Revert "staging: tidspbridge: replace iommu custom for opensource implementation"

This reverts commit d95ec7e2fd5cebf2f1caf3f572fa5e0a820ac5b1.

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Omar Ramirez Luna <omar.ramirez@ti.com>

drivers/staging/tidspbridge/core/_tiomap.h

@@ -23,8 +23,6 @@
 #include <plat/clockdomain.h>
 #include <mach-omap2/prm-regbits-34xx.h>
 #include <mach-omap2/cm-regbits-34xx.h>
-#include <plat/iommu.h>
-#include <plat/iovmm.h>
 #include <dspbridge/devdefs.h>
 #include <hw_defs.h>
 #include <dspbridge/dspioctl.h>	/* for bridge_ioctl_extproc defn */
@@ -330,7 +328,7 @@ struct bridge_dev_context {
 	u32 dw_internal_size;	/* Internal memory size */
 
 	struct omap_mbox *mbox;		/* Mail box handle */
-	struct iommu *dsp_mmu;      /* iommu for iva2 handler */
+
 	struct cfg_hostres *resources;	/* Host Resources */
 
 	/*

drivers/staging/tidspbridge/core/io_sm.c

@@ -291,7 +291,6 @@ int bridge_io_on_loaded(struct io_mgr *hio_mgr)
 	struct cod_manager *cod_man;
 	struct chnl_mgr *hchnl_mgr;
 	struct msg_mgr *hmsg_mgr;
-	struct iommu *mmu;
 	u32 ul_shm_base;
 	u32 ul_shm_base_offset;
 	u32 ul_shm_limit;
@@ -314,6 +313,7 @@ int bridge_io_on_loaded(struct io_mgr *hio_mgr)
 	struct bridge_ioctl_extproc ae_proc[BRDIOCTL_NUMOFMMUTLB];
 	struct cfg_hostres *host_res;
 	struct bridge_dev_context *pbridge_context;
+	u32 map_attrs;
 	u32 shm0_end;
 	u32 ul_dyn_ext_base;
 	u32 ul_seg1_size = 0;
@@ -337,20 +337,6 @@ int bridge_io_on_loaded(struct io_mgr *hio_mgr)
 		status = -EFAULT;
 		goto func_end;
 	}
-	mmu = pbridge_context->dsp_mmu;
-
-	if (mmu)
-		iommu_put(mmu);
-	mmu = iommu_get("iva2");
-
-	if (IS_ERR_OR_NULL(mmu)) {
-		dev_err(bridge, "iommu_get failed!\n");
-		pbridge_context->dsp_mmu = NULL;
-		status = -EFAULT;
-		goto func_end;
-	}
-	pbridge_context->dsp_mmu = mmu;
-
 	status = dev_get_cod_mgr(hio_mgr->hdev_obj, &cod_man);
 	if (!cod_man) {
 		status = -EFAULT;
@@ -490,16 +476,55 @@ int bridge_io_on_loaded(struct io_mgr *hio_mgr)
 	gpp_va_curr = ul_gpp_va;
 	num_bytes = ul_seg1_size;
 
-	va_curr = iommu_kmap(mmu, va_curr, pa_curr, num_bytes,
-					IOVMF_ENDIAN_LITTLE | IOVMF_ELSZ_32);
-	if (IS_ERR_VALUE(va_curr)) {
-		status = (int)va_curr;
-		goto func_end;
-	}
+	/*
+	 * Try to fit into TLB entries. If not possible, push them to page
+	 * tables. It is quite possible that if sections are not on
+	 * bigger page boundary, we may end up making several small pages.
+	 * So, push them onto page tables, if that is the case.
+	 */
+	map_attrs = 0x00000000;
+	map_attrs = DSP_MAPLITTLEENDIAN;
+	map_attrs |= DSP_MAPPHYSICALADDR;
+	map_attrs |= DSP_MAPELEMSIZE32;
+	map_attrs |= DSP_MAPDONOTLOCK;
 
-	pa_curr += ul_pad_size + num_bytes;
-	va_curr += ul_pad_size + num_bytes;
-	gpp_va_curr += ul_pad_size + num_bytes;
+	while (num_bytes) {
+		/*
+		 * To find the max. page size with which both PA & VA are
+		 * aligned.
+		 */
+		all_bits = pa_curr | va_curr;
+		dev_dbg(bridge, "all_bits %x, pa_curr %x, va_curr %x, "
+			"num_bytes %x\n", all_bits, pa_curr, va_curr,
+			num_bytes);
+		for (i = 0; i < 4; i++) {
+			if ((num_bytes >= page_size[i]) && ((all_bits &
+							     (page_size[i] -
+							      1)) == 0)) {
+				status =
+				    hio_mgr->intf_fxns->
+				    pfn_brd_mem_map(hio_mgr->hbridge_context,
+						    pa_curr, va_curr,
+						    page_size[i], map_attrs,
+						    NULL);
+				if (status)
+					goto func_end;
+				pa_curr += page_size[i];
+				va_curr += page_size[i];
+				gpp_va_curr += page_size[i];
+				num_bytes -= page_size[i];
+				/*
+				 * Don't try smaller sizes. Hopefully we have
+				 * reached an address aligned to a bigger page
+				 * size.
+				 */
+				break;
+			}
+		}
+	}
+	pa_curr += ul_pad_size;
+	va_curr += ul_pad_size;
+	gpp_va_curr += ul_pad_size;
 
 	/* Configure the TLB entries for the next cacheable segment */
 	num_bytes = ul_seg_size;
@@ -541,6 +566,22 @@ int bridge_io_on_loaded(struct io_mgr *hio_mgr)
 					ae_proc[ndx].ul_dsp_va *
 					hio_mgr->word_size, page_size[i]);
 				ndx++;
+			} else {
+				status =
+				    hio_mgr->intf_fxns->
+				    pfn_brd_mem_map(hio_mgr->hbridge_context,
+						    pa_curr, va_curr,
+						    page_size[i], map_attrs,
+						    NULL);
+				dev_dbg(bridge,
+					"shm MMU PTE entry PA %x"
+					" VA %x DSP_VA %x Size %x\n",
+					ae_proc[ndx].ul_gpp_pa,
+					ae_proc[ndx].ul_gpp_va,
+					ae_proc[ndx].ul_dsp_va *
+					hio_mgr->word_size, page_size[i]);
+				if (status)
+					goto func_end;
 			}
 			pa_curr += page_size[i];
 			va_curr += page_size[i];
@@ -593,29 +634,37 @@ int bridge_io_on_loaded(struct io_mgr *hio_mgr)
 					"DSP_VA 0x%x\n", ae_proc[ndx].ul_gpp_pa,
 					ae_proc[ndx].ul_dsp_va);
 				ndx++;
+			} else {
+				status = hio_mgr->intf_fxns->pfn_brd_mem_map
+				    (hio_mgr->hbridge_context,
+				     hio_mgr->ext_proc_info.ty_tlb[i].
+				     ul_gpp_phys,
+				     hio_mgr->ext_proc_info.ty_tlb[i].
+				     ul_dsp_virt, 0x100000, map_attrs,
+				     NULL);
 			}
 		}
 		if (status)
 			goto func_end;
 	}
 
+	map_attrs = 0x00000000;
+	map_attrs = DSP_MAPLITTLEENDIAN;
+	map_attrs |= DSP_MAPPHYSICALADDR;
+	map_attrs |= DSP_MAPELEMSIZE32;
+	map_attrs |= DSP_MAPDONOTLOCK;
+
 	/* Map the L4 peripherals */
 	i = 0;
 	while (l4_peripheral_table[i].phys_addr) {
-		status = iommu_kmap(mmu, l4_peripheral_table[i].
-			dsp_virt_addr, l4_peripheral_table[i].phys_addr,
-			PAGE_SIZE, IOVMF_ENDIAN_LITTLE | IOVMF_ELSZ_32);
-		if (IS_ERR_VALUE(status))
-			break;
+		status = hio_mgr->intf_fxns->pfn_brd_mem_map
+		    (hio_mgr->hbridge_context, l4_peripheral_table[i].phys_addr,
+		     l4_peripheral_table[i].dsp_virt_addr, HW_PAGE_SIZE4KB,
+		     map_attrs, NULL);
+		if (status)
+			goto func_end;
 		i++;
 	}
-	if (IS_ERR_VALUE(status)) {
-		while (i--)
-			iommu_kunmap(mmu, l4_peripheral_table[i].
-						dsp_virt_addr);
-		goto func_end;
-	}
-	status = 0;
 
 	for (i = ndx; i < BRDIOCTL_NUMOFMMUTLB; i++) {
 		ae_proc[i].ul_dsp_va = 0;

drivers/staging/tidspbridge/core/tiomap3430.c

@@ -114,7 +114,7 @@ static int bridge_brd_mem_map(struct bridge_dev_context *dev_ctxt,
 				  u32 ul_num_bytes, u32 ul_map_attr,
 				  struct page **mapped_pages);
 static int bridge_brd_mem_un_map(struct bridge_dev_context *dev_ctxt,
-				     u32 da);
+				     u32 virt_addr, u32 ul_num_bytes);
 static int bridge_dev_create(struct bridge_dev_context
 					**dev_cntxt,
 					struct dev_object *hdev_obj,
@@ -122,8 +122,6 @@ static int bridge_dev_create(struct bridge_dev_context
 static int bridge_dev_ctrl(struct bridge_dev_context *dev_context,
 				  u32 dw_cmd, void *pargs);
 static int bridge_dev_destroy(struct bridge_dev_context *dev_ctxt);
-static int get_io_pages(struct mm_struct *mm, u32 uva, unsigned pages,
-						struct page **usr_pgs);
 static u32 user_va2_pa(struct mm_struct *mm, u32 address);
 static int pte_update(struct bridge_dev_context *dev_ctxt, u32 pa,
 			     u32 va, u32 size,
@@ -373,7 +371,6 @@ static int bridge_brd_start(struct bridge_dev_context *dev_ctxt,
 {
 	int status = 0;
 	struct bridge_dev_context *dev_context = dev_ctxt;
-	struct iommu *mmu;
 	u32 dw_sync_addr = 0;
 	u32 ul_shm_base;	/* Gpp Phys SM base addr(byte) */
 	u32 ul_shm_base_virt;	/* Dsp Virt SM base addr */
@@ -393,7 +390,6 @@ static int bridge_brd_start(struct bridge_dev_context *dev_ctxt,
 	struct omap_dsp_platform_data *pdata =
 		omap_dspbridge_dev->dev.platform_data;
 
-	mmu = dev_context->dsp_mmu;
 	/* The device context contains all the mmu setup info from when the
 	 * last dsp base image was loaded. The first entry is always
 	 * SHMMEM base. */
@@ -444,10 +440,29 @@ static int bridge_brd_start(struct bridge_dev_context *dev_ctxt,
 		}
 	}
 	if (!status) {
+		/* Reset and Unreset the RST2, so that BOOTADDR is copied to
+		 * IVA2 SYSC register */
+		(*pdata->dsp_prm_rmw_bits)(OMAP3430_RST2_IVA2_MASK,
+			OMAP3430_RST2_IVA2_MASK, OMAP3430_IVA2_MOD, OMAP2_RM_RSTCTRL);
+		udelay(100);
+		(*pdata->dsp_prm_rmw_bits)(OMAP3430_RST2_IVA2_MASK, 0,
+					OMAP3430_IVA2_MOD, OMAP2_RM_RSTCTRL);
+		udelay(100);
+
+		/* Disbale the DSP MMU */
+		hw_mmu_disable(resources->dw_dmmu_base);
+		/* Disable TWL */
+		hw_mmu_twl_disable(resources->dw_dmmu_base);
+
 		/* Only make TLB entry if both addresses are non-zero */
 		for (entry_ndx = 0; entry_ndx < BRDIOCTL_NUMOFMMUTLB;
 		     entry_ndx++) {
 			struct bridge_ioctl_extproc *e = &dev_context->atlb_entry[entry_ndx];
+			struct hw_mmu_map_attrs_t map_attrs = {
+				.endianism = e->endianism,
+				.element_size = e->elem_size,
+				.mixed_size = e->mixed_mode,
+			};
 
 			if (!e->ul_gpp_pa || !e->ul_dsp_va)
 				continue;
@@ -459,8 +474,13 @@ static int bridge_brd_start(struct bridge_dev_context *dev_ctxt,
 					e->ul_dsp_va,
 					e->ul_size);
 
-			iommu_kmap(mmu, e->ul_dsp_va, e->ul_gpp_pa, e->ul_size,
-				IOVMF_ENDIAN_LITTLE | IOVMF_ELSZ_32);
+			hw_mmu_tlb_add(dev_context->dw_dsp_mmu_base,
+					e->ul_gpp_pa,
+					e->ul_dsp_va,
+					e->ul_size,
+					itmp_entry_ndx,
+					&map_attrs, 1, 1);
+
 			itmp_entry_ndx++;
 		}
 	}
@@ -468,13 +488,29 @@ static int bridge_brd_start(struct bridge_dev_context *dev_ctxt,
 	/* Lock the above TLB entries and get the BIOS and load monitor timer
 	 * information */
 	if (!status) {
+		hw_mmu_num_locked_set(resources->dw_dmmu_base, itmp_entry_ndx);
+		hw_mmu_victim_num_set(resources->dw_dmmu_base, itmp_entry_ndx);
+		hw_mmu_ttb_set(resources->dw_dmmu_base,
+			       dev_context->pt_attrs->l1_base_pa);
+		hw_mmu_twl_enable(resources->dw_dmmu_base);
+		/* Enable the SmartIdle and AutoIdle bit for MMU_SYSCONFIG */
+
+		temp = __raw_readl((resources->dw_dmmu_base) + 0x10);
+		temp = (temp & 0xFFFFFFEF) | 0x11;
+		__raw_writel(temp, (resources->dw_dmmu_base) + 0x10);
+
+		/* Let the DSP MMU run */
+		hw_mmu_enable(resources->dw_dmmu_base);
+
 		/* Enable the BIOS clock */
 		(void)dev_get_symbol(dev_context->hdev_obj,
 				     BRIDGEINIT_BIOSGPTIMER, &ul_bios_gp_timer);
 		(void)dev_get_symbol(dev_context->hdev_obj,
 				     BRIDGEINIT_LOADMON_GPTIMER,
 				     &ul_load_monitor_timer);
+	}
 
+	if (!status) {
 		if (ul_load_monitor_timer != 0xFFFF) {
 			clk_cmd = (BPWR_ENABLE_CLOCK << MBX_PM_CLK_CMDSHIFT) |
 			    ul_load_monitor_timer;
@@ -483,7 +519,9 @@ static int bridge_brd_start(struct bridge_dev_context *dev_ctxt,
 			dev_dbg(bridge, "Not able to get the symbol for Load "
 				"Monitor Timer\n");
 		}
+	}
 
+	if (!status) {
 		if (ul_bios_gp_timer != 0xFFFF) {
 			clk_cmd = (BPWR_ENABLE_CLOCK << MBX_PM_CLK_CMDSHIFT) |
 			    ul_bios_gp_timer;
@@ -492,7 +530,9 @@ static int bridge_brd_start(struct bridge_dev_context *dev_ctxt,
 			dev_dbg(bridge,
 				"Not able to get the symbol for BIOS Timer\n");
 		}
+	}
 
+	if (!status) {
 		/* Set the DSP clock rate */
 		(void)dev_get_symbol(dev_context->hdev_obj,
 				     "_BRIDGEINIT_DSP_FREQ", &ul_dsp_clk_addr);
@@ -545,6 +585,9 @@ static int bridge_brd_start(struct bridge_dev_context *dev_ctxt,
 
 		/* Let DSP go */
 		dev_dbg(bridge, "%s Unreset\n", __func__);
+		/* Enable DSP MMU Interrupts */
+		hw_mmu_event_enable(resources->dw_dmmu_base,
+				    HW_MMU_ALL_INTERRUPTS);
 		/* release the RST1, DSP starts executing now .. */
 		(*pdata->dsp_prm_rmw_bits)(OMAP3430_RST1_IVA2_MASK, 0,
 					OMAP3430_IVA2_MOD, OMAP2_RM_RSTCTRL);
@@ -645,8 +688,6 @@ static int bridge_brd_stop(struct bridge_dev_context *dev_ctxt)
 		omap_mbox_put(dev_context->mbox);
 		dev_context->mbox = NULL;
 	}
-	if (dev_context->dsp_mmu)
-		dev_context->dsp_mmu = (iommu_put(dev_context->dsp_mmu), NULL);
 	/* Reset IVA2 clocks*/
 	(*pdata->dsp_prm_write)(OMAP3430_RST1_IVA2_MASK | OMAP3430_RST2_IVA2_MASK |
 			OMAP3430_RST3_IVA2_MASK, OMAP3430_IVA2_MOD, OMAP2_RM_RSTCTRL);
@@ -1095,81 +1136,217 @@ static int bridge_brd_mem_write(struct bridge_dev_context *dev_ctxt,
  *
  *  TODO: Disable MMU while updating the page tables (but that'll stall DSP)
  */
-static int bridge_brd_mem_map(struct bridge_dev_context *dev_ctx,
-			u32 uva, u32 da, u32 size, u32 attr,
-			struct page **usr_pgs)
-
+static int bridge_brd_mem_map(struct bridge_dev_context *dev_ctxt,
+				  u32 ul_mpu_addr, u32 virt_addr,
+				  u32 ul_num_bytes, u32 ul_map_attr,
+				  struct page **mapped_pages)
 {
-	int res, w;
-	unsigned pages, i;
-	struct iommu *mmu = dev_ctx->dsp_mmu;
+	u32 attrs;
+	int status = 0;
+	struct bridge_dev_context *dev_context = dev_ctxt;
+	struct hw_mmu_map_attrs_t hw_attrs;
 	struct vm_area_struct *vma;
 	struct mm_struct *mm = current->mm;
-	struct sg_table *sgt;
-	struct scatterlist *sg;
-
-	if (!size || !usr_pgs)
+	u32 write = 0;
+	u32 num_usr_pgs = 0;
+	struct page *mapped_page, *pg;
+	s32 pg_num;
+	u32 va = virt_addr;
+	struct task_struct *curr_task = current;
+	u32 pg_i = 0;
+	u32 mpu_addr, pa;
+
+	dev_dbg(bridge,
+		"%s hDevCtxt %p, pa %x, va %x, size %x, ul_map_attr %x\n",
+		__func__, dev_ctxt, ul_mpu_addr, virt_addr, ul_num_bytes,
+		ul_map_attr);
+	if (ul_num_bytes == 0)
 		return -EINVAL;
 
-	pages = size / PG_SIZE4K;
+	if (ul_map_attr & DSP_MAP_DIR_MASK) {
+		attrs = ul_map_attr;
+	} else {
+		/* Assign default attributes */
+		attrs = ul_map_attr | (DSP_MAPVIRTUALADDR | DSP_MAPELEMSIZE16);
+	}
+	/* Take mapping properties */
+	if (attrs & DSP_MAPBIGENDIAN)
+		hw_attrs.endianism = HW_BIG_ENDIAN;
+	else
+		hw_attrs.endianism = HW_LITTLE_ENDIAN;
+
+	hw_attrs.mixed_size = (enum hw_mmu_mixed_size_t)
+	    ((attrs & DSP_MAPMIXEDELEMSIZE) >> 2);
+	/* Ignore element_size if mixed_size is enabled */
+	if (hw_attrs.mixed_size == 0) {
+		if (attrs & DSP_MAPELEMSIZE8) {
+			/* Size is 8 bit */
+			hw_attrs.element_size = HW_ELEM_SIZE8BIT;
+		} else if (attrs & DSP_MAPELEMSIZE16) {
+			/* Size is 16 bit */
+			hw_attrs.element_size = HW_ELEM_SIZE16BIT;
+		} else if (attrs & DSP_MAPELEMSIZE32) {
+			/* Size is 32 bit */
+			hw_attrs.element_size = HW_ELEM_SIZE32BIT;
+		} else if (attrs & DSP_MAPELEMSIZE64) {
+			/* Size is 64 bit */
+			hw_attrs.element_size = HW_ELEM_SIZE64BIT;
+		} else {
+			/*
+			 * Mixedsize isn't enabled, so size can't be
+			 * zero here
+			 */
+			return -EINVAL;
+		}
+	}
+	if (attrs & DSP_MAPDONOTLOCK)
+		hw_attrs.donotlockmpupage = 1;
+	else
+		hw_attrs.donotlockmpupage = 0;
 
+	if (attrs & DSP_MAPVMALLOCADDR) {
+		return mem_map_vmalloc(dev_ctxt, ul_mpu_addr, virt_addr,
+				       ul_num_bytes, &hw_attrs);
+	}
+	/*
+	 * Do OS-specific user-va to pa translation.
+	 * Combine physically contiguous regions to reduce TLBs.
+	 * Pass the translated pa to pte_update.
+	 */
+	if ((attrs & DSP_MAPPHYSICALADDR)) {
+		status = pte_update(dev_context, ul_mpu_addr, virt_addr,
+				    ul_num_bytes, &hw_attrs);
+		goto func_cont;
+	}
+
+	/*
+	 * Important Note: ul_mpu_addr is mapped from user application process
+	 * to current process - it must lie completely within the current
+	 * virtual memory address space in order to be of use to us here!
+	 */
 	down_read(&mm->mmap_sem);
-	vma = find_vma(mm, uva);
-	while (vma && (uva + size > vma->vm_end))
-		vma = find_vma(mm, vma->vm_end + 1);
+	vma = find_vma(mm, ul_mpu_addr);
+	if (vma)
+		dev_dbg(bridge,
+			"VMAfor UserBuf: ul_mpu_addr=%x, ul_num_bytes=%x, "
+			"vm_start=%lx, vm_end=%lx, vm_flags=%lx\n", ul_mpu_addr,
+			ul_num_bytes, vma->vm_start, vma->vm_end,
+			vma->vm_flags);
 
+	/*
+	 * It is observed that under some circumstances, the user buffer is
+	 * spread across several VMAs. So loop through and check if the entire
+	 * user buffer is covered
+	 */
+	while ((vma) && (ul_mpu_addr + ul_num_bytes > vma->vm_end)) {
+		/* jump to the next VMA region */
+		vma = find_vma(mm, vma->vm_end + 1);
+		dev_dbg(bridge,
+			"VMA for UserBuf ul_mpu_addr=%x ul_num_bytes=%x, "
+			"vm_start=%lx, vm_end=%lx, vm_flags=%lx\n", ul_mpu_addr,
+			ul_num_bytes, vma->vm_start, vma->vm_end,
+			vma->vm_flags);
+	}
 	if (!vma) {
 		pr_err("%s: Failed to get VMA region for 0x%x (%d)\n",
-						__func__, uva, size);
+		       __func__, ul_mpu_addr, ul_num_bytes);
+		status = -EINVAL;
 		up_read(&mm->mmap_sem);
-		return -EINVAL;
+		goto func_cont;
 	}
-	if (vma->vm_flags & (VM_WRITE | VM_MAYWRITE))
-		w = 1;
-
-	if (vma->vm_flags & VM_IO)
-		i = get_io_pages(mm, uva, pages, usr_pgs);
-	else
-		i = get_user_pages(current, mm, uva, pages, w, 1,
-							usr_pgs, NULL);
-	up_read(&mm->mmap_sem);
 
-	if (i < 0)
-		return i;
+	if (vma->vm_flags & VM_IO) {
+		num_usr_pgs = ul_num_bytes / PG_SIZE4K;
+		mpu_addr = ul_mpu_addr;
+
+		/* Get the physical addresses for user buffer */
+		for (pg_i = 0; pg_i < num_usr_pgs; pg_i++) {
+			pa = user_va2_pa(mm, mpu_addr);
+			if (!pa) {
+				status = -EPERM;
+				pr_err("DSPBRIDGE: VM_IO mapping physical"
+				       "address is invalid\n");
+				break;
+			}
+			if (pfn_valid(__phys_to_pfn(pa))) {
+				pg = PHYS_TO_PAGE(pa);
+				get_page(pg);
+				if (page_count(pg) < 1) {
+					pr_err("Bad page in VM_IO buffer\n");
+					bad_page_dump(pa, pg);
+				}
+			}
+			status = pte_set(dev_context->pt_attrs, pa,
+					 va, HW_PAGE_SIZE4KB, &hw_attrs);
+			if (status)
+				break;
 
-	if (i < pages) {
-		res = -EFAULT;
-		goto err_pages;
+			va += HW_PAGE_SIZE4KB;
+			mpu_addr += HW_PAGE_SIZE4KB;
+			pa += HW_PAGE_SIZE4KB;
+		}
+	} else {
+		num_usr_pgs = ul_num_bytes / PG_SIZE4K;
+		if (vma->vm_flags & (VM_WRITE | VM_MAYWRITE))
+			write = 1;
+
+		for (pg_i = 0; pg_i < num_usr_pgs; pg_i++) {
+			pg_num = get_user_pages(curr_task, mm, ul_mpu_addr, 1,
+						write, 1, &mapped_page, NULL);
+			if (pg_num > 0) {
+				if (page_count(mapped_page) < 1) {
+					pr_err("Bad page count after doing"
+					       "get_user_pages on"
+					       "user buffer\n");
+					bad_page_dump(page_to_phys(mapped_page),
+						      mapped_page);
+				}
+				status = pte_set(dev_context->pt_attrs,
+						 page_to_phys(mapped_page), va,
+						 HW_PAGE_SIZE4KB, &hw_attrs);
+				if (status)
+					break;
+
+				if (mapped_pages)
+					mapped_pages[pg_i] = mapped_page;
+
+				va += HW_PAGE_SIZE4KB;
+				ul_mpu_addr += HW_PAGE_SIZE4KB;
+			} else {
+				pr_err("DSPBRIDGE: get_user_pages FAILED,"
+				       "MPU addr = 0x%x,"
+				       "vma->vm_flags = 0x%lx,"
+				       "get_user_pages Err"
+				       "Value = %d, Buffer"
+				       "size=0x%x\n", ul_mpu_addr,
+				       vma->vm_flags, pg_num, ul_num_bytes);
+				status = -EPERM;
+				break;
+			}
+		}
 	}
-
-	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
-	if (!sgt) {
-		res = -ENOMEM;
-		goto err_pages;
+	up_read(&mm->mmap_sem);
+func_cont:
+	if (status) {
+		/*
+		 * Roll out the mapped pages incase it failed in middle of
+		 * mapping
+		 */
+		if (pg_i) {
+			bridge_brd_mem_un_map(dev_context, virt_addr,
+					   (pg_i * PG_SIZE4K));
+		}
+		status = -EPERM;
 	}
-
-	res = sg_alloc_table(sgt, pages, GFP_KERNEL);
-
-	if (res < 0)
-		goto err_sg;
-
-	for_each_sg(sgt->sgl, sg, sgt->nents, i)
-		sg_set_page(sg, usr_pgs[i], PAGE_SIZE, 0);
-
-	da = iommu_vmap(mmu, da, sgt, IOVMF_ENDIAN_LITTLE | IOVMF_ELSZ_32);
-
-	if (!IS_ERR_VALUE(da))
-		return 0;
-	res = (int)da;
-
-	sg_free_table(sgt);
-err_sg:
-	kfree(sgt);
-	i = pages;
-err_pages:
-	while (i--)
-		put_page(usr_pgs[i]);
-	return res;
+	/*
+	 * In any case, flush the TLB
+	 * This is called from here instead from pte_update to avoid unnecessary
+	 * repetition while mapping non-contiguous physical regions of a virtual
+	 * region
+	 */
+	flush_all(dev_context);
+	dev_dbg(bridge, "%s status %x\n", __func__, status);
+	return status;
 }
 
 /*
@@ -1180,43 +1357,194 @@ err_pages:
  *      So, instead of looking up the PTE address for every 4K block,
  *      we clear consecutive PTEs until we unmap all the bytes
  */
-static int bridge_brd_mem_un_map(struct bridge_dev_context *dev_ctx, u32 da)
+static int bridge_brd_mem_un_map(struct bridge_dev_context *dev_ctxt,
+				     u32 virt_addr, u32 ul_num_bytes)
 {
-	unsigned i;
-	struct sg_table *sgt;
-	struct scatterlist *sg;
+	u32 l1_base_va;
+	u32 l2_base_va;
+	u32 l2_base_pa;
+	u32 l2_page_num;
+	u32 pte_val;
+	u32 pte_size;
+	u32 pte_count;
+	u32 pte_addr_l1;
+	u32 pte_addr_l2 = 0;
+	u32 rem_bytes;
+	u32 rem_bytes_l2;
+	u32 va_curr;
+	struct page *pg = NULL;
+	int status = 0;
+	struct bridge_dev_context *dev_context = dev_ctxt;
+	struct pg_table_attrs *pt = dev_context->pt_attrs;
+	u32 temp;
+	u32 paddr;
+	u32 numof4k_pages = 0;
 
-	sgt = iommu_vunmap(dev_ctx->dsp_mmu, da);
-	if (!sgt)
-		return -EFAULT;
+	va_curr = virt_addr;
+	rem_bytes = ul_num_bytes;
+	rem_bytes_l2 = 0;
+	l1_base_va = pt->l1_base_va;
+	pte_addr_l1 = hw_mmu_pte_addr_l1(l1_base_va, va_curr);
+	dev_dbg(bridge, "%s dev_ctxt %p, va %x, NumBytes %x l1_base_va %x, "
+		"pte_addr_l1 %x\n", __func__, dev_ctxt, virt_addr,
+		ul_num_bytes, l1_base_va, pte_addr_l1);
+
+	while (rem_bytes && !status) {
+		u32 va_curr_orig = va_curr;
+		/* Find whether the L1 PTE points to a valid L2 PT */
+		pte_addr_l1 = hw_mmu_pte_addr_l1(l1_base_va, va_curr);
+		pte_val = *(u32 *) pte_addr_l1;
+		pte_size = hw_mmu_pte_size_l1(pte_val);
 
-	for_each_sg(sgt->sgl, sg, sgt->nents, i)
-		put_page(sg_page(sg));
-	sg_free_table(sgt);
-	kfree(sgt);
+		if (pte_size != HW_MMU_COARSE_PAGE_SIZE)
+			goto skip_coarse_page;
 
-	return 0;
-}
+		/*
+		 * Get the L2 PA from the L1 PTE, and find
+		 * corresponding L2 VA
+		 */
+		l2_base_pa = hw_mmu_pte_coarse_l1(pte_val);
+		l2_base_va = l2_base_pa - pt->l2_base_pa + pt->l2_base_va;
+		l2_page_num =
+		    (l2_base_pa - pt->l2_base_pa) / HW_MMU_COARSE_PAGE_SIZE;
+		/*
+		 * Find the L2 PTE address from which we will start
+		 * clearing, the number of PTEs to be cleared on this
+		 * page, and the size of VA space that needs to be
+		 * cleared on this L2 page
+		 */
+		pte_addr_l2 = hw_mmu_pte_addr_l2(l2_base_va, va_curr);
+		pte_count = pte_addr_l2 & (HW_MMU_COARSE_PAGE_SIZE - 1);
+		pte_count = (HW_MMU_COARSE_PAGE_SIZE - pte_count) / sizeof(u32);
+		if (rem_bytes < (pte_count * PG_SIZE4K))
+			pte_count = rem_bytes / PG_SIZE4K;
+		rem_bytes_l2 = pte_count * PG_SIZE4K;
 
+		/*
+		 * Unmap the VA space on this L2 PT. A quicker way
+		 * would be to clear pte_count entries starting from
+		 * pte_addr_l2. However, below code checks that we don't
+		 * clear invalid entries or less than 64KB for a 64KB
+		 * entry. Similar checking is done for L1 PTEs too
+		 * below
+		 */
+		while (rem_bytes_l2 && !status) {
+			pte_val = *(u32 *) pte_addr_l2;
+			pte_size = hw_mmu_pte_size_l2(pte_val);
+			/* va_curr aligned to pte_size? */
+			if (pte_size == 0 || rem_bytes_l2 < pte_size ||
+			    va_curr & (pte_size - 1)) {
+				status = -EPERM;
+				break;
+			}
 
-static int get_io_pages(struct mm_struct *mm, u32 uva, unsigned pages,
-						struct page **usr_pgs)
-{
-	u32 pa;
-	int i;
-	struct page *pg;
+			/* Collect Physical addresses from VA */
+			paddr = (pte_val & ~(pte_size - 1));
+			if (pte_size == HW_PAGE_SIZE64KB)
+				numof4k_pages = 16;
+			else
+				numof4k_pages = 1;
+			temp = 0;
+			while (temp++ < numof4k_pages) {
+				if (!pfn_valid(__phys_to_pfn(paddr))) {
+					paddr += HW_PAGE_SIZE4KB;
+					continue;
+				}
+				pg = PHYS_TO_PAGE(paddr);
+				if (page_count(pg) < 1) {
+					pr_info("DSPBRIDGE: UNMAP function: "
+						"COUNT 0 FOR PA 0x%x, size = "
+						"0x%x\n", paddr, ul_num_bytes);
+					bad_page_dump(paddr, pg);
+				} else {
+					set_page_dirty(pg);
+					page_cache_release(pg);
+				}
+				paddr += HW_PAGE_SIZE4KB;
+			}
+			if (hw_mmu_pte_clear(pte_addr_l2, va_curr, pte_size)) {
+				status = -EPERM;
+				goto EXIT_LOOP;
+			}
 
-	for (i = 0; i < pages; i++) {
-		pa = user_va2_pa(mm, uva);
+			status = 0;
+			rem_bytes_l2 -= pte_size;
+			va_curr += pte_size;
+			pte_addr_l2 += (pte_size >> 12) * sizeof(u32);
+		}
+		spin_lock(&pt->pg_lock);
+		if (rem_bytes_l2 == 0) {
+			pt->pg_info[l2_page_num].num_entries -= pte_count;
+			if (pt->pg_info[l2_page_num].num_entries == 0) {
+				/*
+				 * Clear the L1 PTE pointing to the L2 PT
+				 */
+				if (!hw_mmu_pte_clear(l1_base_va, va_curr_orig,
+						     HW_MMU_COARSE_PAGE_SIZE))
+					status = 0;
+				else {
+					status = -EPERM;
+					spin_unlock(&pt->pg_lock);
+					goto EXIT_LOOP;
+				}
+			}
+			rem_bytes -= pte_count * PG_SIZE4K;
+		} else
+			status = -EPERM;
 
-		if (!pfn_valid(__phys_to_pfn(pa)))
+		spin_unlock(&pt->pg_lock);
+		continue;
+skip_coarse_page:
+		/* va_curr aligned to pte_size? */
+		/* pte_size = 1 MB or 16 MB */
+		if (pte_size == 0 || rem_bytes < pte_size ||
+		    va_curr & (pte_size - 1)) {
+			status = -EPERM;
 			break;
+		}
 
-		pg = PHYS_TO_PAGE(pa);
-		usr_pgs[i] = pg;
-		get_page(pg);
+		if (pte_size == HW_PAGE_SIZE1MB)
+			numof4k_pages = 256;
+		else
+			numof4k_pages = 4096;
+		temp = 0;
+		/* Collect Physical addresses from VA */
+		paddr = (pte_val & ~(pte_size - 1));
+		while (temp++ < numof4k_pages) {
+			if (pfn_valid(__phys_to_pfn(paddr))) {
+				pg = PHYS_TO_PAGE(paddr);
+				if (page_count(pg) < 1) {
+					pr_info("DSPBRIDGE: UNMAP function: "
+						"COUNT 0 FOR PA 0x%x, size = "
+						"0x%x\n", paddr, ul_num_bytes);
+					bad_page_dump(paddr, pg);
+				} else {
+					set_page_dirty(pg);
+					page_cache_release(pg);
+				}
+			}
+			paddr += HW_PAGE_SIZE4KB;
+		}
+		if (!hw_mmu_pte_clear(l1_base_va, va_curr, pte_size)) {
+			status = 0;
+			rem_bytes -= pte_size;
+			va_curr += pte_size;
+		} else {
+			status = -EPERM;
+			goto EXIT_LOOP;
+		}
 	}
-	return i;
+	/*
+	 * It is better to flush the TLB here, so that any stale old entries
+	 * get flushed
+	 */
+EXIT_LOOP:
+	flush_all(dev_context);
+	dev_dbg(bridge,
+		"%s: va_curr %x, pte_addr_l1 %x pte_addr_l2 %x rem_bytes %x,"
+		" rem_bytes_l2 %x status %x\n", __func__, va_curr, pte_addr_l1,
+		pte_addr_l2, rem_bytes, rem_bytes_l2, status);
+	return status;
 }
 
 /*

drivers/staging/tidspbridge/core/ue_deh.c

@@ -115,6 +115,12 @@ int bridge_deh_create(struct deh_mgr **ret_deh,
 	/* Fill in context structure */
 	deh->hbridge_context = hbridge_context;
 
+	/* Install ISR function for DSP MMU fault */
+	status = request_irq(INT_DSP_MMU_IRQ, mmu_fault_isr, 0,
+			"DspBridge\tiommu fault", deh);
+	if (status < 0)
+		goto err;
+
 	*ret_deh = deh;
 	return 0;
 
@@ -134,6 +140,8 @@ int bridge_deh_destroy(struct deh_mgr *deh)
 		ntfy_delete(deh->ntfy_obj);
 		kfree(deh->ntfy_obj);
 	}
+	/* Disable DSP MMU fault */
+	free_irq(INT_DSP_MMU_IRQ, deh);
 
 	/* Free DPC object */
 	tasklet_kill(&deh->dpc_tasklet);

drivers/staging/tidspbridge/include/dspbridge/dspdefs.h

@@ -201,7 +201,7 @@ typedef int(*fxn_brd_memmap) (struct bridge_dev_context
  */
 typedef int(*fxn_brd_memunmap) (struct bridge_dev_context
 				       * dev_ctxt,
-				       u32 da);
+				       u32 virt_addr, u32 ul_num_bytes);
 
 /*
  *  ======== bridge_brd_stop ========

drivers/staging/tidspbridge/rmgr/proc.c

@@ -1723,7 +1723,7 @@ int proc_un_map(void *hprocessor, void *map_addr,
 	/* Remove mapping from the page tables. */
 	if (!status) {
 		status = (*p_proc_object->intf_fxns->pfn_brd_mem_un_map)
-		    (p_proc_object->hbridge_context, va_align);
+		    (p_proc_object->hbridge_context, va_align, size_align);
 	}
 
 	mutex_unlock(&proc_lock);