* Re: [PATCH] [dma-coherent] Fix integer overflow in the reserved-memory dma allocation
2020-03-09 13:56 ` Robin Murphy
@ 2020-03-09 14:45 ` Kevin Grandemange
2020-03-09 15:22 ` [PATCH v2] " Kevin Grandemange
2020-03-12 15:41 ` [PATCH v3] " Kevin Grandemange
2 siblings, 0 replies; 6+ messages in thread
From: Kevin Grandemange @ 2020-03-09 14:45 UTC (permalink / raw)
To: Robin Murphy, Christoph Hellwig; +Cc: iommu
On 3/9/20 2:56 PM, Robin Murphy wrote:
> On 09/03/2020 11:02 am, Kevin Grandemange wrote:
>> pageno is an int and the PAGE_SHIFT shift is done on an int,
>> overflowing if the memory is bigger than 2G
>>
>> This can be reproduced using for example a reserved-memory of 4G
>
> Nit: the example shows 16GB, not 4.
>
Will fix. My usecase is an external memory that can span 16GB / 4GB depending on the configuration.
>> reserved-memory {
>> #address-cells = <2>;
>> #size-cells = <2>;
>> ranges;
>>
>> reserved_dma: buffer@0 {
>> compatible = "shared-dma-pool";
>> no-map;
>> reg = <0x5 0x00000000 0x4 0x0>;
>> };
>> };
>>
>> Signed-off-by: Kevin Grandemange <kevin.grandemange@allegrodvt.com>
>> ---
>> kernel/dma/coherent.c | 6 ++++--
>> 1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
>> index 551b0eb7028a..c20c6c0635b7 100644
>> --- a/kernel/dma/coherent.c
>> +++ b/kernel/dma/coherent.c
>> @@ -130,6 +130,7 @@ static void *__dma_alloc_from_coherent(struct device *dev,
>> int order = get_order(size);
>> unsigned long flags;
>> int pageno;
>> + ssize_t mem_offset;
>> void *ret;
>> spin_lock_irqsave(&mem->spinlock, flags);
>> @@ -144,8 +145,9 @@ static void *__dma_alloc_from_coherent(struct device *dev,
>> /*
>> * Memory was found in the coherent area.
>> */
>> - *dma_handle = dma_get_device_base(dev, mem) + (pageno << PAGE_SHIFT);
>> - ret = mem->virt_base + (pageno << PAGE_SHIFT);
>> + mem_offset = (ssize_t)pageno << PAGE_SHIFT;
>
> Isn't this still capable of overflowing for 32-bit architectures where ssize_t is most likely to be long and LONG_MAX == INT_MAX (before we even get to PAE and friends)? Logically, dma_addr_t would be the most appropriate type here.
You are right, I was only thinking in my usecase (arm64) and wrongly assuming ssize_t had the same properties as size_t.
I will change that to dma_addr_t.
>
> FWIW since you have to have an explicit cast either way, I don't see much point in introducing the local variable vs. just adding the cast into the existing expression.
The local variable was just for checkpatch.pl line length check, I agree with you, it doesn't add much.
>
> Thanks,
> Robin.
>
>> + *dma_handle = dma_get_device_base(dev, mem) + mem_offset;
>> + ret = mem->virt_base + mem_offset;
>> spin_unlock_irqrestore(&mem->spinlock, flags);
>> memset(ret, 0, size);
>> return ret;
>>
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v2] [dma-coherent] Fix integer overflow in the reserved-memory dma allocation
2020-03-09 13:56 ` Robin Murphy
2020-03-09 14:45 ` Kevin Grandemange
@ 2020-03-09 15:22 ` Kevin Grandemange
2020-03-12 15:41 ` [PATCH v3] " Kevin Grandemange
2 siblings, 0 replies; 6+ messages in thread
From: Kevin Grandemange @ 2020-03-09 15:22 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: iommu, Kevin Grandemange, Robin Murphy
pageno is an int and the PAGE_SHIFT shift is done on an int,
overflowing if the memory is bigger than 2G
This can be reproduced using for example a reserved-memory of 4G
reserved-memory {
#address-cells = <2>;
#size-cells = <2>;
ranges;
reserved_dma: buffer@0 {
compatible = "shared-dma-pool";
no-map;
reg = <0x5 0x00000000 0x1 0x0>;
};
};
Signed-off-by: Kevin Grandemange <kevin.grandemange@allegrodvt.com>
---
Changes v1 -> v2:
- removed mem_offset tmp variable
- use dma_addr_t instead of ssize_t
- Fix reserved-memory size in the dts example
kernel/dma/coherent.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
index 551b0eb7028a..b8e4fa441493 100644
--- a/kernel/dma/coherent.c
+++ b/kernel/dma/coherent.c
@@ -144,8 +144,8 @@ static void *__dma_alloc_from_coherent(struct device *dev,
/*
* Memory was found in the coherent area.
*/
- *dma_handle = dma_get_device_base(dev, mem) + (pageno << PAGE_SHIFT);
- ret = mem->virt_base + (pageno << PAGE_SHIFT);
+ *dma_handle = dma_get_device_base(dev, mem) + ((dma_addr_t)pageno << PAGE_SHIFT);
+ ret = mem->virt_base + ((dma_addr_t)pageno << PAGE_SHIFT);
spin_unlock_irqrestore(&mem->spinlock, flags);
memset(ret, 0, size);
return ret;
--
2.20.1
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v3] [dma-coherent] Fix integer overflow in the reserved-memory dma allocation
2020-03-09 13:56 ` Robin Murphy
2020-03-09 14:45 ` Kevin Grandemange
2020-03-09 15:22 ` [PATCH v2] " Kevin Grandemange
@ 2020-03-12 15:41 ` Kevin Grandemange
2020-03-16 9:43 ` Christoph Hellwig
2 siblings, 1 reply; 6+ messages in thread
From: Kevin Grandemange @ 2020-03-12 15:41 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: iommu, Kevin Grandemange, Robin Murphy
pageno is an int and the PAGE_SHIFT shift is done on an int,
overflowing if the memory is bigger than 2G
This can be reproduced using for example a reserved-memory of 4G
reserved-memory {
#address-cells = <2>;
#size-cells = <2>;
ranges;
reserved_dma: buffer@0 {
compatible = "shared-dma-pool";
no-map;
reg = <0x5 0x00000000 0x1 0x0>;
};
};
Signed-off-by: Kevin Grandemange <kevin.grandemange@allegrodvt.com>
---
Changes v1 -> v2:
- removed mem_offset tmp variable
- use dma_addr_t instead of ssize_t
- Fix reserved-memory size in the dts example
Changes v2 -> v3:
- Fix several other site where PAGE_SHIFT shifts are done on ints.
kernel/dma/coherent.c | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
index 551b0eb7028a..d322cb786e7e 100644
--- a/kernel/dma/coherent.c
+++ b/kernel/dma/coherent.c
@@ -134,7 +134,7 @@ static void *__dma_alloc_from_coherent(struct device *dev,
spin_lock_irqsave(&mem->spinlock, flags);
- if (unlikely(size > (mem->size << PAGE_SHIFT)))
+ if (unlikely(size > ((dma_addr_t)mem->size << PAGE_SHIFT)))
goto err;
pageno = bitmap_find_free_region(mem->bitmap, mem->size, order);
@@ -144,8 +144,8 @@ static void *__dma_alloc_from_coherent(struct device *dev,
/*
* Memory was found in the coherent area.
*/
- *dma_handle = dma_get_device_base(dev, mem) + (pageno << PAGE_SHIFT);
- ret = mem->virt_base + (pageno << PAGE_SHIFT);
+ *dma_handle = dma_get_device_base(dev, mem) + ((dma_addr_t)pageno << PAGE_SHIFT);
+ ret = mem->virt_base + ((dma_addr_t)pageno << PAGE_SHIFT);
spin_unlock_irqrestore(&mem->spinlock, flags);
memset(ret, 0, size);
return ret;
@@ -194,7 +194,7 @@ static int __dma_release_from_coherent(struct dma_coherent_mem *mem,
int order, void *vaddr)
{
if (mem && vaddr >= mem->virt_base && vaddr <
- (mem->virt_base + (mem->size << PAGE_SHIFT))) {
+ (mem->virt_base + ((dma_addr_t)mem->size << PAGE_SHIFT))) {
int page = (vaddr - mem->virt_base) >> PAGE_SHIFT;
unsigned long flags;
@@ -238,7 +238,7 @@ static int __dma_mmap_from_coherent(struct dma_coherent_mem *mem,
struct vm_area_struct *vma, void *vaddr, size_t size, int *ret)
{
if (mem && vaddr >= mem->virt_base && vaddr + size <=
- (mem->virt_base + (mem->size << PAGE_SHIFT))) {
+ (mem->virt_base + ((dma_addr_t)mem->size << PAGE_SHIFT))) {
unsigned long off = vma->vm_pgoff;
int start = (vaddr - mem->virt_base) >> PAGE_SHIFT;
int user_count = vma_pages(vma);
@@ -248,7 +248,7 @@ static int __dma_mmap_from_coherent(struct dma_coherent_mem *mem,
if (off < count && user_count <= count - off) {
unsigned long pfn = mem->pfn_base + start + off;
*ret = remap_pfn_range(vma, vma->vm_start, pfn,
- user_count << PAGE_SHIFT,
+ (unsigned long)user_count << PAGE_SHIFT,
vma->vm_page_prot);
}
return 1;
--
2.20.1
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu
^ permalink raw reply related [flat|nested] 6+ messages in thread