After recent pinctl patches we see a warning when booting davinci
due to a bad memory allocation:
------------[ cut here ]------------
WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:3511 __alloc_pages_nodemask+0x16c/0xb18
Modules linked in:
CPU: 0 PID: 1 Comm: swapper Not tainted 4.9.0-rc2-00023-g22d5127-dirty #1019
Hardware name: Generic DA850/OMAP-L138/AM18x
Backtrace:
[<c000d670>] (dump_backtrace) from [<c000d794>] (show_stack+0x18/0x1c)
[<c000d77c>] (show_stack) from [<c021a0d0>] (dump_stack+0x20/0x28)
[<c021a0b0>] (dump_stack) from [<c001bb10>] (__warn+0xe8/0x100)
[<c001ba28>] (__warn) from [<c001bb50>] (warn_slowpath_null+0x28/0x30)
[<c001bb28>] (warn_slowpath_null) from [<c0097e7c>] (__alloc_pages_nodemask+0x16c/0xb18)
[<c0097d10>] (__alloc_pages_nodemask) from [<c00afef4>] (kmalloc_order+0x20/0x58)
[<c00afed4>] (kmalloc_order) from [<c00ce7ac>] (__kmalloc_track_caller+0x188/0x190)
[<c00ce624>] (__kmalloc_track_caller) from [<c02a762c>] (devm_kmalloc+0x24/0x70)
[<c02a7608>] (devm_kmalloc) from [<c0247d10>] (pcs_dt_node_to_map+0x1d0/0xa40)
[<c0245ec8>] (pinctrl_dt_to_map) from [<c0242fd0>] (pinctrl_get+0xe8/0x484)
[snip]
This series fixes this error.
Changes form v1 -> v2
* Add an error message, and correct also other places where the
issue is seen
* Add patch to parse for bits instead of pins
Axel Haslam (2):
pinctrl: single: check for any error when getting rows
pinctrl: single: search for the bits property when parsing bits
drivers/pinctrl/pinctrl-single.c | 14 +++++++++-----
1 file changed, 9 insertions(+), 5 deletions(-)
--
2.10.1
pinctrl_count_index_with_args returns -ENOENT not
-EINVAL. The return check would pass, and we would
try to kzalloc with a negative error size throwing
a warning.
Instead of checking for -EINVAL specifically, lets
check for any error and avoid negative size allocations.
Signed-off-by: Axel Haslam <[email protected]>
---
drivers/pinctrl/pinctrl-single.c | 12 ++++++++----
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/drivers/pinctrl/pinctrl-single.c b/drivers/pinctrl/pinctrl-single.c
index 539f31c..f36a9f1 100644
--- a/drivers/pinctrl/pinctrl-single.c
+++ b/drivers/pinctrl/pinctrl-single.c
@@ -1133,8 +1133,10 @@ static int pcs_parse_one_pinctrl_entry(struct pcs_device *pcs,
struct pcs_function *function;
rows = pinctrl_count_index_with_args(np, name);
- if (rows == -EINVAL)
- return rows;
+ if (rows <= 0) {
+ dev_err(pcs->dev, "Ivalid number of rows: %d\n", rows);
+ return -EINVAL;
+ }
vals = devm_kzalloc(pcs->dev, sizeof(*vals) * rows, GFP_KERNEL);
if (!vals)
@@ -1228,8 +1230,10 @@ static int pcs_parse_bits_in_pinctrl_entry(struct pcs_device *pcs,
struct pcs_function *function;
rows = pinctrl_count_index_with_args(np, name);
- if (rows == -EINVAL)
- return rows;
+ if (rows <= 0) {
+ dev_err(pcs->dev, "Invalid number of rows: %d\n", rows);
+ return -EINVAL;
+ }
npins_in_row = pcs->width / pcs->bits_per_pin;
--
2.10.1
The pcs_parse_bits_in_pinctrl_entry function should search
for the "pinctrl-single,bits" and not "pinctrl-single,pins"
Signed-off-by: Axel Haslam <[email protected]>
---
drivers/pinctrl/pinctrl-single.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/pinctrl/pinctrl-single.c b/drivers/pinctrl/pinctrl-single.c
index f36a9f1..2b196e5 100644
--- a/drivers/pinctrl/pinctrl-single.c
+++ b/drivers/pinctrl/pinctrl-single.c
@@ -1223,7 +1223,7 @@ static int pcs_parse_bits_in_pinctrl_entry(struct pcs_device *pcs,
unsigned *num_maps,
const char **pgnames)
{
- const char *name = "pinctrl-single,pins";
+ const char *name = "pinctrl-single,bits";
struct pcs_func_vals *vals;
int rows, *pins, found = 0, res = -ENOMEM, i;
int npins_in_row;
--
2.10.1
* Axel Haslam <[email protected]> [161109 07:54]:
> After recent pinctl patches we see a warning when booting davinci
> due to a bad memory allocation:
>
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:3511 __alloc_pages_nodemask+0x16c/0xb18
> Modules linked in:
> CPU: 0 PID: 1 Comm: swapper Not tainted 4.9.0-rc2-00023-g22d5127-dirty #1019
> Hardware name: Generic DA850/OMAP-L138/AM18x
> Backtrace:
> [<c000d670>] (dump_backtrace) from [<c000d794>] (show_stack+0x18/0x1c)
> [<c000d77c>] (show_stack) from [<c021a0d0>] (dump_stack+0x20/0x28)
> [<c021a0b0>] (dump_stack) from [<c001bb10>] (__warn+0xe8/0x100)
> [<c001ba28>] (__warn) from [<c001bb50>] (warn_slowpath_null+0x28/0x30)
> [<c001bb28>] (warn_slowpath_null) from [<c0097e7c>] (__alloc_pages_nodemask+0x16c/0xb18)
> [<c0097d10>] (__alloc_pages_nodemask) from [<c00afef4>] (kmalloc_order+0x20/0x58)
> [<c00afed4>] (kmalloc_order) from [<c00ce7ac>] (__kmalloc_track_caller+0x188/0x190)
> [<c00ce624>] (__kmalloc_track_caller) from [<c02a762c>] (devm_kmalloc+0x24/0x70)
> [<c02a7608>] (devm_kmalloc) from [<c0247d10>] (pcs_dt_node_to_map+0x1d0/0xa40)
> [<c0245ec8>] (pinctrl_dt_to_map) from [<c0242fd0>] (pinctrl_get+0xe8/0x484)
> [snip]
>
> This series fixes this error.
Thanks for fixing these and sorry about breaking pinctrl-bits:
Acked-by: Tony Lindgren <[email protected]>
On Wed, Nov 9, 2016 at 3:54 PM, Axel Haslam <[email protected]> wrote:
> pinctrl_count_index_with_args returns -ENOENT not
> -EINVAL. The return check would pass, and we would
> try to kzalloc with a negative error size throwing
> a warning.
>
> Instead of checking for -EINVAL specifically, lets
> check for any error and avoid negative size allocations.
>
> Signed-off-by: Axel Haslam <[email protected]>
Patch applied with Tony's ACK.
Yours,
Linus Walleij
On Wed, Nov 9, 2016 at 3:54 PM, Axel Haslam <[email protected]> wrote:
> The pcs_parse_bits_in_pinctrl_entry function should search
> for the "pinctrl-single,bits" and not "pinctrl-single,pins"
>
> Signed-off-by: Axel Haslam <[email protected]>
Patch applied with Tony's ACK.
Yours,
Linus Walleij