Hi Greentime and Vicent,
can you take a look at the (untested) patch below? It converts nds32
to use the generic remapping DMA allocator, which is also used by
arm64 and csky.
Replace the code that sets up uncached PTEs with the generic vmap based
remapping code. It also provides an atomic pool for allocations from
non-blocking context, which we not properly supported by the existing
nds32 code.
Signed-off-by: Christoph Hellwig <[email protected]>
---
arch/nds32/Kconfig | 2 +
arch/nds32/kernel/dma.c | 325 ++--------------------------------------
2 files changed, 13 insertions(+), 314 deletions(-)
diff --git a/arch/nds32/Kconfig b/arch/nds32/Kconfig
index 3299e287a477..643ea6b4bfa2 100644
--- a/arch/nds32/Kconfig
+++ b/arch/nds32/Kconfig
@@ -7,12 +7,14 @@
config NDS32
def_bool y
select ARCH_32BIT_OFF_T
+ select ARCH_HAS_DMA_PREP_COHERENT
select ARCH_HAS_SYNC_DMA_FOR_CPU
select ARCH_HAS_SYNC_DMA_FOR_DEVICE
select ARCH_WANT_FRAME_POINTERS if FTRACE
select CLKSRC_MMIO
select CLONE_BACKWARDS
select COMMON_CLK
+ select DMA_DIRECT_REMAP
select GENERIC_ATOMIC64
select GENERIC_CPU_DEVICES
select GENERIC_CLOCKEVENTS
diff --git a/arch/nds32/kernel/dma.c b/arch/nds32/kernel/dma.c
index d0dbd4fe9645..490e3720d694 100644
--- a/arch/nds32/kernel/dma.c
+++ b/arch/nds32/kernel/dma.c
@@ -3,327 +3,13 @@
#include <linux/types.h>
#include <linux/mm.h>
-#include <linux/string.h>
#include <linux/dma-noncoherent.h>
-#include <linux/io.h>
#include <linux/cache.h>
#include <linux/highmem.h>
-#include <linux/slab.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>
#include <asm/proc-fns.h>
-/*
- * This is the page table (2MB) covering uncached, DMA consistent allocations
- */
-static pte_t *consistent_pte;
-static DEFINE_RAW_SPINLOCK(consistent_lock);
-
-/*
- * VM region handling support.
- *
- * This should become something generic, handling VM region allocations for
- * vmalloc and similar (ioremap, module space, etc).
- *
- * I envisage vmalloc()'s supporting vm_struct becoming:
- *
- * struct vm_struct {
- * struct vm_region region;
- * unsigned long flags;
- * struct page **pages;
- * unsigned int nr_pages;
- * unsigned long phys_addr;
- * };
- *
- * get_vm_area() would then call vm_region_alloc with an appropriate
- * struct vm_region head (eg):
- *
- * struct vm_region vmalloc_head = {
- * .vm_list = LIST_HEAD_INIT(vmalloc_head.vm_list),
- * .vm_start = VMALLOC_START,
- * .vm_end = VMALLOC_END,
- * };
- *
- * However, vmalloc_head.vm_start is variable (typically, it is dependent on
- * the amount of RAM found at boot time.) I would imagine that get_vm_area()
- * would have to initialise this each time prior to calling vm_region_alloc().
- */
-struct arch_vm_region {
- struct list_head vm_list;
- unsigned long vm_start;
- unsigned long vm_end;
- struct page *vm_pages;
-};
-
-static struct arch_vm_region consistent_head = {
- .vm_list = LIST_HEAD_INIT(consistent_head.vm_list),
- .vm_start = CONSISTENT_BASE,
- .vm_end = CONSISTENT_END,
-};
-
-static struct arch_vm_region *vm_region_alloc(struct arch_vm_region *head,
- size_t size, int gfp)
-{
- unsigned long addr = head->vm_start, end = head->vm_end - size;
- unsigned long flags;
- struct arch_vm_region *c, *new;
-
- new = kmalloc(sizeof(struct arch_vm_region), gfp);
- if (!new)
- goto out;
-
- raw_spin_lock_irqsave(&consistent_lock, flags);
-
- list_for_each_entry(c, &head->vm_list, vm_list) {
- if ((addr + size) < addr)
- goto nospc;
- if ((addr + size) <= c->vm_start)
- goto found;
- addr = c->vm_end;
- if (addr > end)
- goto nospc;
- }
-
-found:
- /*
- * Insert this entry _before_ the one we found.
- */
- list_add_tail(&new->vm_list, &c->vm_list);
- new->vm_start = addr;
- new->vm_end = addr + size;
-
- raw_spin_unlock_irqrestore(&consistent_lock, flags);
- return new;
-
-nospc:
- raw_spin_unlock_irqrestore(&consistent_lock, flags);
- kfree(new);
-out:
- return NULL;
-}
-
-static struct arch_vm_region *vm_region_find(struct arch_vm_region *head,
- unsigned long addr)
-{
- struct arch_vm_region *c;
-
- list_for_each_entry(c, &head->vm_list, vm_list) {
- if (c->vm_start == addr)
- goto out;
- }
- c = NULL;
-out:
- return c;
-}
-
-void *arch_dma_alloc(struct device *dev, size_t size, dma_addr_t *handle,
- gfp_t gfp, unsigned long attrs)
-{
- struct page *page;
- struct arch_vm_region *c;
- unsigned long order;
- u64 mask = ~0ULL, limit;
- pgprot_t prot = pgprot_noncached(PAGE_KERNEL);
-
- if (!consistent_pte) {
- pr_err("%s: not initialized\n", __func__);
- dump_stack();
- return NULL;
- }
-
- if (dev) {
- mask = dev->coherent_dma_mask;
-
- /*
- * Sanity check the DMA mask - it must be non-zero, and
- * must be able to be satisfied by a DMA allocation.
- */
- if (mask == 0) {
- dev_warn(dev, "coherent DMA mask is unset\n");
- goto no_page;
- }
-
- }
-
- /*
- * Sanity check the allocation size.
- */
- size = PAGE_ALIGN(size);
- limit = (mask + 1) & ~mask;
- if ((limit && size >= limit) ||
- size >= (CONSISTENT_END - CONSISTENT_BASE)) {
- pr_warn("coherent allocation too big "
- "(requested %#x mask %#llx)\n", size, mask);
- goto no_page;
- }
-
- order = get_order(size);
-
- if (mask != 0xffffffff)
- gfp |= GFP_DMA;
-
- page = alloc_pages(gfp, order);
- if (!page)
- goto no_page;
-
- /*
- * Invalidate any data that might be lurking in the
- * kernel direct-mapped region for device DMA.
- */
- {
- unsigned long kaddr = (unsigned long)page_address(page);
- memset(page_address(page), 0, size);
- cpu_dma_wbinval_range(kaddr, kaddr + size);
- }
-
- /*
- * Allocate a virtual address in the consistent mapping region.
- */
- c = vm_region_alloc(&consistent_head, size,
- gfp & ~(__GFP_DMA | __GFP_HIGHMEM));
- if (c) {
- pte_t *pte = consistent_pte + CONSISTENT_OFFSET(c->vm_start);
- struct page *end = page + (1 << order);
-
- c->vm_pages = page;
-
- /*
- * Set the "dma handle"
- */
- *handle = page_to_phys(page);
-
- do {
- BUG_ON(!pte_none(*pte));
-
- /*
- * x86 does not mark the pages reserved...
- */
- SetPageReserved(page);
- set_pte(pte, mk_pte(page, prot));
- page++;
- pte++;
- } while (size -= PAGE_SIZE);
-
- /*
- * Free the otherwise unused pages.
- */
- while (page < end) {
- __free_page(page);
- page++;
- }
-
- return (void *)c->vm_start;
- }
-
- if (page)
- __free_pages(page, order);
-no_page:
- *handle = ~0;
- return NULL;
-}
-
-void arch_dma_free(struct device *dev, size_t size, void *cpu_addr,
- dma_addr_t handle, unsigned long attrs)
-{
- struct arch_vm_region *c;
- unsigned long flags, addr;
- pte_t *ptep;
-
- size = PAGE_ALIGN(size);
-
- raw_spin_lock_irqsave(&consistent_lock, flags);
-
- c = vm_region_find(&consistent_head, (unsigned long)cpu_addr);
- if (!c)
- goto no_area;
-
- if ((c->vm_end - c->vm_start) != size) {
- pr_err("%s: freeing wrong coherent size (%ld != %d)\n",
- __func__, c->vm_end - c->vm_start, size);
- dump_stack();
- size = c->vm_end - c->vm_start;
- }
-
- ptep = consistent_pte + CONSISTENT_OFFSET(c->vm_start);
- addr = c->vm_start;
- do {
- pte_t pte = ptep_get_and_clear(&init_mm, addr, ptep);
- unsigned long pfn;
-
- ptep++;
- addr += PAGE_SIZE;
-
- if (!pte_none(pte) && pte_present(pte)) {
- pfn = pte_pfn(pte);
-
- if (pfn_valid(pfn)) {
- struct page *page = pfn_to_page(pfn);
-
- /*
- * x86 does not mark the pages reserved...
- */
- ClearPageReserved(page);
-
- __free_page(page);
- continue;
- }
- }
-
- pr_crit("%s: bad page in kernel page table\n", __func__);
- } while (size -= PAGE_SIZE);
-
- flush_tlb_kernel_range(c->vm_start, c->vm_end);
-
- list_del(&c->vm_list);
-
- raw_spin_unlock_irqrestore(&consistent_lock, flags);
-
- kfree(c);
- return;
-
-no_area:
- raw_spin_unlock_irqrestore(&consistent_lock, flags);
- pr_err("%s: trying to free invalid coherent area: %p\n",
- __func__, cpu_addr);
- dump_stack();
-}
-
-/*
- * Initialise the consistent memory allocation.
- */
-static int __init consistent_init(void)
-{
- pgd_t *pgd;
- pmd_t *pmd;
- pte_t *pte;
- int ret = 0;
-
- do {
- pgd = pgd_offset(&init_mm, CONSISTENT_BASE);
- pmd = pmd_alloc(&init_mm, pgd, CONSISTENT_BASE);
- if (!pmd) {
- pr_err("%s: no pmd tables\n", __func__);
- ret = -ENOMEM;
- break;
- }
- /* The first level mapping may be created in somewhere.
- * It's not necessary to warn here. */
- /* WARN_ON(!pmd_none(*pmd)); */
-
- pte = pte_alloc_kernel(pmd, CONSISTENT_BASE);
- if (!pte) {
- ret = -ENOMEM;
- break;
- }
-
- consistent_pte = pte;
- } while (0);
-
- return ret;
-}
-
-core_initcall(consistent_init);
-
static inline void cache_op(phys_addr_t paddr, size_t size,
void (*fn)(unsigned long start, unsigned long end))
{
@@ -389,3 +75,14 @@ void arch_sync_dma_for_cpu(struct device *dev, phys_addr_t paddr,
BUG();
}
}
+
+void arch_dma_prep_coherent(struct page *page, size_t size)
+{
+ cache_op(page_to_phys(page), size, cpu_dma_wbinval_range);
+}
+
+static int __init atomic_pool_init(void)
+{
+ return dma_atomic_pool_init(GFP_KERNEL, pgprot_noncached(PAGE_KERNEL));
+}
+postcore_initcall(atomic_pool_init);
--
2.20.1
Christoph Hellwig <[email protected]> 於 2019年6月14日 週五 下午6:09寫道：
>
> Hi Greentime and Vicent,
>
> can you take a look at the (untested) patch below? It converts nds32
> to use the generic remapping DMA allocator, which is also used by
> arm64 and csky.
Hi Christoph,
It looks good to me. I just verified in nds32 platform and it works fine.
Should I put it in my next-tree or you will pick it up in your tree? :)
Tested-by: Greentime Hu <[email protected]>
Reviewed-by: Greentime Hu <[email protected]>
On Fri, Jun 14, 2019 at 07:35:29PM +0800, Greentime Hu wrote:
> It looks good to me. I just verified in nds32 platform and it works fine.
> Should I put it in my next-tree or you will pick it up in your tree? :)
Either way works for me, let me know what you prefer.
Christoph Hellwig <[email protected]> 於 2019年6月14日 週五 下午8:22寫道：
>
> On Fri, Jun 14, 2019 at 07:35:29PM +0800, Greentime Hu wrote:
> > It looks good to me. I just verified in nds32 platform and it works fine.
> > Should I put it in my next-tree or you will pick it up in your tree? :)
>
> Either way works for me, let me know what you prefer.
I prefer to put in your tree. Thanks. :)
Thanks,
applied to dma-mapping for-next.