* [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map()
@ 2024-10-10 11:18 Harshit Mogalapalli
2024-10-10 17:17 ` Christophe JAILLET
2024-10-10 19:13 ` Linus Walleij
0 siblings, 2 replies; 6+ messages in thread
From: Harshit Mogalapalli @ 2024-10-10 11:18 UTC (permalink / raw)
To: Linus Walleij, Chen Wang, Inochi Amaoto, Harshit Mogalapalli,
linux-gpio, linux-kernel
Cc: dan.carpenter, kernel-janitors, error27
'map' is allocated using devm_* which takes care of freeing the allocated
data, but in error paths there is a call to pinctrl_utils_free_map()
which also does kfree(map) which leads to a double free.
Use kcalloc() instead of devm_kcalloc() as freeing is manually handled.
Fixes: a29d8e93e710 ("pinctrl: sophgo: add support for CV1800B SoC")
Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
---
This is based on static analysis with smatch, only compile tested.
---
drivers/pinctrl/sophgo/pinctrl-cv18xx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/pinctrl/sophgo/pinctrl-cv18xx.c b/drivers/pinctrl/sophgo/pinctrl-cv18xx.c
index d18fc5aa84f7..57f2674e75d6 100644
--- a/drivers/pinctrl/sophgo/pinctrl-cv18xx.c
+++ b/drivers/pinctrl/sophgo/pinctrl-cv18xx.c
@@ -221,7 +221,7 @@ static int cv1800_pctrl_dt_node_to_map(struct pinctrl_dev *pctldev,
if (!grpnames)
return -ENOMEM;
- map = devm_kcalloc(dev, ngroups * 2, sizeof(*map), GFP_KERNEL);
+ map = kcalloc(ngroups * 2, sizeof(*map), GFP_KERNEL);
if (!map)
return -ENOMEM;
--
2.39.3
^ permalink raw reply related [flat|nested] 6+ messages in thread* Re: [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map() 2024-10-10 11:18 [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map() Harshit Mogalapalli @ 2024-10-10 17:17 ` Christophe JAILLET 2024-10-10 18:33 ` Dan Carpenter 2024-10-10 19:11 ` Harshit Mogalapalli 2024-10-10 19:13 ` Linus Walleij 1 sibling, 2 replies; 6+ messages in thread From: Christophe JAILLET @ 2024-10-10 17:17 UTC (permalink / raw) To: Harshit Mogalapalli, Linus Walleij, Chen Wang, Inochi Amaoto, linux-gpio, linux-kernel Cc: dan.carpenter, kernel-janitors, error27 Le 10/10/2024 à 13:18, Harshit Mogalapalli a écrit : > 'map' is allocated using devm_* which takes care of freeing the allocated > data, but in error paths there is a call to pinctrl_utils_free_map() > which also does kfree(map) which leads to a double free. > > Use kcalloc() instead of devm_kcalloc() as freeing is manually handled. > > Fixes: a29d8e93e710 ("pinctrl: sophgo: add support for CV1800B SoC") > Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com> > --- > This is based on static analysis with smatch, only compile tested. > --- > drivers/pinctrl/sophgo/pinctrl-cv18xx.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/pinctrl/sophgo/pinctrl-cv18xx.c b/drivers/pinctrl/sophgo/pinctrl-cv18xx.c > index d18fc5aa84f7..57f2674e75d6 100644 > --- a/drivers/pinctrl/sophgo/pinctrl-cv18xx.c > +++ b/drivers/pinctrl/sophgo/pinctrl-cv18xx.c > @@ -221,7 +221,7 @@ static int cv1800_pctrl_dt_node_to_map(struct pinctrl_dev *pctldev, > if (!grpnames) > return -ENOMEM; > > - map = devm_kcalloc(dev, ngroups * 2, sizeof(*map), GFP_KERNEL); > + map = kcalloc(ngroups * 2, sizeof(*map), GFP_KERNEL); > if (!map) > return -ENOMEM; > Hi, drivers/pinctrl/nuvoton/pinctrl-ma35.c seems to have the same issue. CJ ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map() 2024-10-10 17:17 ` Christophe JAILLET @ 2024-10-10 18:33 ` Dan Carpenter 2024-10-10 20:32 ` Christophe JAILLET 2024-10-10 19:11 ` Harshit Mogalapalli 1 sibling, 1 reply; 6+ messages in thread From: Dan Carpenter @ 2024-10-10 18:33 UTC (permalink / raw) To: Christophe JAILLET Cc: Harshit Mogalapalli, Linus Walleij, Chen Wang, Inochi Amaoto, linux-gpio, linux-kernel, kernel-janitors, error27 On Thu, Oct 10, 2024 at 07:17:19PM +0200, Christophe JAILLET wrote: > Le 10/10/2024 à 13:18, Harshit Mogalapalli a écrit : > > 'map' is allocated using devm_* which takes care of freeing the allocated > > data, but in error paths there is a call to pinctrl_utils_free_map() > > which also does kfree(map) which leads to a double free. > > > > Use kcalloc() instead of devm_kcalloc() as freeing is manually handled. > > > > Fixes: a29d8e93e710 ("pinctrl: sophgo: add support for CV1800B SoC") > > Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com> > > --- > > This is based on static analysis with smatch, only compile tested. > > --- > > drivers/pinctrl/sophgo/pinctrl-cv18xx.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/pinctrl/sophgo/pinctrl-cv18xx.c b/drivers/pinctrl/sophgo/pinctrl-cv18xx.c > > index d18fc5aa84f7..57f2674e75d6 100644 > > --- a/drivers/pinctrl/sophgo/pinctrl-cv18xx.c > > +++ b/drivers/pinctrl/sophgo/pinctrl-cv18xx.c > > @@ -221,7 +221,7 @@ static int cv1800_pctrl_dt_node_to_map(struct pinctrl_dev *pctldev, > > if (!grpnames) > > return -ENOMEM; > > - map = devm_kcalloc(dev, ngroups * 2, sizeof(*map), GFP_KERNEL); > > + map = kcalloc(ngroups * 2, sizeof(*map), GFP_KERNEL); > > if (!map) > > return -ENOMEM; > > Hi, > > drivers/pinctrl/nuvoton/pinctrl-ma35.c seems to have the same issue. > Yep. That one is too complicated for Smatch to find. In this case, the kfree() happened in the cleanup so it was in the same function. regards, dan carpenter ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map() 2024-10-10 18:33 ` Dan Carpenter @ 2024-10-10 20:32 ` Christophe JAILLET 0 siblings, 0 replies; 6+ messages in thread From: Christophe JAILLET @ 2024-10-10 20:32 UTC (permalink / raw) To: Dan Carpenter Cc: Harshit Mogalapalli, Linus Walleij, Chen Wang, Inochi Amaoto, linux-gpio, linux-kernel, kernel-janitors, error27 Le 10/10/2024 à 20:33, Dan Carpenter a écrit : > On Thu, Oct 10, 2024 at 07:17:19PM +0200, Christophe JAILLET wrote: >> Le 10/10/2024 à 13:18, Harshit Mogalapalli a écrit : >>> 'map' is allocated using devm_* which takes care of freeing the allocated >>> data, but in error paths there is a call to pinctrl_utils_free_map() >>> which also does kfree(map) which leads to a double free. >>> >>> Use kcalloc() instead of devm_kcalloc() as freeing is manually handled. >>> >>> Fixes: a29d8e93e710 ("pinctrl: sophgo: add support for CV1800B SoC") >>> Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com> >>> --- >>> This is based on static analysis with smatch, only compile tested. >>> --- >>> drivers/pinctrl/sophgo/pinctrl-cv18xx.c | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/drivers/pinctrl/sophgo/pinctrl-cv18xx.c b/drivers/pinctrl/sophgo/pinctrl-cv18xx.c >>> index d18fc5aa84f7..57f2674e75d6 100644 >>> --- a/drivers/pinctrl/sophgo/pinctrl-cv18xx.c >>> +++ b/drivers/pinctrl/sophgo/pinctrl-cv18xx.c >>> @@ -221,7 +221,7 @@ static int cv1800_pctrl_dt_node_to_map(struct pinctrl_dev *pctldev, >>> if (!grpnames) >>> return -ENOMEM; >>> - map = devm_kcalloc(dev, ngroups * 2, sizeof(*map), GFP_KERNEL); >>> + map = kcalloc(ngroups * 2, sizeof(*map), GFP_KERNEL); >>> if (!map) >>> return -ENOMEM; >> >> Hi, >> >> drivers/pinctrl/nuvoton/pinctrl-ma35.c seems to have the same issue. >> > > Yep. That one is too complicated for Smatch to find. In this case, the kfree() > happened in the cleanup so it was in the same function. > > regards, > dan carpenter > > For the records, I spotted the other case with coccinelle: @ok1@ identifier X, fct; @@ struct pinctrl_ops X = { .dt_node_to_map = fct, .dt_free_map = pinconf_generic_dt_free_map, }; @test1 depends on ok1@ identifier fct = ok1.fct; identifier fct2 =~ "devm"; expression x =~ "map"; @@ int fct(...) { ... * x = fct2(...); ... } @ok2@ identifier X, fct; @@ struct pinctrl_ops X = { .dt_node_to_map = fct, .dt_free_map = pinctrl_utils_free_map, }; @test2 depends on ok2@ identifier fct = ok2.fct; identifier fct2 =~ "devm"; expression x =~ "map"; @@ int fct(...) { ... * x = fct2(...); ... } CJ ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map() 2024-10-10 17:17 ` Christophe JAILLET 2024-10-10 18:33 ` Dan Carpenter @ 2024-10-10 19:11 ` Harshit Mogalapalli 1 sibling, 0 replies; 6+ messages in thread From: Harshit Mogalapalli @ 2024-10-10 19:11 UTC (permalink / raw) To: Christophe JAILLET, Linus Walleij, Chen Wang, Inochi Amaoto, linux-gpio, linux-kernel Cc: dan.carpenter, kernel-janitors, error27 Hi Christophe, > > Hi, > > drivers/pinctrl/nuvoton/pinctrl-ma35.c seems to have the same issue. > Thank you! I will work on fix for this, thanks!! Regards, Harshit > CJ > > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map() 2024-10-10 11:18 [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map() Harshit Mogalapalli 2024-10-10 17:17 ` Christophe JAILLET @ 2024-10-10 19:13 ` Linus Walleij 1 sibling, 0 replies; 6+ messages in thread From: Linus Walleij @ 2024-10-10 19:13 UTC (permalink / raw) To: Harshit Mogalapalli Cc: Chen Wang, Inochi Amaoto, linux-gpio, linux-kernel, dan.carpenter, kernel-janitors, error27 On Thu, Oct 10, 2024 at 1:24 PM Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com> wrote: > 'map' is allocated using devm_* which takes care of freeing the allocated > data, but in error paths there is a call to pinctrl_utils_free_map() > which also does kfree(map) which leads to a double free. > > Use kcalloc() instead of devm_kcalloc() as freeing is manually handled. > > Fixes: a29d8e93e710 ("pinctrl: sophgo: add support for CV1800B SoC") > Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com> Patch applied for fixes! Yours, Linus Walleij ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2024-10-10 20:32 UTC | newest] Thread overview: 6+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2024-10-10 11:18 [PATCH] pinctrl: sophgo: fix double free in cv1800_pctrl_dt_node_to_map() Harshit Mogalapalli 2024-10-10 17:17 ` Christophe JAILLET 2024-10-10 18:33 ` Dan Carpenter 2024-10-10 20:32 ` Christophe JAILLET 2024-10-10 19:11 ` Harshit Mogalapalli 2024-10-10 19:13 ` Linus Walleij
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).