Use huge_pages_needed for private hugetlb mappings
Subsystems (2)
RFC 1/1 Use huge_pages_needed for private hugetlb mappings
[RFC] Use huge_pages_needed algorithm for MAP_PRIVATE mappings
Since Ken Chen and David Gibson's strict reservation patch can handle only shared mappings, restore the huge_pages_needed function and use it for private mappings.
* Evil Note *
This patch depends on the following two patches:
Ken Chen - [patch] hugetlb strict commit accounting - v3 David Gibson - hugepage: Serialize hugepage allocation and instantiation
Signed-off-by Adam Litke <[email protected]>
--- inode.c | 46 +++++++++++++++++++++++++++++++++++++++++++++- 1 files changed, 45 insertions(+), 1 deletion(-)
fs/hugetlbfs/inode.c +45 -1
@@ -56,9 +56,48 @@ static void huge_pagevec_release(structpagevec_reinit (pvec); } +/* +* huge_pages_needed tries to determine the number of new huge pages that +* will be required to fully populate this VMA. This will be equal to +* the size of the VMA in huge pages minus the number of huge pages +* (covered by this VMA) that are found in the page cache. +* +* Result is in bytes to be compatible with is_hugepage_mem_enough() +*/ +static unsigned long +huge_pages_needed (struct address_space *mapping,struct vm_area_struct *vma) +{ +int i; +struct pagevec pvec; +unsigned long start = vma->vm_start; +unsigned long end = vma->vm_end; +unsigned long hugepages = (end - start) >>HPAGE_SHIFT ; +pgoff_t next = vma->vm_pgoff >> (HPAGE_SHIFT -PAGE_SHIFT ); +pgoff_t endpg = next + hugepages; + +pagevec_init (&pvec,0 ); +while (next < endpg) { +if (!pagevec_lookup (&pvec, mapping, next,PAGEVEC_SIZE )) +break ; +for (i =0 ; i <pagevec_count (&pvec); i++) { +struct page *page = pvec.pages[i]; +if (page->index > next) + next = page->index; +if (page->index >= endpg) +break ; + next++; + hugepages--; + } +huge_pagevec_release (&pvec); + } +return hugepages <<HPAGE_SHIFT ; +} +static int hugetlbfs_file_mmap (struct file *file,struct vm_area_struct *vma) {struct inode *inode = file->f_dentry->d_inode; +struct address_space *mapping = inode->i_mapping; +unsigned long bytes;loff_t len, vma_len;int ret;
@@ -86,9 +125,14 @@ static int hugetlbfs_file_mmap(struct fiif (!(vma->vm_flags &VM_WRITE ) && len > inode->i_size) goto out; -if (vma->vm_flags &VM_MAYSHARE ) +if (vma->vm_flags &VM_MAYSHARE ) {if (hugetlb_reserve_pages (inode, vma)) goto out; + }else { + bytes =huge_pages_needed (mapping, vma); +if (!is_hugepage_mem_enough (bytes)) + goto out; + } ret =0 ;hugetlb_prefault_arch_hook (vma->vm_mm);
Signature · 8 lines
Adam Litke - (agl at us.ibm.com) IBM Linux Technology Center -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>