2017-07-07 23:53:31

by Kani, Toshimitsu

Subject: [PATCH] libnvdimm: fix badblock range handling of ARS range

__add_badblock_range() does not account sector alignment when
it sets 'num_sectors'. Therefore, an ARS error record range
spanning across two sectors is set to a single sector length,
which leaves the 2nd sector unprotected.

Change __add_badblock_range() to set 'num_sectors' properly.

Fixes: 0caeef63e6d2f866d85bb507bf63e0ce8ec91cef
Signed-off-by: Toshi Kani <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Vishal Verma <[email protected]>
Cc: <[email protected]>
---
drivers/nvdimm/core.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/nvdimm/core.c b/drivers/nvdimm/core.c
index 2dee908..932c3994 100644
--- a/drivers/nvdimm/core.c
+++ b/drivers/nvdimm/core.c
@@ -421,14 +421,15 @@ static void set_badblock(struct badblocks *bb, sector_t s, int num)
static void __add_badblock_range(struct badblocks *bb, u64 ns_offset, u64 len)
{
const unsigned int sector_size = 512;
- sector_t start_sector;
+ sector_t start_sector, end_sector;
u64 num_sectors;
u32 rem;

start_sector = div_u64(ns_offset, sector_size);
- num_sectors = div_u64_rem(len, sector_size, &rem);
+ end_sector = div_u64_rem(ns_offset + len, sector_size, &rem);
if (rem)
- num_sectors++;
+ end_sector++;
+ num_sectors = end_sector - start_sector;

if (unlikely(num_sectors > (u64)INT_MAX)) {
u64 remaining = num_sectors;

2017-07-10 21:32:56

by Verma, Vishal L

Subject: Re: [PATCH] libnvdimm: fix badblock range handling of ARS range

On 07/07, Toshi Kani wrote:
> __add_badblock_range() does not account sector alignment when
> it sets 'num_sectors'. Therefore, an ARS error record range
> spanning across two sectors is set to a single sector length,
> which leaves the 2nd sector unprotected.
>
> Change __add_badblock_range() to set 'num_sectors' properly.
>
> Fixes: 0caeef63e6d2f866d85bb507bf63e0ce8ec91cef
> Signed-off-by: Toshi Kani <[email protected]>
> Cc: Dan Williams <[email protected]>
> Cc: Vishal Verma <[email protected]>
> Cc: <[email protected]>
> ---
> drivers/nvdimm/core.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)

Good find! I don't know why I assumed a poison entry spanning past the
end of a sector wouldn't span past multiple sectors :)

Reviewed-by: Vishal Verma <[email protected]>

>
> diff --git a/drivers/nvdimm/core.c b/drivers/nvdimm/core.c
> index 2dee908..932c3994 100644
> --- a/drivers/nvdimm/core.c
> +++ b/drivers/nvdimm/core.c
> @@ -421,14 +421,15 @@ static void set_badblock(struct badblocks *bb, sector_t s, int num)
> static void __add_badblock_range(struct badblocks *bb, u64 ns_offset, u64 len)
> {
> const unsigned int sector_size = 512;
> - sector_t start_sector;
> + sector_t start_sector, end_sector;
> u64 num_sectors;
> u32 rem;
>
> start_sector = div_u64(ns_offset, sector_size);
> - num_sectors = div_u64_rem(len, sector_size, &rem);
> + end_sector = div_u64_rem(ns_offset + len, sector_size, &rem);
> if (rem)
> - num_sectors++;
> + end_sector++;
> + num_sectors = end_sector - start_sector;
>
> if (unlikely(num_sectors > (u64)INT_MAX)) {
> u64 remaining = num_sectors;