Re: [PATCH] mm: Generalize HUGETLB_PAGE_SIZE_VARIABLE
On Mon, Mar 01, 2021 at 01:13:41PM +0530, Anshuman Khandual wrote: > > doesn't this need a 'if HUGETLB_PAGE' > > While making HUGETLB_PAGE_SIZE_VARIABLE a generic option, also made it > dependent on HUGETLB_PAGE. Should not that gate HUGETLB_PAGE_SIZE_VARIABLE > when HUGETLB_PAGE is not available irrespective of the select statement on > the platforms ? depends doesn't properly work for variables that are selected.
[PATCH 14/14] swiotlb: remove swiotlb_nr_tbl
All callers just use it to check if swiotlb is active at all, for which
they can just use is_swiotlb_active. In the longer run drivers need
to stop using is_swiotlb_active as well, but let's do the simple step
first.
Signed-off-by: Christoph Hellwig
---
drivers/gpu/drm/i915/gem/i915_gem_internal.c | 2 +-
drivers/gpu/drm/nouveau/nouveau_ttm.c| 2 +-
drivers/pci/xen-pcifront.c | 2 +-
include/linux/swiotlb.h | 1 -
kernel/dma/swiotlb.c | 7 +--
5 files changed, 4 insertions(+), 10 deletions(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_internal.c
b/drivers/gpu/drm/i915/gem/i915_gem_internal.c
index ad22f42541bda6..a9d65fc8aa0eab 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_internal.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_internal.c
@@ -42,7 +42,7 @@ static int i915_gem_object_get_pages_internal(struct
drm_i915_gem_object *obj)
max_order = MAX_ORDER;
#ifdef CONFIG_SWIOTLB
- if (swiotlb_nr_tbl()) {
+ if (is_swiotlb_active()) {
unsigned int max_segment;
max_segment = swiotlb_max_segment();
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c
b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index a37bc3d7b38b3b..9662522aa0664a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -321,7 +321,7 @@ nouveau_ttm_init(struct nouveau_drm *drm)
}
#if IS_ENABLED(CONFIG_SWIOTLB) && IS_ENABLED(CONFIG_X86)
- need_swiotlb = !!swiotlb_nr_tbl();
+ need_swiotlb = is_swiotlb_active();
#endif
ret = ttm_bo_device_init(>ttm.bdev, _bo_driver,
diff --git a/drivers/pci/xen-pcifront.c b/drivers/pci/xen-pcifront.c
index c6fe0cfec0f681..a549e822033fd6 100644
--- a/drivers/pci/xen-pcifront.c
+++ b/drivers/pci/xen-pcifront.c
@@ -693,7 +693,7 @@ static int pcifront_connect_and_init_dma(struct
pcifront_device *pdev)
spin_unlock(_dev_lock);
- if (!err && !swiotlb_nr_tbl()) {
+ if (!err && !is_swiotlb_active()) {
err = pci_xen_swiotlb_init_late();
if (err)
dev_err(>xdev->dev, "Could not setup SWIOTLB!\n");
diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 63f7a63f61d098..216854a5e5134b 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -37,7 +37,6 @@ enum swiotlb_force {
extern void swiotlb_init(int verbose);
int swiotlb_init_with_tbl(char *tlb, unsigned long nslabs, int verbose);
-extern unsigned long swiotlb_nr_tbl(void);
unsigned long swiotlb_size_or_default(void);
extern int swiotlb_late_init_with_tbl(char *tlb, unsigned long nslabs);
extern int swiotlb_late_init_with_default_size(size_t default_size);
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index b7bcd7b804bfe8..809d5fdc204675 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -94,12 +94,6 @@ setup_io_tlb_npages(char *str)
}
early_param("swiotlb", setup_io_tlb_npages);
-unsigned long swiotlb_nr_tbl(void)
-{
- return io_tlb_default_mem ? io_tlb_default_mem->nslabs : 0;
-}
-EXPORT_SYMBOL_GPL(swiotlb_nr_tbl);
-
unsigned int swiotlb_max_segment(void)
{
return io_tlb_default_mem ? max_segment : 0;
@@ -652,6 +646,7 @@ bool is_swiotlb_active(void)
{
return io_tlb_default_mem != NULL;
}
+EXPORT_SYMBOL_GPL(is_swiotlb_active);
#ifdef CONFIG_DEBUG_FS
--
2.29.2
[PATCH 13/14] swiotlb: dynamically allocate io_tlb_default_mem
Instead of allocating ->list and ->orig_addr separately just do one
dynamic allocation for the actual io_tlb_mem structure. This simplifies
a lot of the initialization code, and also allows to just check
io_tlb_default_mem to see if swiotlb is in use.
Signed-off-by: Christoph Hellwig
---
drivers/xen/swiotlb-xen.c | 22 +--
include/linux/swiotlb.h | 18 ++-
kernel/dma/swiotlb.c | 300 +-
3 files changed, 117 insertions(+), 223 deletions(-)
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 5329ad54a5f34e..4c89afc0df6289 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -158,17 +158,14 @@ static const char *xen_swiotlb_error(enum xen_swiotlb_err
err)
int __ref xen_swiotlb_init(void)
{
enum xen_swiotlb_err m_ret = XEN_SWIOTLB_UNKNOWN;
- unsigned long nslabs, bytes, order;
- unsigned int repeat = 3;
+ unsigned long bytes = swiotlb_size_or_default();
+ unsigned long nslabs = bytes >> IO_TLB_SHIFT;
+ unsigned int order, repeat = 3;
int rc = -ENOMEM;
char *start;
- nslabs = swiotlb_nr_tbl();
- if (!nslabs)
- nslabs = DEFAULT_NSLABS;
retry:
m_ret = XEN_SWIOTLB_ENOMEM;
- bytes = nslabs << IO_TLB_SHIFT;
order = get_order(bytes);
/*
@@ -221,19 +218,16 @@ int __ref xen_swiotlb_init(void)
#ifdef CONFIG_X86
void __init xen_swiotlb_init_early(void)
{
- unsigned long nslabs, bytes;
+ unsigned long bytes = swiotlb_size_or_default();
+ unsigned long nslabs = bytes >> IO_TLB_SHIFT;
unsigned int repeat = 3;
char *start;
int rc;
- nslabs = swiotlb_nr_tbl();
- if (!nslabs)
- nslabs = DEFAULT_NSLABS;
retry:
/*
* Get IO TLB memory from any location.
*/
- bytes = nslabs << IO_TLB_SHIFT;
start = memblock_alloc(PAGE_ALIGN(bytes), PAGE_SIZE);
if (!start)
panic("%s: Failed to allocate %lu bytes align=0x%lx\n",
@@ -248,8 +242,8 @@ void __init xen_swiotlb_init_early(void)
if (repeat--) {
/* Min is 2MB */
nslabs = max(1024UL, (nslabs >> 1));
- pr_info("Lowering to %luMB\n",
- (nslabs << IO_TLB_SHIFT) >> 20);
+ bytes = nslabs << IO_TLB_SHIFT;
+ pr_info("Lowering to %luMB\n", bytes >> 20);
goto retry;
}
panic("%s (rc:%d)", xen_swiotlb_error(XEN_SWIOTLB_EFIXUP), rc);
@@ -548,7 +542,7 @@ xen_swiotlb_sync_sg_for_device(struct device *dev, struct
scatterlist *sgl,
static int
xen_swiotlb_dma_supported(struct device *hwdev, u64 mask)
{
- return xen_phys_to_dma(hwdev, io_tlb_default_mem.end - 1) <= mask;
+ return xen_phys_to_dma(hwdev, io_tlb_default_mem->end - 1) <= mask;
}
const struct dma_map_ops xen_swiotlb_dma_ops = {
diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 5ec5378b17c333..63f7a63f61d098 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -90,28 +90,30 @@ struct io_tlb_mem {
phys_addr_t end;
unsigned long nslabs;
unsigned long used;
- unsigned int *list;
unsigned int index;
- phys_addr_t *orig_addr;
- size_t *alloc_size;
spinlock_t lock;
struct dentry *debugfs;
bool late_alloc;
+ struct io_tlb_slot {
+ phys_addr_t orig_addr;
+ size_t alloc_size;
+ unsigned int list;
+ } slots[];
};
-extern struct io_tlb_mem io_tlb_default_mem;
+extern struct io_tlb_mem *io_tlb_default_mem;
static inline bool is_swiotlb_buffer(phys_addr_t paddr)
{
- struct io_tlb_mem *mem = _tlb_default_mem;
+ struct io_tlb_mem *mem = io_tlb_default_mem;
- return paddr >= mem->start && paddr < mem->end;
+ return mem && paddr >= mem->start && paddr < mem->end;
}
void __init swiotlb_exit(void);
unsigned int swiotlb_max_segment(void);
size_t swiotlb_max_mapping_size(struct device *dev);
bool is_swiotlb_active(void);
-void __init swiotlb_adjust_size(unsigned long new_size);
+void __init swiotlb_adjust_size(unsigned long size);
#else
#define swiotlb_force SWIOTLB_NO_FORCE
static inline bool is_swiotlb_buffer(phys_addr_t paddr)
@@ -135,7 +137,7 @@ static inline bool is_swiotlb_active(void)
return false;
}
-static inline void swiotlb_adjust_size(unsigned long new_size)
+static inline void swiotlb_adjust_size(unsigned long size)
{
}
#endif /* CONFIG_SWIOTLB */
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index 6aa84fa3b1467e..b7bcd7b804bfe8 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -63,7 +63,7 @@
enum swiotlb_force swiotlb_force;
-struct io_tlb_mem io_tlb_default_mem;
+struct io_tlb_mem *io_tlb_default_mem;
/*
* Max segment that we can
[PATCH 12/14] swiotlb: move global variables into a new io_tlb_mem structure
From: Claire Chang
Added a new struct, io_tlb_mem, as the IO TLB memory pool descriptor and
moved relevant global variables into that struct.
This will be useful later to allow for restricted DMA pool.
Signed-off-by: Claire Chang
[hch: rebased]
Signed-off-by: Christoph Hellwig
---
drivers/xen/swiotlb-xen.c | 2 +-
include/linux/swiotlb.h | 43 -
kernel/dma/swiotlb.c | 354 +-
3 files changed, 203 insertions(+), 196 deletions(-)
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 4ecfce2c6f7263..5329ad54a5f34e 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -548,7 +548,7 @@ xen_swiotlb_sync_sg_for_device(struct device *dev, struct
scatterlist *sgl,
static int
xen_swiotlb_dma_supported(struct device *hwdev, u64 mask)
{
- return xen_phys_to_dma(hwdev, io_tlb_end - 1) <= mask;
+ return xen_phys_to_dma(hwdev, io_tlb_default_mem.end - 1) <= mask;
}
const struct dma_map_ops xen_swiotlb_dma_ops = {
diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 0696bdc8072e97..5ec5378b17c333 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -6,6 +6,7 @@
#include
#include
#include
+#include
struct device;
struct page;
@@ -61,11 +62,49 @@ dma_addr_t swiotlb_map(struct device *dev, phys_addr_t phys,
#ifdef CONFIG_SWIOTLB
extern enum swiotlb_force swiotlb_force;
-extern phys_addr_t io_tlb_start, io_tlb_end;
+
+/**
+ * struct io_tlb_mem - IO TLB Memory Pool Descriptor
+ *
+ * @start: The start address of the swiotlb memory pool. Used to do a quick
+ * range check to see if the memory was in fact allocated by this
+ * API.
+ * @end: The end address of the swiotlb memory pool. Used to do a quick
+ * range check to see if the memory was in fact allocated by this
+ * API.
+ * @nslabs:The number of IO TLB blocks (in groups of 64) between @start and
+ * @end. This is command line adjustable via setup_io_tlb_npages.
+ * @used: The number of used IO TLB block.
+ * @list: The free list describing the number of free entries available
+ * from each index.
+ * @index: The index to start searching in the next round.
+ * @orig_addr: The original address corresponding to a mapped entry.
+ * @alloc_size:Size of the allocated buffer.
+ * @lock: The lock to protect the above data structures in the map and
+ * unmap calls.
+ * @debugfs: The dentry to debugfs.
+ * @late_alloc:%true if allocated using the page allocator
+ */
+struct io_tlb_mem {
+ phys_addr_t start;
+ phys_addr_t end;
+ unsigned long nslabs;
+ unsigned long used;
+ unsigned int *list;
+ unsigned int index;
+ phys_addr_t *orig_addr;
+ size_t *alloc_size;
+ spinlock_t lock;
+ struct dentry *debugfs;
+ bool late_alloc;
+};
+extern struct io_tlb_mem io_tlb_default_mem;
static inline bool is_swiotlb_buffer(phys_addr_t paddr)
{
- return paddr >= io_tlb_start && paddr < io_tlb_end;
+ struct io_tlb_mem *mem = _tlb_default_mem;
+
+ return paddr >= mem->start && paddr < mem->end;
}
void __init swiotlb_exit(void);
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index ebe7c123e27e52..6aa84fa3b1467e 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -59,32 +59,11 @@
*/
#define IO_TLB_MIN_SLABS ((1<<20) >> IO_TLB_SHIFT)
-enum swiotlb_force swiotlb_force;
-
-/*
- * Used to do a quick range check in swiotlb_tbl_unmap_single and
- * swiotlb_tbl_sync_single_*, to see if the memory was in fact allocated by
this
- * API.
- */
-phys_addr_t io_tlb_start, io_tlb_end;
-
-/*
- * The number of IO TLB blocks (in groups of 64) between io_tlb_start and
- * io_tlb_end. This is command line adjustable via setup_io_tlb_npages.
- */
-static unsigned long io_tlb_nslabs;
+#define INVALID_PHYS_ADDR (~(phys_addr_t)0)
-/*
- * The number of used IO TLB block
- */
-static unsigned long io_tlb_used;
+enum swiotlb_force swiotlb_force;
-/*
- * This is a free list describing the number of free entries available from
- * each index
- */
-static unsigned int *io_tlb_list;
-static unsigned int io_tlb_index;
+struct io_tlb_mem io_tlb_default_mem;
/*
* Max segment that we can provide which (if pages are contingous) will
@@ -92,32 +71,15 @@ static unsigned int io_tlb_index;
*/
static unsigned int max_segment;
-/*
- * We need to save away the original address corresponding to a mapped entry
- * for the sync operations.
- */
-#define INVALID_PHYS_ADDR (~(phys_addr_t)0)
-static phys_addr_t *io_tlb_orig_addr;
-
-/*
- * The mapped buffer's size should be validated during a sync operation.
- */
-static size_t *io_tlb_alloc_size;
-
-/*
- * Protect the above data structures in the map and unmap calls
- */
-static DEFINE_SPINLOCK(io_tlb_lock);
-
-static int late_alloc;
-
static int __init
[PATCH 11/14] xen-swiotlb: remove the unused size argument from xen_swiotlb_fixup
Signed-off-by: Christoph Hellwig
---
drivers/xen/swiotlb-xen.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 00adeb95ebb9df..4ecfce2c6f7263 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -104,8 +104,7 @@ static int is_xen_swiotlb_buffer(struct device *dev,
dma_addr_t dma_addr)
return 0;
}
-static int
-xen_swiotlb_fixup(void *buf, size_t size, unsigned long nslabs)
+static int xen_swiotlb_fixup(void *buf, unsigned long nslabs)
{
int i, rc;
int dma_bits;
@@ -195,7 +194,7 @@ int __ref xen_swiotlb_init(void)
/*
* And replace that memory with pages under 4GB.
*/
- rc = xen_swiotlb_fixup(start, bytes, nslabs);
+ rc = xen_swiotlb_fixup(start, nslabs);
if (rc) {
free_pages((unsigned long)start, order);
m_ret = XEN_SWIOTLB_EFIXUP;
@@ -243,7 +242,7 @@ void __init xen_swiotlb_init_early(void)
/*
* And replace that memory with pages under 4GB.
*/
- rc = xen_swiotlb_fixup(start, bytes, nslabs);
+ rc = xen_swiotlb_fixup(start, nslabs);
if (rc) {
memblock_free(__pa(start), PAGE_ALIGN(bytes));
if (repeat--) {
--
2.29.2
[PATCH 10/14] xen-swiotlb: split xen_swiotlb_init
Split xen_swiotlb_init into a normal an an early case. That makes both
much simpler and more readable, and also allows marking the early
code as __init and x86-only.
Signed-off-by: Christoph Hellwig
---
arch/arm/xen/mm.c | 2 +-
arch/x86/xen/pci-swiotlb-xen.c | 4 +-
drivers/xen/swiotlb-xen.c | 124 +++--
include/xen/swiotlb-xen.h | 3 +-
4 files changed, 75 insertions(+), 58 deletions(-)
diff --git a/arch/arm/xen/mm.c b/arch/arm/xen/mm.c
index 467fa225c3d0ed..aae950cd053fea 100644
--- a/arch/arm/xen/mm.c
+++ b/arch/arm/xen/mm.c
@@ -140,7 +140,7 @@ static int __init xen_mm_init(void)
struct gnttab_cache_flush cflush;
if (!xen_initial_domain())
return 0;
- xen_swiotlb_init(1, false);
+ xen_swiotlb_init();
cflush.op = 0;
cflush.a.dev_bus_addr = 0;
diff --git a/arch/x86/xen/pci-swiotlb-xen.c b/arch/x86/xen/pci-swiotlb-xen.c
index 19ae3e4fe4e98e..54f9aa7e845739 100644
--- a/arch/x86/xen/pci-swiotlb-xen.c
+++ b/arch/x86/xen/pci-swiotlb-xen.c
@@ -59,7 +59,7 @@ int __init pci_xen_swiotlb_detect(void)
void __init pci_xen_swiotlb_init(void)
{
if (xen_swiotlb) {
- xen_swiotlb_init(1, true /* early */);
+ xen_swiotlb_init_early();
dma_ops = _swiotlb_dma_ops;
#ifdef CONFIG_PCI
@@ -76,7 +76,7 @@ int pci_xen_swiotlb_init_late(void)
if (xen_swiotlb)
return 0;
- rc = xen_swiotlb_init(1, false /* late */);
+ rc = xen_swiotlb_init();
if (rc)
return rc;
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 060eeb056486f5..00adeb95ebb9df 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -156,96 +156,112 @@ static const char *xen_swiotlb_error(enum
xen_swiotlb_err err)
#define DEFAULT_NSLABS ALIGN(SZ_64M >> IO_TLB_SHIFT, IO_TLB_SEGSIZE)
-int __ref xen_swiotlb_init(int verbose, bool early)
+int __ref xen_swiotlb_init(void)
{
- unsigned long bytes, order;
- int rc = -ENOMEM;
enum xen_swiotlb_err m_ret = XEN_SWIOTLB_UNKNOWN;
+ unsigned long nslabs, bytes, order;
unsigned int repeat = 3;
+ int rc = -ENOMEM;
char *start;
- unsigned long nslabs;
nslabs = swiotlb_nr_tbl();
-retry:
if (!nslabs)
nslabs = DEFAULT_NSLABS;
+retry:
+ m_ret = XEN_SWIOTLB_ENOMEM;
bytes = nslabs << IO_TLB_SHIFT;
order = get_order(bytes);
/*
* Get IO TLB memory from any location.
*/
- if (early) {
- start = memblock_alloc(PAGE_ALIGN(bytes),
- PAGE_SIZE);
- if (!start)
- panic("%s: Failed to allocate %lu bytes align=0x%lx\n",
- __func__, PAGE_ALIGN(bytes), PAGE_SIZE);
- } else {
#define SLABS_PER_PAGE (1 << (PAGE_SHIFT - IO_TLB_SHIFT))
#define IO_TLB_MIN_SLABS ((1<<20) >> IO_TLB_SHIFT)
- while ((SLABS_PER_PAGE << order) > IO_TLB_MIN_SLABS) {
- start = (void *)xen_get_swiotlb_free_pages(order);
- if (start)
- break;
- order--;
- }
- if (order != get_order(bytes)) {
- pr_warn("Warning: only able to allocate %ld MB for
software IO TLB\n",
- (PAGE_SIZE << order) >> 20);
- nslabs = SLABS_PER_PAGE << order;
- bytes = nslabs << IO_TLB_SHIFT;
- }
+ while ((SLABS_PER_PAGE << order) > IO_TLB_MIN_SLABS) {
+ start = (void *)xen_get_swiotlb_free_pages(order);
+ if (start)
+ break;
+ order--;
}
- if (!start) {
- m_ret = XEN_SWIOTLB_ENOMEM;
+ if (!start)
goto error;
+ if (order != get_order(bytes)) {
+ pr_warn("Warning: only able to allocate %ld MB for software IO
TLB\n",
+ (PAGE_SIZE << order) >> 20);
+ nslabs = SLABS_PER_PAGE << order;
+ bytes = nslabs << IO_TLB_SHIFT;
}
+
/*
* And replace that memory with pages under 4GB.
*/
- rc = xen_swiotlb_fixup(start,
- bytes,
- nslabs);
+ rc = xen_swiotlb_fixup(start, bytes, nslabs);
if (rc) {
- if (early)
- memblock_free(__pa(start),
- PAGE_ALIGN(bytes));
- else {
- free_pages((unsigned long)start, order);
- start = NULL;
- }
+ free_pages((unsigned long)start, order);
m_ret = XEN_SWIOTLB_EFIXUP;
goto error;
}
- if
[PATCH 09/14] swiotlb: lift the double initialization protection from xen-swiotlb
Lift the double initialization protection from xen-swiotlb to the core
code to avoid exposing too many swiotlb internals. Also upgrade the
check to a warning as it should not happen.
Signed-off-by: Christoph Hellwig
---
drivers/xen/swiotlb-xen.c | 7 ---
kernel/dma/swiotlb.c | 8
2 files changed, 8 insertions(+), 7 deletions(-)
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 1a31ddf7139799..060eeb056486f5 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -172,12 +172,6 @@ int __ref xen_swiotlb_init(int verbose, bool early)
bytes = nslabs << IO_TLB_SHIFT;
order = get_order(bytes);
- /*
-* IO TLB memory already allocated. Just use it.
-*/
- if (io_tlb_start != 0)
- goto end;
-
/*
* Get IO TLB memory from any location.
*/
@@ -232,7 +226,6 @@ int __ref xen_swiotlb_init(int verbose, bool early)
} else
rc = swiotlb_late_init_with_tbl(start, nslabs);
-end:
if (!rc)
swiotlb_set_max_segment(PAGE_SIZE);
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index 388d9be35b5795..ebe7c123e27e52 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -229,6 +229,10 @@ int __init swiotlb_init_with_tbl(char *tlb, unsigned long
nslabs, int verbose)
unsigned long i, bytes;
size_t alloc_size;
+ /* protect against double initialization */
+ if (WARN_ON_ONCE(io_tlb_start))
+ return -ENOMEM;
+
bytes = nslabs << IO_TLB_SHIFT;
io_tlb_nslabs = nslabs;
@@ -367,6 +371,10 @@ swiotlb_late_init_with_tbl(char *tlb, unsigned long nslabs)
{
unsigned long i, bytes;
+ /* protect against double initialization */
+ if (WARN_ON_ONCE(io_tlb_start))
+ return -ENOMEM;
+
bytes = nslabs << IO_TLB_SHIFT;
io_tlb_nslabs = nslabs;
--
2.29.2
[PATCH 08/14] xen-swiotlb: remove xen_io_tlb_start and xen_io_tlb_nslabs
The xen_io_tlb_start and xen_io_tlb_nslabs variables ar now only used in
xen_swiotlb_init, so replace them with local variables.
Signed-off-by: Christoph Hellwig
---
drivers/xen/swiotlb-xen.c | 57 +--
1 file changed, 25 insertions(+), 32 deletions(-)
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 5352655432e724..1a31ddf7139799 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -40,14 +40,7 @@
#include
#define MAX_DMA_BITS 32
-/*
- * Used to do a quick range check in swiotlb_tbl_unmap_single and
- * swiotlb_tbl_sync_single_*, to see if the memory was in fact allocated by
this
- * API.
- */
-static char *xen_io_tlb_start;
-static unsigned long xen_io_tlb_nslabs;
/*
* Quick lookup value of the bus address of the IOTLB.
*/
@@ -169,75 +162,75 @@ int __ref xen_swiotlb_init(int verbose, bool early)
int rc = -ENOMEM;
enum xen_swiotlb_err m_ret = XEN_SWIOTLB_UNKNOWN;
unsigned int repeat = 3;
+ char *start;
+ unsigned long nslabs;
- xen_io_tlb_nslabs = swiotlb_nr_tbl();
+ nslabs = swiotlb_nr_tbl();
retry:
- if (!xen_io_tlb_nslabs)
- xen_io_tlb_nslabs = DEFAULT_NSLABS;
- bytes = xen_io_tlb_nslabs << IO_TLB_SHIFT;
+ if (!nslabs)
+ nslabs = DEFAULT_NSLABS;
+ bytes = nslabs << IO_TLB_SHIFT;
order = get_order(bytes);
/*
* IO TLB memory already allocated. Just use it.
*/
- if (io_tlb_start != 0) {
- xen_io_tlb_start = phys_to_virt(io_tlb_start);
+ if (io_tlb_start != 0)
goto end;
- }
/*
* Get IO TLB memory from any location.
*/
if (early) {
- xen_io_tlb_start = memblock_alloc(PAGE_ALIGN(bytes),
+ start = memblock_alloc(PAGE_ALIGN(bytes),
PAGE_SIZE);
- if (!xen_io_tlb_start)
+ if (!start)
panic("%s: Failed to allocate %lu bytes align=0x%lx\n",
__func__, PAGE_ALIGN(bytes), PAGE_SIZE);
} else {
#define SLABS_PER_PAGE (1 << (PAGE_SHIFT - IO_TLB_SHIFT))
#define IO_TLB_MIN_SLABS ((1<<20) >> IO_TLB_SHIFT)
while ((SLABS_PER_PAGE << order) > IO_TLB_MIN_SLABS) {
- xen_io_tlb_start = (void
*)xen_get_swiotlb_free_pages(order);
- if (xen_io_tlb_start)
+ start = (void *)xen_get_swiotlb_free_pages(order);
+ if (start)
break;
order--;
}
if (order != get_order(bytes)) {
pr_warn("Warning: only able to allocate %ld MB for
software IO TLB\n",
(PAGE_SIZE << order) >> 20);
- xen_io_tlb_nslabs = SLABS_PER_PAGE << order;
- bytes = xen_io_tlb_nslabs << IO_TLB_SHIFT;
+ nslabs = SLABS_PER_PAGE << order;
+ bytes = nslabs << IO_TLB_SHIFT;
}
}
- if (!xen_io_tlb_start) {
+ if (!start) {
m_ret = XEN_SWIOTLB_ENOMEM;
goto error;
}
/*
* And replace that memory with pages under 4GB.
*/
- rc = xen_swiotlb_fixup(xen_io_tlb_start,
+ rc = xen_swiotlb_fixup(start,
bytes,
- xen_io_tlb_nslabs);
+ nslabs);
if (rc) {
if (early)
- memblock_free(__pa(xen_io_tlb_start),
+ memblock_free(__pa(start),
PAGE_ALIGN(bytes));
else {
- free_pages((unsigned long)xen_io_tlb_start, order);
- xen_io_tlb_start = NULL;
+ free_pages((unsigned long)start, order);
+ start = NULL;
}
m_ret = XEN_SWIOTLB_EFIXUP;
goto error;
}
if (early) {
- if (swiotlb_init_with_tbl(xen_io_tlb_start, xen_io_tlb_nslabs,
+ if (swiotlb_init_with_tbl(start, nslabs,
verbose))
panic("Cannot allocate SWIOTLB buffer");
rc = 0;
} else
- rc = swiotlb_late_init_with_tbl(xen_io_tlb_start,
xen_io_tlb_nslabs);
+ rc = swiotlb_late_init_with_tbl(start, nslabs);
end:
if (!rc)
@@ -246,17 +239,17 @@ int __ref xen_swiotlb_init(int verbose, bool early)
return rc;
error:
if (repeat--) {
- xen_io_tlb_nslabs = max(1024UL, /* Min is 2MB */
- (xen_io_tlb_nslabs >> 1));
+ nslabs = max(1024UL, /* Min is 2MB */
+
[PATCH 07/14] xen-swiotlb: remove xen_set_nslabs
The xen_set_nslabs function is a little weird, as it has just one
caller, that caller passes a global variable as the argument,
which is then overriden in the function and a derivative of it
returned. Just add a cpp symbol for the default size using a readable
constant and open code the remaining three lines in the caller.
Signed-off-by: Christoph Hellwig
---
drivers/xen/swiotlb-xen.c | 19 +++
1 file changed, 7 insertions(+), 12 deletions(-)
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index e99f0614dcb979..5352655432e724 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -138,16 +138,6 @@ xen_swiotlb_fixup(void *buf, size_t size, unsigned long
nslabs)
} while (i < nslabs);
return 0;
}
-static unsigned long xen_set_nslabs(unsigned long nr_tbl)
-{
- if (!nr_tbl) {
- xen_io_tlb_nslabs = (64 * 1024 * 1024 >> IO_TLB_SHIFT);
- xen_io_tlb_nslabs = ALIGN(xen_io_tlb_nslabs, IO_TLB_SEGSIZE);
- } else
- xen_io_tlb_nslabs = nr_tbl;
-
- return xen_io_tlb_nslabs << IO_TLB_SHIFT;
-}
enum xen_swiotlb_err {
XEN_SWIOTLB_UNKNOWN = 0,
@@ -170,6 +160,9 @@ static const char *xen_swiotlb_error(enum xen_swiotlb_err
err)
}
return "";
}
+
+#define DEFAULT_NSLABS ALIGN(SZ_64M >> IO_TLB_SHIFT, IO_TLB_SEGSIZE)
+
int __ref xen_swiotlb_init(int verbose, bool early)
{
unsigned long bytes, order;
@@ -179,8 +172,10 @@ int __ref xen_swiotlb_init(int verbose, bool early)
xen_io_tlb_nslabs = swiotlb_nr_tbl();
retry:
- bytes = xen_set_nslabs(xen_io_tlb_nslabs);
- order = get_order(xen_io_tlb_nslabs << IO_TLB_SHIFT);
+ if (!xen_io_tlb_nslabs)
+ xen_io_tlb_nslabs = DEFAULT_NSLABS;
+ bytes = xen_io_tlb_nslabs << IO_TLB_SHIFT;
+ order = get_order(bytes);
/*
* IO TLB memory already allocated. Just use it.
--
2.29.2
[PATCH 06/14] xen-swiotlb: use io_tlb_end in xen_swiotlb_dma_supported
Use the existing variable that holds the physical address for
xen_io_tlb_end to simplify xen_swiotlb_dma_supported a bit, and remove
the otherwise unused xen_io_tlb_end variable and the xen_virt_to_bus
helper.
Signed-off-by: Christoph Hellwig
Reviewed-by: Konrad Rzeszutek Wilk
---
drivers/xen/swiotlb-xen.c | 10 ++
1 file changed, 2 insertions(+), 8 deletions(-)
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index bffb35993c9d5f..e99f0614dcb979 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -46,7 +46,7 @@
* API.
*/
-static char *xen_io_tlb_start, *xen_io_tlb_end;
+static char *xen_io_tlb_start;
static unsigned long xen_io_tlb_nslabs;
/*
* Quick lookup value of the bus address of the IOTLB.
@@ -82,11 +82,6 @@ static inline phys_addr_t xen_dma_to_phys(struct device *dev,
return xen_bus_to_phys(dev, dma_to_phys(dev, dma_addr));
}
-static inline dma_addr_t xen_virt_to_bus(struct device *dev, void *address)
-{
- return xen_phys_to_dma(dev, virt_to_phys(address));
-}
-
static inline int range_straddles_page_boundary(phys_addr_t p, size_t size)
{
unsigned long next_bfn, xen_pfn = XEN_PFN_DOWN(p);
@@ -250,7 +245,6 @@ int __ref xen_swiotlb_init(int verbose, bool early)
rc = swiotlb_late_init_with_tbl(xen_io_tlb_start,
xen_io_tlb_nslabs);
end:
- xen_io_tlb_end = xen_io_tlb_start + bytes;
if (!rc)
swiotlb_set_max_segment(PAGE_SIZE);
@@ -558,7 +552,7 @@ xen_swiotlb_sync_sg_for_device(struct device *dev, struct
scatterlist *sgl,
static int
xen_swiotlb_dma_supported(struct device *hwdev, u64 mask)
{
- return xen_virt_to_bus(hwdev, xen_io_tlb_end - 1) <= mask;
+ return xen_phys_to_dma(hwdev, io_tlb_end - 1) <= mask;
}
const struct dma_map_ops xen_swiotlb_dma_ops = {
--
2.29.2
[PATCH 05/14] xen-swiotlb: use is_swiotlb_buffer in is_xen_swiotlb_buffer
Use the is_swiotlb_buffer to check if a physical address is
a swiotlb buffer. This works because xen-swiotlb does use the
same buffer as the main swiotlb code, and xen_io_tlb_{start,end}
are just the addresses for it that went through phys_to_virt.
Signed-off-by: Christoph Hellwig
Reviewed-by: Konrad Rzeszutek Wilk
---
drivers/xen/swiotlb-xen.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 4e8a4e14942afd..bffb35993c9d5f 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -111,10 +111,8 @@ static int is_xen_swiotlb_buffer(struct device *dev,
dma_addr_t dma_addr)
* have the same virtual address as another address
* in our domain. Therefore _only_ check address within our domain.
*/
- if (pfn_valid(PFN_DOWN(paddr))) {
- return paddr >= virt_to_phys(xen_io_tlb_start) &&
- paddr < virt_to_phys(xen_io_tlb_end);
- }
+ if (pfn_valid(PFN_DOWN(paddr)))
+ return is_swiotlb_buffer(paddr);
return 0;
}
--
2.29.2
[PATCH 04/14] swiotlb: split swiotlb_tbl_sync_single
Split swiotlb_tbl_sync_single into two separate funtions for the to device
and to cpu synchronization.
Signed-off-by: Christoph Hellwig
---
drivers/iommu/dma-iommu.c | 12 ++--
drivers/xen/swiotlb-xen.c | 4 ++--
include/linux/swiotlb.h | 17 -
kernel/dma/direct.c | 8
kernel/dma/direct.h | 4 ++--
kernel/dma/swiotlb.c | 34 +++---
6 files changed, 33 insertions(+), 46 deletions(-)
diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index da2bd8f0885e6e..b57a0e3e21f6c7 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -749,7 +749,7 @@ static void iommu_dma_sync_single_for_cpu(struct device
*dev,
arch_sync_dma_for_cpu(phys, size, dir);
if (is_swiotlb_buffer(phys))
- swiotlb_tbl_sync_single(dev, phys, size, dir, SYNC_FOR_CPU);
+ swiotlb_sync_single_for_cpu(dev, phys, size, dir);
}
static void iommu_dma_sync_single_for_device(struct device *dev,
@@ -762,7 +762,7 @@ static void iommu_dma_sync_single_for_device(struct device
*dev,
phys = iommu_iova_to_phys(iommu_get_dma_domain(dev), dma_handle);
if (is_swiotlb_buffer(phys))
- swiotlb_tbl_sync_single(dev, phys, size, dir, SYNC_FOR_DEVICE);
+ swiotlb_sync_single_for_device(dev, phys, size, dir);
if (!dev_is_dma_coherent(dev))
arch_sync_dma_for_device(phys, size, dir);
@@ -783,8 +783,8 @@ static void iommu_dma_sync_sg_for_cpu(struct device *dev,
arch_sync_dma_for_cpu(sg_phys(sg), sg->length, dir);
if (is_swiotlb_buffer(sg_phys(sg)))
- swiotlb_tbl_sync_single(dev, sg_phys(sg), sg->length,
- dir, SYNC_FOR_CPU);
+ swiotlb_sync_single_for_cpu(dev, sg_phys(sg),
+ sg->length, dir);
}
}
@@ -800,8 +800,8 @@ static void iommu_dma_sync_sg_for_device(struct device *dev,
for_each_sg(sgl, sg, nelems, i) {
if (is_swiotlb_buffer(sg_phys(sg)))
- swiotlb_tbl_sync_single(dev, sg_phys(sg), sg->length,
- dir, SYNC_FOR_DEVICE);
+ swiotlb_sync_single_for_device(dev, sg_phys(sg),
+ sg->length, dir);
if (!dev_is_dma_coherent(dev))
arch_sync_dma_for_device(sg_phys(sg), sg->length, dir);
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index d47f1b311caac0..4e8a4e14942afd 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -462,7 +462,7 @@ xen_swiotlb_sync_single_for_cpu(struct device *dev,
dma_addr_t dma_addr,
}
if (is_xen_swiotlb_buffer(dev, dma_addr))
- swiotlb_tbl_sync_single(dev, paddr, size, dir, SYNC_FOR_CPU);
+ swiotlb_sync_single_for_cpu(dev, paddr, size, dir);
}
static void
@@ -472,7 +472,7 @@ xen_swiotlb_sync_single_for_device(struct device *dev,
dma_addr_t dma_addr,
phys_addr_t paddr = xen_dma_to_phys(dev, dma_addr);
if (is_xen_swiotlb_buffer(dev, dma_addr))
- swiotlb_tbl_sync_single(dev, paddr, size, dir, SYNC_FOR_DEVICE);
+ swiotlb_sync_single_for_device(dev, paddr, size, dir);
if (!dev_is_dma_coherent(dev)) {
if (pfn_valid(PFN_DOWN(dma_to_phys(dev, dma_addr
diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 59f421d041ed9e..0696bdc8072e97 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -42,14 +42,6 @@ extern int swiotlb_late_init_with_tbl(char *tlb, unsigned
long nslabs);
extern int swiotlb_late_init_with_default_size(size_t default_size);
extern void __init swiotlb_update_mem_attributes(void);
-/*
- * Enumeration for sync targets
- */
-enum dma_sync_target {
- SYNC_FOR_CPU = 0,
- SYNC_FOR_DEVICE = 1,
-};
-
phys_addr_t swiotlb_tbl_map_single(struct device *hwdev, phys_addr_t phys,
size_t mapping_size, size_t alloc_size,
enum dma_data_direction dir, unsigned long attrs);
@@ -60,11 +52,10 @@ extern void swiotlb_tbl_unmap_single(struct device *hwdev,
enum dma_data_direction dir,
unsigned long attrs);
-extern void swiotlb_tbl_sync_single(struct device *hwdev,
- phys_addr_t tlb_addr,
- size_t size, enum dma_data_direction dir,
- enum dma_sync_target target);
-
+void swiotlb_sync_single_for_device(struct device *dev, phys_addr_t tlb_addr,
+ size_t size, enum dma_data_direction dir);
+void swiotlb_sync_single_for_cpu(struct device *dev, phys_addr_t tlb_addr,
+ size_t size, enum
[PATCH 03/14] swiotlb: move orig addr and size validation into swiotlb_bounce
Move the code to find and validate the original buffer address and size
from the callers into swiotlb_bounce. This means a tiny bit of extra
work in the swiotlb_map path, but avoids code duplication and a leads to
a better code structure.
Signed-off-by: Christoph Hellwig
---
kernel/dma/swiotlb.c | 59 +---
1 file changed, 23 insertions(+), 36 deletions(-)
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index 03aa614565e417..a9063092f6f566 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -460,12 +460,25 @@ void __init swiotlb_exit(void)
/*
* Bounce: copy the swiotlb buffer from or back to the original dma location
*/
-static void swiotlb_bounce(phys_addr_t orig_addr, phys_addr_t tlb_addr,
- size_t size, enum dma_data_direction dir)
+static void swiotlb_bounce(struct device *dev, phys_addr_t tlb_addr, size_t
size,
+ enum dma_data_direction dir)
{
+ int index = (tlb_addr - io_tlb_start) >> IO_TLB_SHIFT;
+ size_t alloc_size = io_tlb_alloc_size[index];
+ phys_addr_t orig_addr = io_tlb_orig_addr[index];
unsigned long pfn = PFN_DOWN(orig_addr);
unsigned char *vaddr = phys_to_virt(tlb_addr);
+ if (orig_addr == INVALID_PHYS_ADDR)
+ return;
+
+ if (size > alloc_size) {
+ dev_WARN_ONCE(dev, 1,
+ "Buffer overflow detected. Allocation size: %zu.
Mapping size: %zu.\n",
+ alloc_size, size);
+ size = alloc_size;
+ }
+
if (PageHighMem(pfn_to_page(pfn))) {
/* The buffer does not have a mapping. Map it in and copy */
unsigned int offset = orig_addr & ~PAGE_MASK;
@@ -644,21 +657,10 @@ phys_addr_t swiotlb_tbl_map_single(struct device *dev,
phys_addr_t orig_addr,
tlb_addr = slot_addr(io_tlb_start, index) + offset;
if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC) &&
(dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL))
- swiotlb_bounce(orig_addr, tlb_addr, mapping_size,
DMA_TO_DEVICE);
+ swiotlb_bounce(dev, tlb_addr, mapping_size, DMA_TO_DEVICE);
return tlb_addr;
}
-static void validate_sync_size_and_truncate(struct device *hwdev, size_t
alloc_size, size_t *size)
-{
- if (*size > alloc_size) {
- /* Warn and truncate mapping_size */
- dev_WARN_ONCE(hwdev, 1,
- "Attempt for buffer overflow. Original size: %zu.
Mapping size: %zu.\n",
- alloc_size, *size);
- *size = alloc_size;
- }
-}
-
/*
* tlb_addr is the physical address of the bounce buffer to unmap.
*/
@@ -669,19 +671,15 @@ void swiotlb_tbl_unmap_single(struct device *hwdev,
phys_addr_t tlb_addr,
unsigned long flags;
unsigned int offset = swiotlb_align_offset(hwdev, tlb_addr);
int index = (tlb_addr - offset - io_tlb_start) >> IO_TLB_SHIFT;
- phys_addr_t orig_addr = io_tlb_orig_addr[index];
- size_t alloc_size = io_tlb_alloc_size[index];
- int i, count, nslots = nr_slots(alloc_size + offset);
-
- validate_sync_size_and_truncate(hwdev, alloc_size, _size);
+ int nslots = nr_slots(io_tlb_alloc_size[index] + offset);
+ int count, i;
/*
* First, sync the memory before unmapping the entry
*/
- if (orig_addr != INVALID_PHYS_ADDR &&
- !(attrs & DMA_ATTR_SKIP_CPU_SYNC) &&
- ((dir == DMA_FROM_DEVICE) || (dir == DMA_BIDIRECTIONAL)))
- swiotlb_bounce(orig_addr, tlb_addr, mapping_size,
DMA_FROM_DEVICE);
+ if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC) &&
+ (dir == DMA_FROM_DEVICE || dir == DMA_BIDIRECTIONAL))
+ swiotlb_bounce(hwdev, tlb_addr, mapping_size, DMA_FROM_DEVICE);
/*
* Return the buffer to the free list by setting the corresponding
@@ -721,27 +719,16 @@ void swiotlb_tbl_sync_single(struct device *hwdev,
phys_addr_t tlb_addr,
size_t size, enum dma_data_direction dir,
enum dma_sync_target target)
{
- int index = (tlb_addr - io_tlb_start) >> IO_TLB_SHIFT;
- size_t alloc_size = io_tlb_alloc_size[index];
- phys_addr_t orig_addr = io_tlb_orig_addr[index];
-
- if (orig_addr == INVALID_PHYS_ADDR)
- return;
-
- validate_sync_size_and_truncate(hwdev, alloc_size, );
-
switch (target) {
case SYNC_FOR_CPU:
if (likely(dir == DMA_FROM_DEVICE || dir == DMA_BIDIRECTIONAL))
- swiotlb_bounce(orig_addr, tlb_addr,
- size, DMA_FROM_DEVICE);
+ swiotlb_bounce(hwdev, tlb_addr, size, DMA_FROM_DEVICE);
else
BUG_ON(dir != DMA_TO_DEVICE);
break;
case SYNC_FOR_DEVICE:
if (likely(dir == DMA_TO_DEVICE ||
[PATCH 02/14] swiotlb: remove the alloc_size parameter to swiotlb_tbl_unmap_single
Now that swiotlb remembers the allocation size there is no need to pass
it back to swiotlb_tbl_unmap_single.
Signed-off-by: Christoph Hellwig
---
drivers/iommu/dma-iommu.c | 11 +++---
drivers/xen/swiotlb-xen.c | 4 ++--
include/linux/swiotlb.h | 1 -
kernel/dma/direct.h | 2 +-
kernel/dma/swiotlb.c | 45 ---
5 files changed, 29 insertions(+), 34 deletions(-)
diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 9ab6ee22c11088..da2bd8f0885e6e 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -493,8 +493,6 @@ static void __iommu_dma_unmap_swiotlb(struct device *dev,
dma_addr_t dma_addr,
unsigned long attrs)
{
struct iommu_domain *domain = iommu_get_dma_domain(dev);
- struct iommu_dma_cookie *cookie = domain->iova_cookie;
- struct iova_domain *iovad = >iovad;
phys_addr_t phys;
phys = iommu_iova_to_phys(domain, dma_addr);
@@ -504,8 +502,7 @@ static void __iommu_dma_unmap_swiotlb(struct device *dev,
dma_addr_t dma_addr,
__iommu_dma_unmap(dev, dma_addr, size);
if (unlikely(is_swiotlb_buffer(phys)))
- swiotlb_tbl_unmap_single(dev, phys, size,
- iova_align(iovad, size), dir, attrs);
+ swiotlb_tbl_unmap_single(dev, phys, size, dir, attrs);
}
static bool dev_is_untrusted(struct device *dev)
@@ -580,10 +577,8 @@ static dma_addr_t __iommu_dma_map_swiotlb(struct device
*dev, phys_addr_t phys,
}
iova = __iommu_dma_map(dev, phys, aligned_size, prot, dma_mask);
- if ((iova == DMA_MAPPING_ERROR) && is_swiotlb_buffer(phys))
- swiotlb_tbl_unmap_single(dev, phys, org_size,
- aligned_size, dir, attrs);
-
+ if (iova == DMA_MAPPING_ERROR && is_swiotlb_buffer(phys))
+ swiotlb_tbl_unmap_single(dev, phys, org_size, dir, attrs);
return iova;
}
diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 2b385c1b4a99cb..d47f1b311caac0 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -406,7 +406,7 @@ static dma_addr_t xen_swiotlb_map_page(struct device *dev,
struct page *page,
* Ensure that the address returned is DMA'ble
*/
if (unlikely(!dma_capable(dev, dev_addr, size, true))) {
- swiotlb_tbl_unmap_single(dev, map, size, size, dir,
+ swiotlb_tbl_unmap_single(dev, map, size, dir,
attrs | DMA_ATTR_SKIP_CPU_SYNC);
return DMA_MAPPING_ERROR;
}
@@ -445,7 +445,7 @@ static void xen_swiotlb_unmap_page(struct device *hwdev,
dma_addr_t dev_addr,
/* NOTE: We use dev_addr here, not paddr! */
if (is_xen_swiotlb_buffer(hwdev, dev_addr))
- swiotlb_tbl_unmap_single(hwdev, paddr, size, size, dir, attrs);
+ swiotlb_tbl_unmap_single(hwdev, paddr, size, dir, attrs);
}
static void
diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 5857a937c63722..59f421d041ed9e 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -57,7 +57,6 @@ phys_addr_t swiotlb_tbl_map_single(struct device *hwdev,
phys_addr_t phys,
extern void swiotlb_tbl_unmap_single(struct device *hwdev,
phys_addr_t tlb_addr,
size_t mapping_size,
-size_t alloc_size,
enum dma_data_direction dir,
unsigned long attrs);
diff --git a/kernel/dma/direct.h b/kernel/dma/direct.h
index b9861557873768..e1bf721591c0cf 100644
--- a/kernel/dma/direct.h
+++ b/kernel/dma/direct.h
@@ -114,6 +114,6 @@ static inline void dma_direct_unmap_page(struct device
*dev, dma_addr_t addr,
dma_direct_sync_single_for_cpu(dev, addr, size, dir);
if (unlikely(is_swiotlb_buffer(phys)))
- swiotlb_tbl_unmap_single(dev, phys, size, size, dir, attrs);
+ swiotlb_tbl_unmap_single(dev, phys, size, dir, attrs);
}
#endif /* _KERNEL_DMA_DIRECT_H */
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index c10e855a03bc16..03aa614565e417 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -102,7 +102,7 @@ static phys_addr_t *io_tlb_orig_addr;
/*
* The mapped buffer's size should be validated during a sync operation.
*/
-static size_t *io_tlb_orig_size;
+static size_t *io_tlb_alloc_size;
/*
* Protect the above data structures in the map and unmap calls
@@ -253,15 +253,15 @@ int __init swiotlb_init_with_tbl(char *tlb, unsigned long
nslabs, int verbose)
__func__, alloc_size, PAGE_SIZE);
alloc_size = PAGE_ALIGN(io_tlb_nslabs * sizeof(size_t));
- io_tlb_orig_size = memblock_alloc(alloc_size, PAGE_SIZE);
- if (!io_tlb_orig_size)
+ io_tlb_alloc_size =
swiotlb cleanups v2
Hi Konrad, this series contains a bunch of swiotlb cleanups, mostly to reduce the amount of internals exposed to code outside of swiotlb.c, which should helper to prepare for supporting multiple different bounce buffer pools. Changes since v1: - rebased to v5.12-rc1 - a few more cleanups - merge and forward port the patch from Claire to move all the global variables into a struct to prepare for multiple instances
[PATCH 01/14] powerpc/svm: stop using io_tlb_start
Use the local variable that is passed to swiotlb_init_with_tbl for
freeing the memory in the failure case to isolate the code a little
better from swiotlb internals.
Signed-off-by: Christoph Hellwig
---
arch/powerpc/platforms/pseries/svm.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/arch/powerpc/platforms/pseries/svm.c
b/arch/powerpc/platforms/pseries/svm.c
index 7b739cc7a8a93e..1d829e257996fb 100644
--- a/arch/powerpc/platforms/pseries/svm.c
+++ b/arch/powerpc/platforms/pseries/svm.c
@@ -55,9 +55,9 @@ void __init svm_swiotlb_init(void)
if (vstart && !swiotlb_init_with_tbl(vstart, io_tlb_nslabs, false))
return;
- if (io_tlb_start)
- memblock_free_early(io_tlb_start,
- PAGE_ALIGN(io_tlb_nslabs << IO_TLB_SHIFT));
+
+ memblock_free_early(__pa(vstart),
+ PAGE_ALIGN(io_tlb_nslabs << IO_TLB_SHIFT));
panic("SVM: Cannot allocate SWIOTLB buffer");
}
--
2.29.2
Re: [PATCH] mm: Generalize HUGETLB_PAGE_SIZE_VARIABLE
On 3/1/21 11:53 AM, Christoph Hellwig wrote: > On Mon, Mar 01, 2021 at 11:20:53AM +0530, Anshuman Khandual wrote: >> HUGETLB_PAGE_SIZE_VARIABLE need not be defined for each individual >> platform subscribing it. Instead just make it generic. >> >> Cc: Michael Ellerman >> Cc: Benjamin Herrenschmidt >> Cc: Paul Mackerras >> Cc: Andrew Morton >> Cc: Christoph Hellwig >> Cc: linux-i...@vger.kernel.org >> Cc: linuxppc-dev@lists.ozlabs.org >> Cc: linux...@kvack.org >> Cc: linux-ker...@vger.kernel.org >> Suggested-by: Christoph Hellwig >> Signed-off-by: Anshuman Khandual >> --- >> This change was originally suggested in an earilier discussion. This >> applies on v5.12-rc1 and has been build tested on all applicable >> platforms i.e ia64 and powerpc. >> >> https://patchwork.kernel.org/project/linux-mm/patch/1613024531-19040-3-git-send-email-anshuman.khand...@arm.com/ >> >> arch/ia64/Kconfig| 6 +- >> arch/powerpc/Kconfig | 6 +- >> mm/Kconfig | 8 >> 3 files changed, 10 insertions(+), 10 deletions(-) >> >> diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig >> index 2ad7a8d29fcc..6b3e3f6c29ae 100644 >> --- a/arch/ia64/Kconfig >> +++ b/arch/ia64/Kconfig >> @@ -32,6 +32,7 @@ config IA64 >> select TTY >> select HAVE_ARCH_TRACEHOOK >> select HAVE_VIRT_CPU_ACCOUNTING >> +select HUGETLB_PAGE_SIZE_VARIABLE > > doesn't this need a 'if HUGETLB_PAGE' While making HUGETLB_PAGE_SIZE_VARIABLE a generic option, also made it dependent on HUGETLB_PAGE. Should not that gate HUGETLB_PAGE_SIZE_VARIABLE when HUGETLB_PAGE is not available irrespective of the select statement on the platforms ? > > or did you verify that HUGETLB_PAGE_SIZE_VARIABLE checks are always > nested inside of HUGETLB_PAGE ones? >
[PATCH kernel v2] powerpc/iommu: Annotate nested lock for lockdep
The IOMMU table is divided into pools for concurrent mappings and each
pool has a separate spinlock. When taking the ownership of an IOMMU group
to pass through a device to a VM, we lock these spinlocks which triggers
a false negative warning in lockdep (below).
This fixes it by annotating the large pool's spinlock as a nest lock
which makes lockdep not complaining when locking nested locks if
the nest lock is locked already.
===
WARNING: possible recursive locking detected
5.11.0-le_syzkaller_a+fstn1 #100 Not tainted
qemu-system-ppc/4129 is trying to acquire lock:
c000119bddb0 (&(p->lock)/1){}-{2:2}, at: iommu_take_ownership+0xac/0x1e0
but task is already holding lock:
c000119bdd30 (&(p->lock)/1){}-{2:2}, at: iommu_take_ownership+0xac/0x1e0
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
lock(&(p->lock)/1);
lock(&(p->lock)/1);
===
Signed-off-by: Alexey Kardashevskiy
---
Changes:
v2:
* fixed iommu_release_ownership() as well
---
arch/powerpc/kernel/iommu.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/powerpc/kernel/iommu.c b/arch/powerpc/kernel/iommu.c
index c1a5c366a664..d0df3e5ff5e0 100644
--- a/arch/powerpc/kernel/iommu.c
+++ b/arch/powerpc/kernel/iommu.c
@@ -1089,7 +1089,7 @@ int iommu_take_ownership(struct iommu_table *tbl)
spin_lock_irqsave(>large_pool.lock, flags);
for (i = 0; i < tbl->nr_pools; i++)
- spin_lock(>pools[i].lock);
+ spin_lock_nest_lock(>pools[i].lock, >large_pool.lock);
iommu_table_release_pages(tbl);
@@ -1117,7 +1117,7 @@ void iommu_release_ownership(struct iommu_table *tbl)
spin_lock_irqsave(>large_pool.lock, flags);
for (i = 0; i < tbl->nr_pools; i++)
- spin_lock(>pools[i].lock);
+ spin_lock_nest_lock(>pools[i].lock, >large_pool.lock);
memset(tbl->it_map, 0, sz);
--
2.17.1
Re: [PATCH] mm: Generalize HUGETLB_PAGE_SIZE_VARIABLE
On Mon, Mar 01, 2021 at 11:20:53AM +0530, Anshuman Khandual wrote: > HUGETLB_PAGE_SIZE_VARIABLE need not be defined for each individual > platform subscribing it. Instead just make it generic. > > Cc: Michael Ellerman > Cc: Benjamin Herrenschmidt > Cc: Paul Mackerras > Cc: Andrew Morton > Cc: Christoph Hellwig > Cc: linux-i...@vger.kernel.org > Cc: linuxppc-dev@lists.ozlabs.org > Cc: linux...@kvack.org > Cc: linux-ker...@vger.kernel.org > Suggested-by: Christoph Hellwig > Signed-off-by: Anshuman Khandual > --- > This change was originally suggested in an earilier discussion. This > applies on v5.12-rc1 and has been build tested on all applicable > platforms i.e ia64 and powerpc. > > https://patchwork.kernel.org/project/linux-mm/patch/1613024531-19040-3-git-send-email-anshuman.khand...@arm.com/ > > arch/ia64/Kconfig| 6 +- > arch/powerpc/Kconfig | 6 +- > mm/Kconfig | 8 > 3 files changed, 10 insertions(+), 10 deletions(-) > > diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig > index 2ad7a8d29fcc..6b3e3f6c29ae 100644 > --- a/arch/ia64/Kconfig > +++ b/arch/ia64/Kconfig > @@ -32,6 +32,7 @@ config IA64 > select TTY > select HAVE_ARCH_TRACEHOOK > select HAVE_VIRT_CPU_ACCOUNTING > + select HUGETLB_PAGE_SIZE_VARIABLE doesn't this need a 'if HUGETLB_PAGE' or did you verify that HUGETLB_PAGE_SIZE_VARIABLE checks are always nested inside of HUGETLB_PAGE ones?
[PATCH] mm: Generalize HUGETLB_PAGE_SIZE_VARIABLE
HUGETLB_PAGE_SIZE_VARIABLE need not be defined for each individual platform subscribing it. Instead just make it generic. Cc: Michael Ellerman Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Andrew Morton Cc: Christoph Hellwig Cc: linux-i...@vger.kernel.org Cc: linuxppc-dev@lists.ozlabs.org Cc: linux...@kvack.org Cc: linux-ker...@vger.kernel.org Suggested-by: Christoph Hellwig Signed-off-by: Anshuman Khandual --- This change was originally suggested in an earilier discussion. This applies on v5.12-rc1 and has been build tested on all applicable platforms i.e ia64 and powerpc. https://patchwork.kernel.org/project/linux-mm/patch/1613024531-19040-3-git-send-email-anshuman.khand...@arm.com/ arch/ia64/Kconfig| 6 +- arch/powerpc/Kconfig | 6 +- mm/Kconfig | 8 3 files changed, 10 insertions(+), 10 deletions(-) diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig index 2ad7a8d29fcc..6b3e3f6c29ae 100644 --- a/arch/ia64/Kconfig +++ b/arch/ia64/Kconfig @@ -32,6 +32,7 @@ config IA64 select TTY select HAVE_ARCH_TRACEHOOK select HAVE_VIRT_CPU_ACCOUNTING + select HUGETLB_PAGE_SIZE_VARIABLE select VIRT_TO_BUS select GENERIC_IRQ_PROBE select GENERIC_PENDING_IRQ if SMP @@ -82,11 +83,6 @@ config STACKTRACE_SUPPORT config GENERIC_LOCKBREAK def_bool n -config HUGETLB_PAGE_SIZE_VARIABLE - bool - depends on HUGETLB_PAGE - default y - config GENERIC_CALIBRATE_DELAY bool default y diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 3778ad17f56a..b8565bed284f 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -232,6 +232,7 @@ config PPC select HAVE_HARDLOCKUP_DETECTOR_PERFif PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !HAVE_HARDLOCKUP_DETECTOR_ARCH select HAVE_PERF_REGS select HAVE_PERF_USER_STACK_DUMP + select HUGETLB_PAGE_SIZE_VARIABLE if PPC_BOOK3S_64 select MMU_GATHER_RCU_TABLE_FREE select MMU_GATHER_PAGE_SIZE select HAVE_REGS_AND_STACK_ACCESS_API @@ -416,11 +417,6 @@ config HIGHMEM source "kernel/Kconfig.hz" -config HUGETLB_PAGE_SIZE_VARIABLE - bool - depends on HUGETLB_PAGE && PPC_BOOK3S_64 - default y - config MATH_EMULATION bool "Math emulation" depends on 4xx || PPC_8xx || PPC_MPC832x || BOOKE diff --git a/mm/Kconfig b/mm/Kconfig index 24c045b24b95..e604a87862a4 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -274,6 +274,14 @@ config ARCH_ENABLE_HUGEPAGE_MIGRATION config ARCH_ENABLE_THP_MIGRATION bool +config HUGETLB_PAGE_SIZE_VARIABLE + def_bool n + depends on HUGETLB_PAGE + help + When there are multiple HugeTLB sizes available on a platform + and pageblock_order could then be a dynamic value instead of + standard HUGETLB_PAGE_ORDER. + config CONTIG_ALLOC def_bool (MEMORY_ISOLATION && COMPACTION) || CMA -- 2.20.1
[PATCH V2] mm/memtest: Add ARCH_USE_MEMTEST
early_memtest() does not get called from all architectures. Hence enabling CONFIG_MEMTEST and providing a valid memtest=[1..N] kernel command line option might not trigger the memory pattern tests as would be expected in normal circumstances. This situation is misleading. The change here prevents the above mentioned problem after introducing a new config option ARCH_USE_MEMTEST that should be subscribed on platforms that call early_memtest(), in order to enable the config CONFIG_MEMTEST. Conversely CONFIG_MEMTEST cannot be enabled on platforms where it would not be tested anyway. Cc: Russell King Cc: Catalin Marinas Cc: Will Deacon Cc: Thomas Bogendoerfer Cc: Michael Ellerman Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Thomas Gleixner Cc: Ingo Molnar Cc: Chris Zankel Cc: Max Filippov Cc: linux-arm-ker...@lists.infradead.org Cc: linux-m...@vger.kernel.org Cc: linuxppc-dev@lists.ozlabs.org Cc: linux-xte...@linux-xtensa.org Cc: linux...@kvack.org Cc: linux-ker...@vger.kernel.org Reviewed-by: Max Filippov Signed-off-by: Anshuman Khandual --- This patch applies on v5.12-rc1 and has been tested on arm64 platform. But it has been just build tested on all other platforms. Changes in V2: - Added ARCH_USE_MEMTEST in the sorted alphabetical order on platforms Changes in V1: https://patchwork.kernel.org/project/linux-mm/patch/1612498242-31579-1-git-send-email-anshuman.khand...@arm.com/ arch/arm/Kconfig | 1 + arch/arm64/Kconfig | 1 + arch/mips/Kconfig| 1 + arch/powerpc/Kconfig | 1 + arch/x86/Kconfig | 1 + arch/xtensa/Kconfig | 1 + lib/Kconfig.debug| 9 - 7 files changed, 14 insertions(+), 1 deletion(-) diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 853aab5ab327..9ab047d4cd0a 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -33,6 +33,7 @@ config ARM select ARCH_SUPPORTS_ATOMIC_RMW select ARCH_USE_BUILTIN_BSWAP select ARCH_USE_CMPXCHG_LOCKREF + select ARCH_USE_MEMTEST select ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT if MMU select ARCH_WANT_IPC_PARSE_VERSION select ARCH_WANT_LD_ORPHAN_WARN diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 1f212b47a48a..d4fe5118e9c8 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -67,6 +67,7 @@ config ARM64 select ARCH_KEEP_MEMBLOCK select ARCH_USE_CMPXCHG_LOCKREF select ARCH_USE_GNU_PROPERTY + select ARCH_USE_MEMTEST select ARCH_USE_QUEUED_RWLOCKS select ARCH_USE_QUEUED_SPINLOCKS select ARCH_USE_SYM_ANNOTATIONS diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig index d89efba3d8a4..93a4f502f962 100644 --- a/arch/mips/Kconfig +++ b/arch/mips/Kconfig @@ -14,6 +14,7 @@ config MIPS select ARCH_SUPPORTS_UPROBES select ARCH_USE_BUILTIN_BSWAP select ARCH_USE_CMPXCHG_LOCKREF if 64BIT + select ARCH_USE_MEMTEST select ARCH_USE_QUEUED_RWLOCKS select ARCH_USE_QUEUED_SPINLOCKS select ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT if MMU diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 386ae12d8523..3778ad17f56a 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -149,6 +149,7 @@ config PPC select ARCH_SUPPORTS_DEBUG_PAGEALLOCif PPC32 || PPC_BOOK3S_64 select ARCH_USE_BUILTIN_BSWAP select ARCH_USE_CMPXCHG_LOCKREF if PPC64 + select ARCH_USE_MEMTEST select ARCH_USE_QUEUED_RWLOCKS if PPC_QUEUED_SPINLOCKS select ARCH_USE_QUEUED_SPINLOCKSif PPC_QUEUED_SPINLOCKS select ARCH_WANT_IPC_PARSE_VERSION diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 2792879d398e..2cb76fd5258e 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -100,6 +100,7 @@ config X86 select ARCH_SUPPORTS_LTO_CLANG if X86_64 select ARCH_SUPPORTS_LTO_CLANG_THIN if X86_64 select ARCH_USE_BUILTIN_BSWAP + select ARCH_USE_MEMTEST select ARCH_USE_QUEUED_RWLOCKS select ARCH_USE_QUEUED_SPINLOCKS select ARCH_USE_SYM_ANNOTATIONS diff --git a/arch/xtensa/Kconfig b/arch/xtensa/Kconfig index a99dc39f6964..ca51896c53df 100644 --- a/arch/xtensa/Kconfig +++ b/arch/xtensa/Kconfig @@ -7,6 +7,7 @@ config XTENSA select ARCH_HAS_SYNC_DMA_FOR_CPU if MMU select ARCH_HAS_SYNC_DMA_FOR_DEVICE if MMU select ARCH_HAS_DMA_SET_UNCACHED if MMU + select ARCH_USE_MEMTEST select ARCH_USE_QUEUED_RWLOCKS select ARCH_USE_QUEUED_SPINLOCKS select ARCH_WANT_FRAME_POINTERS diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index a2d04c00cda2..2c296535a4b3 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -2521,11 +2521,18 @@ config TEST_FPU endif # RUNTIME_TESTING_MENU +config ARCH_USE_MEMTEST + bool + help + An architecture should select this when it uses early_memtest() + during boot process. + config MEMTEST bool "Memtest" + depends on
[Reoprt] Some compile warning on ppc dts
Hi
After run the following commands
make distclean
make allmodconfig ARCH=powerpc CROSS_COMPILE=powerpc64-linux-gnu-
make oldconfig ARCH=powerpc CROSS_COMPILE=powerpc64-linux-gnu-
make -j64 ARCH=powerpc CROSS_COMPILE=powerpc64-linux-gnu-
I get some warning:
arch/powerpc/boot/dts/mpc5200b.dtsi:267.20-280.4: Warning (pci_bridge):
/pci@fd00: missing ranges for PCI bridg
e (or not a bridge)
arch/powerpc/boot/dts/o2dnt2.dtb: Warning (pci_device_bus_num): Failed
prerequisite 'pci_bridge'
arch/powerpc/boot/dts/mpc5200b.dtsi:182.18-186.5: Warning
(spi_bus_bridge): /soc5200@f000/psc@2000: node name f
or SPI buses should be 'spi'
also defined at arch/powerpc/boot/dts/o2d.dtsi:32.12-43.5
arch/powerpc/boot/dts/o2dnt2.dtb: Warning (spi_bus_reg): Failed
prerequisite 'spi_bus_bridge'
...
For the problem about "node name for SPI buses should be 'spi'":
Rename the psc@2000 to spi@2000 in arch/powerpc/boot/dts/o2d.dtsi can
fix it.
diff --git a/arch/powerpc/boot/dts/o2d.dtsi b/arch/powerpc/boot/dts/o2d.dtsi
index 6661955a2be4..cd3dc70cd72e 100644
--- a/arch/powerpc/boot/dts/o2d.dtsi
+++ b/arch/powerpc/boot/dts/o2d.dtsi
@@ -29,7 +29,7 @@ rtc@800 {
>-->--->---status = "disabled";
>-->---};
-
->-->---psc@2000 {>->---// PSC1
+>-->---spi@2000 {>->---// PSC1
>-->--->---compatible =
"fsl,mpc5200b-psc-spi","fsl,mpc5200-psc-spi";
>-->--->---#address-cells = <1>;
>-->--->---#size-cells = <0>;
---
For the problem about "missing ranges for PCI bridge (or not a bridge)":
Ranges should be add in arch/powerpc/boot/dts/mpc5200b.dtsi.
>---pci: pci@fd00 {
>--->---#interrupt-cells = <1>;
>--->---#size-cells = <2>;
>--->---#address-cells = <3>;
>--->---device_type = "pci";
>--->---compatible = "fsl,mpc5200b-pci","fsl,mpc5200-pci";
>--->---reg = <0xfd00 0x100>;
>--->---// interrupt-map-mask = need to add
>--->---// interrupt-map = need to add
>--->---clock-frequency = <0>; // From boot loader
>--->---interrupts = <2 8 0 2 9 0 2 10 0>;
>--->---bus-range = <0 0>;
>--->---// ranges = need to add
>---};
I think the ranges should be add by someone who knows the mpc5200 better.
Can anyone fix this?
--
Regards
Chen Jun
Re: [PATCH V2] powerpc/perf: Fix handling of privilege level checks in perf interrupt context
> On 26-Feb-2021, at 3:05 PM, Peter Zijlstra wrote: > > On Thu, Feb 25, 2021 at 05:10:39AM -0500, Athira Rajeev wrote: >> diff --git a/arch/powerpc/perf/core-book3s.c >> b/arch/powerpc/perf/core-book3s.c >> index 4b4319d8..c8be44c 100644 >> --- a/arch/powerpc/perf/core-book3s.c >> +++ b/arch/powerpc/perf/core-book3s.c >> @@ -222,7 +222,7 @@ static inline void perf_get_data_addr(struct perf_event >> *event, struct pt_regs * >> if (!(mmcra & MMCRA_SAMPLE_ENABLE) || sdar_valid) >> *addrp = mfspr(SPRN_SDAR); >> >> -if (is_kernel_addr(mfspr(SPRN_SDAR)) && perf_allow_kernel(>attr) >> != 0) >> +if (is_kernel_addr(mfspr(SPRN_SDAR)) && event->attr.exclude_kernel) >> *addrp = 0; >> } >> >> @@ -507,7 +507,7 @@ static void power_pmu_bhrb_read(struct perf_event >> *event, struct cpu_hw_events * >> * addresses, hence include a check before filtering >> code >> */ >> if (!(ppmu->flags & PPMU_ARCH_31) && >> -is_kernel_addr(addr) && >> perf_allow_kernel(>attr) != 0) >> +is_kernel_addr(addr) && event->attr.exclude_kernel) >> continue; >> >> /* Branches are read most recent first (ie. mfbhrb 0 is > > Acked-by: Peter Zijlstra (Intel) Thanks Peter for reviewing the patch. Athira.
Re: [PATCH v3 28/32] powerpc/64s: interrupt implement exit logic in C
Excerpts from Christophe Leroy's message of February 27, 2021 8:07 pm: > > > Le 25/02/2020 à 18:35, Nicholas Piggin a écrit : >> Implement the bulk of interrupt return logic in C. The asm return code >> must handle a few cases: restoring full GPRs, and emulating stack store. >> >> The stack store emulation is significantly simplfied, rather than creating >> a new return frame and switching to that before performing the store, it >> uses the PACA to keep a scratch register around to perform thestore. >> >> The asm return code is moved into 64e for now. The new logic has made >> allowance for 64e, but I don't have a full environment that works well >> to test it, and even booting in emulated qemu is not great for stress >> testing. 64e shouldn't be too far off working with this, given a bit >> more testing and auditing of the logic. >> >> This is slightly faster on a POWER9 (page fault speed increases about >> 1.1%), probably due to reduced mtmsrd. > > > This series, and especially this patch has added a awfull number of BUG_ON() > traps. > > We have an issue open at https://github.com/linuxppc/issues/issues/88 since > 2017 for reducing the > number of BUG_ON()s > > And the kernel Documentation is explicit on the willingness to deprecate > BUG_ON(), see > https://www.kernel.org/doc/html/latest/process/deprecated.html?highlight=bug_on > : > > BUG() and BUG_ON() > Use WARN() and WARN_ON() instead, and handle the “impossible” error condition > as gracefully as > possible. While the BUG()-family of APIs were originally designed to act as > an “impossible > situation” assert and to kill a kernel thread “safely”, they turn out to just > be too risky. (e.g. > “In what order do locks need to be released? Have various states been > restored?”) Very commonly, > using BUG() will destabilize a system or entirely break it, which makes it > impossible to debug or > even get viable crash reports. Linus has very strong feelings about this. > > So ... can we do something cleaner with all the BUG_ON()s recently added ? Yeah you're right. Some of it is probably overkill due to paranoia when developing the series. Now we have a bit more confidence we could probably look at cutting down on these. I do get a bit concerned about detecting a problem in some code like this and attempting to just continue, it usually means the system is going to crash pretty badly anyway (and the WARN_ON trap interrupt is probably going to finish you off anyway). So I think removing the more obvious checks entirely (maybe with a PPC DEBUG config option) is the right way to go. Thanks, Nick
