->isp_result = (volatile caddr_t)malloc(len, M_DEVBUF, M_NOWAIT); if (isp->isp_result == 0) { printf("%s: cannot allocate result queue\n", isp->isp_name); return (1); } isp->isp_result_dma = (u_int32_t) kdvma_mapin((caddr_t)isp->isp_result, len, 0); if (isp->isp_result_dma == 0) { printf("%s: could not mapin result queue\n", isp->isp_name); return (1); } return (0); } /* * TODO: If kdvma_mapin fails, try using multiple smaller chunks.. */ static int isp_sbus_dmasetup(struct ispsoftc *isp, struct scsi_xfer *xs, ispreq_t *rq, u_int16_t *iptrp, u_int16_t optr) { struct isp_sbussoftc *sbc = (struct isp_sbussoftc *) isp; ispreq_t *qe; ispcontreq_t *crq; vaddr_t kdvma; int dosleep = (xs->flags & SCSI_NOSLEEP) != 0; qe = (ispreq_t *) ISP_QUEUE_ENTRY(isp->isp_rquest, isp->isp_reqidx); if (xs->datalen == 0) { rq->req_seg_count = 1; goto mbxsync; } if (CPU_ISSUN4M) { kdvma = (vaddr_t) kdvma_mapin((caddr_t)xs->data, xs->datalen, dosleep); if (kdvma == (vaddr_t) 0) { XS_SETERR(xs, HBA_BOTCH); return (CMD_COMPLETE); } } else { kdvma = (vaddr_t) xs->data; } if (sbc->sbus_kdma_allocs[isp_handle_index(rq->req_handle)] != 0) { panic("%s: kdma handle already allocated\n", isp->isp_name); /* NOTREACHED */ } if (XS_CDBLEN(xs) > 12) { crq = (ispcontreq_t *) ISP_QUEUE_ENTRY(isp->isp_rquest, *iptrp); *iptrp = ISP_NXT_QENTRY(*iptrp, RQUEST_QUEUE_LEN(isp)); if (*iptrp == optr) { printf("%s: Request Queue Overflow++\n", isp->isp_name); if (CPU_ISSUN4M) { dvma_mapout(kdvma, (vaddr_t) xs->data, xs->datalen); } XS_SETERR(xs, HBA_BOTCH); return (CMD_EAGAIN); } } else { crq = NULL; } sbc->sbus_kdma_allocs[isp_handle_index(rq->req_handle)] = kdvma; if (xs->flags & SCSI_DATA_IN) { rq->req_flags |= REQFLAG_DATA_IN; } else { rq->req_flags |= REQFLAG_DATA_OUT; } if (crq) { rq->req_seg_count = 2; rq->req_dataseg[0].ds_count = 0; rq->req_dataseg[0].ds_base = 0; bzero((void *)crq, sizeof (*crq)); crq->req_header.rqs_entry_count = 1; crq->req_header.rqs_entry_type = RQSTYPE_DATASEG; crq->req_dataseg[0].ds_count = xs->datalen; crq->req_dataseg[0].ds_base = (u_int32_t) kdvma; ISP_SBUSIFY_ISPHDR(isp, &crq->req_header) } else { rq->req_dataseg[0].ds_count = xs->datalen; rq->req_dataseg[0].ds_base = (u_int32_t) kdvma; rq->req_seg_count = 1; } mbxsync: ISP_SWIZZLE_REQUEST(isp, rq); bcopy(rq, qe, sizeof (ispreq_t)); return (CMD_QUEUED); } static void isp_sbus_dmateardown(struct ispsoftc *isp, XS_T *xs, u_int16_t handle) { struct isp_sbussoftc *sbc = (struct isp_sbussoftc *) isp; vaddr_t kdvma; if (xs->flags & SCSI_DATA_IN) { cpuinfo.cache_flush(xs->data, xs->datalen - xs->resid); } if (sbc->sbus_kdma_allocs[isp_handle_index(handle)] == (vaddr_t) 0) { panic("%s: kdma handle not already allocated\n", isp->isp_name); /* NOTREACHED */ } kdvma = sbc->sbus_kdma_allocs[isp_handle_index(handle)]; sbc->sbus_kdma_allocs[isp_handle_index(handle)] = (vaddr_t) 0; if (CPU_ISSUN4M) { dvma_mapout(kdvma, (vaddr_t) xs->data, xs->datalen); } } static int isp_sbus_intr(void *arg) { u_int16_t isr, sema, mbox; struct ispsoftc *isp = (struct ispsoftc *)arg; isp->isp_intcnt++; if (ISP_READ_ISR(isp, &isr, &sema, &mbox) == 0) { isp->isp_intbogus++; return (0); } else { isp->isp_osinfo.onintstack = 1; isp_intr(isp, isr, sema, mbox); isp->isp_osinfo.onintstack = 0; return (1); } }