Re: [PATCH] mm: fix printk format within cma

["Huang, Ying" <ying.huang@intel.com>]

Zhaoyang Huang <huangzhaoyang@gmail.com> writes:

> On Tue, Apr 18, 2023 at 11:38 AM Matthew Wilcox <willy@infradead.org> wrote:
>>
>> On Tue, Apr 18, 2023 at 11:33:09AM +0800, zhaoyang.huang wrote:
>> > cma and page pointer printed via %p are hash value which make debug to be hard.
>> > change them to %px.
>>
>> Why does printing the page pointer make any sense at all?  Surely the
>> PFN makes much more sense.
> either pfn or a correct page pointer makes sense for debugging, while
> page could be more safe than pfn which expose the paddr directly

You can specify "no_hash_pointers" in kernel command line to print
pointer value for debug.  IIUC, this take care of both security and
debuggability.

Best Regards,
Huang, Ying

>>
>> > [63321.482751] [c7] cma: cma_alloc(): memory range at 000000000b5e462c is busy, retrying
>> > [63321.482786] [c7] cma: cma_alloc(): memory range at 000000000f7d6fae is busy, retrying
>> > [63321.482823] [c7] cma: cma_alloc(): memory range at 00000000e653b59b is busy, retrying
>> > [63322.378890] [c7] cma: cma_release(page 00000000dd53cf48)
>> > [63322.378913] [c7] cma: cma_release(page 00000000315f703d)
>> > [63322.378925] [c7] cma: cma_release(page 00000000791e3a5f)
>> >
>> > Signed-off-by: Zhaoyang Huang <zhaoyang.huang@unisoc.com>
>> > ---
>> >  mm/cma.c | 6 +++---
>> >  1 file changed, 3 insertions(+), 3 deletions(-)
>> >
>> > diff --git a/mm/cma.c b/mm/cma.c
>> > index 4a978e0..dfe9813 100644
>> > --- a/mm/cma.c
>> > +++ b/mm/cma.c
>> > @@ -435,7 +435,7 @@ struct page *cma_alloc(struct cma *cma, unsigned long count,
>> >       if (!cma || !cma->count || !cma->bitmap)
>> >               goto out;
>> >
>> > -     pr_debug("%s(cma %p, count %lu, align %d)\n", __func__, (void *)cma,
>> > +     pr_debug("%s(cma %px, count %lu, align %d)\n", __func__, (void *)cma,
>> >                count, align);
>> >
>> >       if (!count)
>> > @@ -534,7 +534,7 @@ bool cma_pages_valid(struct cma *cma, const struct page *pages,
>> >       pfn = page_to_pfn(pages);
>> >
>> >       if (pfn < cma->base_pfn || pfn >= cma->base_pfn + cma->count) {
>> > -             pr_debug("%s(page %p, count %lu)\n", __func__,
>> > +             pr_debug("%s(page %px, count %lu)\n", __func__,
>> >                                               (void *)pages, count);
>> >               return false;
>> >       }
>> > @@ -560,7 +560,7 @@ bool cma_release(struct cma *cma, const struct page *pages,
>> >       if (!cma_pages_valid(cma, pages, count))
>> >               return false;
>> >
>> > -     pr_debug("%s(page %p, count %lu)\n", __func__, (void *)pages, count);
>> > +     pr_debug("%s(page %px, count %lu)\n", __func__, (void *)pages, count);
>> >
>> >       pfn = page_to_pfn(pages);
>> >
>> > --
>> > 1.9.1
>> >
>> >