Message ID | 20210316100433.17665-1-colin.king@canonical.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | mm: cma: Fix potential null dereference on pointer cma | expand |
On Tue, Mar 16, 2021 at 10:04:33AM +0000, Colin King wrote: > From: Colin Ian King <colin.king@canonical.com> > > At the start of the function there is a null pointer check on cma > and then branch to error handling label 'out'. The subsequent call > to cma_sysfs_fail_pages_count dereferences cma, hence there is a > potential null pointer deference issue. Fix this by only calling > cma_sysfs_fail_pages_count if cma is not null. > > Addresses-Coverity: ("Dereference after null check") > Fixes: dc4764f7e9ac ("mm: cma: support sysfs") > Signed-off-by: Colin Ian King <colin.king@canonical.com> Acked-by: Minchan Kim <minchan@kernel.org>
On 3/16/21 3:04 AM, Colin King wrote: > From: Colin Ian King <colin.king@canonical.com> > > At the start of the function there is a null pointer check on cma > and then branch to error handling label 'out'. The subsequent call > to cma_sysfs_fail_pages_count dereferences cma, hence there is a > potential null pointer deference issue. Fix this by only calling As far as I can tell, it's not possible to actually cause that null failure with the existing kernel code paths. *Might* be worth mentioning that here (unless I'm wrong), but either way, looks good, so: Reviewed-by: John Hubbard <jhubbard@nvidia.com> thanks,
diff --git a/mm/cma.c b/mm/cma.c index 6d4dbafdb318..90e27458ddb7 100644 --- a/mm/cma.c +++ b/mm/cma.c @@ -512,7 +512,8 @@ struct page *cma_alloc(struct cma *cma, size_t count, unsigned int align, cma_sysfs_alloc_pages_count(cma, count); } else { count_vm_event(CMA_ALLOC_FAIL); - cma_sysfs_fail_pages_count(cma, count); + if (cma) + cma_sysfs_fail_pages_count(cma, count); } return page;