* [PATCH] clk: visconti: prevent array overflow in visconti_clk_register_gates()
@ 2022-03-16 8:35 Dan Carpenter
2022-03-17 5:19 ` nobuhiro1.iwamatsu
2022-03-17 19:21 ` Stephen Boyd
0 siblings, 2 replies; 3+ messages in thread
From: Dan Carpenter @ 2022-03-16 8:35 UTC (permalink / raw)
To: Michael Turquette, Nobuhiro Iwamatsu
Cc: Stephen Boyd, linux-clk, kernel-janitors
This code was using -1 to represent that there was no reset function.
Unfortunately, the -1 was stored in u8 so the if (clks[i].rs_id >= 0)
condition was always true. This lead to an out of bounds access in
visconti_clk_register_gates().
Fixes: b4cbe606dc36 ("clk: visconti: Add support common clock driver and reset driver")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
drivers/clk/visconti/clkc.h | 3 +++
drivers/clk/visconti/clkc-tmpv770x.c | 2 +-
drivers/clk/visconti/clkc.c | 2 +-
3 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/drivers/clk/visconti/clkc.h b/drivers/clk/visconti/clkc.h
index 09ed82ff64e4..8756a1ec42ef 100644
--- a/drivers/clk/visconti/clkc.h
+++ b/drivers/clk/visconti/clkc.h
@@ -73,4 +73,7 @@ int visconti_clk_register_gates(struct visconti_clk_provider *data,
int num_gate,
const struct visconti_reset_data *reset,
spinlock_t *lock);
+
+#define NO_RESET 0xFF
+
#endif /* _VISCONTI_CLKC_H_ */
diff --git a/drivers/clk/visconti/clkc-tmpv770x.c b/drivers/clk/visconti/clkc-tmpv770x.c
index c2b2f41a85a4..6c753b2cb558 100644
--- a/drivers/clk/visconti/clkc-tmpv770x.c
+++ b/drivers/clk/visconti/clkc-tmpv770x.c
@@ -176,7 +176,7 @@ static const struct visconti_clk_gate_table clk_gate_tables[] = {
{ TMPV770X_CLK_WRCK, "wrck",
clks_parent_data, ARRAY_SIZE(clks_parent_data),
0, 0x68, 0x168, 9, 32,
- -1, }, /* No reset */
+ NO_RESET, },
{ TMPV770X_CLK_PICKMON, "pickmon",
clks_parent_data, ARRAY_SIZE(clks_parent_data),
0, 0x10, 0x110, 8, 4,
diff --git a/drivers/clk/visconti/clkc.c b/drivers/clk/visconti/clkc.c
index 56a8a4ffebca..d0b193b5d0b3 100644
--- a/drivers/clk/visconti/clkc.c
+++ b/drivers/clk/visconti/clkc.c
@@ -147,7 +147,7 @@ int visconti_clk_register_gates(struct visconti_clk_provider *ctx,
if (!dev_name)
return -ENOMEM;
- if (clks[i].rs_id >= 0) {
+ if (clks[i].rs_id != NO_RESET) {
rson_offset = reset[clks[i].rs_id].rson_offset;
rsoff_offset = reset[clks[i].rs_id].rsoff_offset;
rs_idx = reset[clks[i].rs_id].rs_idx;
--
2.20.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* RE: [PATCH] clk: visconti: prevent array overflow in visconti_clk_register_gates()
2022-03-16 8:35 [PATCH] clk: visconti: prevent array overflow in visconti_clk_register_gates() Dan Carpenter
@ 2022-03-17 5:19 ` nobuhiro1.iwamatsu
2022-03-17 19:21 ` Stephen Boyd
1 sibling, 0 replies; 3+ messages in thread
From: nobuhiro1.iwamatsu @ 2022-03-17 5:19 UTC (permalink / raw)
To: dan.carpenter, mturquette; +Cc: sboyd, linux-clk, kernel-janitors
Hi Dan,
Thanks for your patch.
And I forgot to update and fix on the report you sent me before, sorry.
> -----Original Message-----
> From: Dan Carpenter <dan.carpenter@oracle.com>
> Sent: Wednesday, March 16, 2022 5:36 PM
> To: Michael Turquette <mturquette@baylibre.com>; iwamatsu nobuhiro(岩松
> 信洋 □SWC◯ACT) <nobuhiro1.iwamatsu@toshiba.co.jp>
> Cc: Stephen Boyd <sboyd@kernel.org>; linux-clk@vger.kernel.org;
> kernel-janitors@vger.kernel.org
> Subject: [PATCH] clk: visconti: prevent array overflow in
> visconti_clk_register_gates()
>
> This code was using -1 to represent that there was no reset function.
> Unfortunately, the -1 was stored in u8 so the if (clks[i].rs_id >= 0) condition
> was always true. This lead to an out of bounds access in
> visconti_clk_register_gates().
>
> Fixes: b4cbe606dc36 ("clk: visconti: Add support common clock driver and
> reset driver")
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Looks good to me.
Acked-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
Best regards,
Nobuhiro
> ---
> drivers/clk/visconti/clkc.h | 3 +++
> drivers/clk/visconti/clkc-tmpv770x.c | 2 +-
> drivers/clk/visconti/clkc.c | 2 +-
> 3 files changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/clk/visconti/clkc.h b/drivers/clk/visconti/clkc.h index
> 09ed82ff64e4..8756a1ec42ef 100644
> --- a/drivers/clk/visconti/clkc.h
> +++ b/drivers/clk/visconti/clkc.h
> @@ -73,4 +73,7 @@ int visconti_clk_register_gates(struct
> visconti_clk_provider *data,
> int num_gate,
> const struct visconti_reset_data *reset,
> spinlock_t *lock);
> +
> +#define NO_RESET 0xFF
> +
> #endif /* _VISCONTI_CLKC_H_ */
> diff --git a/drivers/clk/visconti/clkc-tmpv770x.c
> b/drivers/clk/visconti/clkc-tmpv770x.c
> index c2b2f41a85a4..6c753b2cb558 100644
> --- a/drivers/clk/visconti/clkc-tmpv770x.c
> +++ b/drivers/clk/visconti/clkc-tmpv770x.c
> @@ -176,7 +176,7 @@ static const struct visconti_clk_gate_table
> clk_gate_tables[] = {
> { TMPV770X_CLK_WRCK, "wrck",
> clks_parent_data, ARRAY_SIZE(clks_parent_data),
> 0, 0x68, 0x168, 9, 32,
> - -1, }, /* No reset */
> + NO_RESET, },
> { TMPV770X_CLK_PICKMON, "pickmon",
> clks_parent_data, ARRAY_SIZE(clks_parent_data),
> 0, 0x10, 0x110, 8, 4,
> diff --git a/drivers/clk/visconti/clkc.c b/drivers/clk/visconti/clkc.c index
> 56a8a4ffebca..d0b193b5d0b3 100644
> --- a/drivers/clk/visconti/clkc.c
> +++ b/drivers/clk/visconti/clkc.c
> @@ -147,7 +147,7 @@ int visconti_clk_register_gates(struct
> visconti_clk_provider *ctx,
> if (!dev_name)
> return -ENOMEM;
>
> - if (clks[i].rs_id >= 0) {
> + if (clks[i].rs_id != NO_RESET) {
> rson_offset = reset[clks[i].rs_id].rson_offset;
> rsoff_offset = reset[clks[i].rs_id].rsoff_offset;
> rs_idx = reset[clks[i].rs_id].rs_idx;
> --
> 2.20.1
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] clk: visconti: prevent array overflow in visconti_clk_register_gates()
2022-03-16 8:35 [PATCH] clk: visconti: prevent array overflow in visconti_clk_register_gates() Dan Carpenter
2022-03-17 5:19 ` nobuhiro1.iwamatsu
@ 2022-03-17 19:21 ` Stephen Boyd
1 sibling, 0 replies; 3+ messages in thread
From: Stephen Boyd @ 2022-03-17 19:21 UTC (permalink / raw)
To: Dan Carpenter, Michael Turquette, Nobuhiro Iwamatsu
Cc: linux-clk, kernel-janitors
Quoting Dan Carpenter (2022-03-16 01:35:33)
> This code was using -1 to represent that there was no reset function.
> Unfortunately, the -1 was stored in u8 so the if (clks[i].rs_id >= 0)
> condition was always true. This lead to an out of bounds access in
> visconti_clk_register_gates().
>
> Fixes: b4cbe606dc36 ("clk: visconti: Add support common clock driver and reset driver")
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> ---
Applied to clk-next
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2022-03-17 19:22 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-03-16 8:35 [PATCH] clk: visconti: prevent array overflow in visconti_clk_register_gates() Dan Carpenter
2022-03-17 5:19 ` nobuhiro1.iwamatsu
2022-03-17 19:21 ` Stephen Boyd
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox