Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

nv_peer_mem: Fix SGls #75

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

nv_peer_mem: Fix SGls #75

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
47 changes: 38 additions & 9 deletions nv_peer_mem.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -122,6 +122,8 @@ struct nv_mem_context {
unsigned long page_size;
int is_callback;
int sg_allocated;
unsigned long addr;
size_t size;
};


Expand Down Expand Up @@ -211,6 +213,8 @@ static int nv_mem_acquire(unsigned long addr, size_t size, void *peer_mem_privat
nv_mem_context->page_virt_start = addr & GPU_PAGE_MASK;
nv_mem_context->page_virt_end = (addr + size + GPU_PAGE_SIZE - 1) & GPU_PAGE_MASK;
nv_mem_context->mapped_size = nv_mem_context->page_virt_end - nv_mem_context->page_virt_start;
nv_mem_context->size = size;
nv_mem_context->addr = addr;

ret = nvidia_p2p_get_pages(0, 0, nv_mem_context->page_virt_start, nv_mem_context->mapped_size,
&nv_mem_context->page_table, nv_mem_dummy_callback, nv_mem_context);
Expand Down Expand Up @@ -249,13 +253,20 @@ static int nv_dma_map(struct sg_table *sg_head, void *context,
struct nv_mem_context *nv_mem_context =
(struct nv_mem_context *) context;
struct nvidia_p2p_page_table *page_table = nv_mem_context->page_table;
unsigned long page_size;
unsigned long system_page_virt_start, system_page_virt_end;
unsigned int system_page_offset;

if (page_table->page_size != NVIDIA_P2P_PAGE_SIZE_64KB) {
peer_err("nv_dma_map -- assumption of 64KB pages failed size_id=%u\n",
nv_mem_context->page_table->page_size);
return -EINVAL;
}

system_page_virt_start = nv_mem_context->addr & PAGE_MASK;
system_page_virt_end = (nv_mem_context->addr + nv_mem_context->size + PAGE_SIZE - 1) & PAGE_MASK;
system_page_offset = system_page_virt_start - nv_mem_context->page_virt_start;

#if NV_DMA_MAPPING
{
struct nvidia_p2p_dma_mapping *dma_mapping;
Expand Down Expand Up @@ -289,10 +300,21 @@ static int nv_dma_map(struct sg_table *sg_head, void *context,

nv_mem_context->dma_mapping = dma_mapping;
nv_mem_context->sg_allocated = 1;
BUILD_BUG_ON(PAGE_SIZE > GPU_PAGE_SIZE);
for_each_sg(sg_head->sgl, sg, nv_mem_context->npages, i) {
sg_set_page(sg, NULL, nv_mem_context->page_size, 0);
sg->dma_address = dma_mapping->dma_addresses[i];
sg->dma_length = nv_mem_context->page_size;
page_size = nv_mem_context->page_size;
if (i == nv_mem_context->npages - 1)
page_size = ((system_page_virt_end - system_page_virt_start +
Copy link

@drossetti drossetti Aug 20, 2020 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This can be either system page size (4KB x86, 64KB POWER) or 64KB (GPU BAR1), right?
a comment would be useful

Also, that does not work if system page size > GPU_PAGE_SIZE, hypothetically. A BUG_ON/static assert would be useful.

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

And don't forget on ppc64le, sometimes the system is using HUGE pages which is configurable (default 2MB on RHEL7/8, but should always be queried).

system_page_offset - 1) & GPU_PAGE_OFFSET) + 1;

if (i == 0) {
sg_head->sgl[0].dma_address += system_page_offset;
page_size -= system_page_offset;
}

sg_set_page(sg, NULL, page_size, 0);
sg->dma_length = page_size;
}
}
#else
Expand All @@ -311,10 +333,21 @@ static int nv_dma_map(struct sg_table *sg_head, void *context,
return ret;

nv_mem_context->sg_allocated = 1;
BUILD_BUG_ON(PAGE_SIZE > GPU_PAGE_SIZE);
for_each_sg(sg_head->sgl, sg, nv_mem_context->npages, i) {
sg_set_page(sg, NULL, nv_mem_context->page_size, 0);
sg->dma_address = page_table->pages[i]->physical_address;
sg->dma_length = nv_mem_context->page_size;
page_size = nv_mem_context->page_size;
if (i == nv_mem_context->npages - 1)
page_size = ((system_page_virt_end - system_page_virt_start +
system_page_offset - 1) & GPU_PAGE_OFFSET) + 1;

if (i == 0) {
sg_head->sgl[0].dma_address += system_page_offset;
page_size -= system_page_offset;
}

sg_set_page(sg, NULL, page_size, 0);
sg->dma_length = page_size;
}

#endif
Expand Down Expand Up @@ -431,11 +464,7 @@ static int nv_mem_get_pages(unsigned long addr,

static unsigned long nv_mem_get_page_size(void *context)
{
struct nv_mem_context *nv_mem_context =
(struct nv_mem_context *)context;

return nv_mem_context->page_size;

return PAGE_SIZE;
}


Expand Down