soc: qcom: geni: Fix NULL pointer dereference
Subsystems (2)
PATCH 1/1 soc: qcom: geni: Fix NULL pointer dereference
pdev struct doesn't exits for the devices whose status are disabled from DT node, in such cases NULL is returned from 'of_find_device_by_node' Later when we try to get drvdata from pdev struct NULL pointer dereference is triggered.
Add a NULL check for return values to fix the issue.
We were hitting this issue when one of QUP is disabled.
--- drivers/soc/qcom/qcom-geni-se.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/soc/qcom/qcom-geni-se.c +7 -1
@@ -820,6 +820,7 @@ void geni_remove_earlycon_icc_vote(void)struct geni_wrapper *wrapper;struct device_node *parent;struct device_node *child; +struct platform_device *wrapper_pdev;if (!earlycon_wrapper)return ;
@@ -829,7 +830,12 @@ void geni_remove_earlycon_icc_vote(void)for_each_child_of_node (parent, child) {if (!of_device_is_compatible (child,"qcom,geni-se-qup" ))continue ; - wrapper =platform_get_drvdata (of_find_device_by_node (child)); + + wrapper_pdev =of_find_device_by_node (child); +if (!wrapper_pdev) +continue ; + + wrapper =platform_get_drvdata (wrapper_pdev);icc_put (wrapper->to_core.path); wrapper->to_core.path =NULL ;
Signature · 1 line
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,\na Linux Foundation Collaborative Project
Please make sure to cc the linux-arm-msm@vger list for patches of Qualcomm code.
Quoted 1 line · Akash Asthana
pdev struct doesn't exits for the devices whose status are disabled
s/exits/exist/
Quoted 20 lines
from DT node, in such cases NULL is returned from '
of_find_device_by_node' Later when we try to get drvdata from pdev struct NULL pointer dereference is triggered.Add a NULL check for return values to fix the issue.
We were hitting this issue when one of QUP is disabled.
Fixes 048eb908a1f2 ("soc: qcom-geni-se: Add interconnect support to fix earlycon crash")Reported-by Sai Prakash Ranjan <[email protected]>Signed-off-by Akash Asthana <[email protected]>--- drivers/soc/qcom/qcom-geni-se.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)drivers/soc/qcom/qcom-geni-se.c +0
-0
drivers/soc/qcom/qcom-geni-se.c void geni_remove_earlycon_icc_vote(void) · line 823
Quoted 4 lines
drivers/soc/qcom/qcom-geni-se.c +1
-0@@ -820,6 +820,7 @@ void geni_remove_earlycon_icc_vote(void)struct geni_wrapper *wrapper;struct device_node *parent;struct device_node *child; +struct platform_device *wrapper_pdev;
nit: since there is no other 'pdev' in this function you could just name it 'pdev', which is less clunky. The variable is only used immediately after it is assigned, so it's clear from the context that it refers to the 'wrapper'.
Hi Matthias,
Quoted 2 lines · Matthias Kaehlcke
Please make sure to cc the linux-arm-msm@vger list for patches of Qualcomm code.
Sure.
Quoted 3 lines · Akash Asthana
Quoted 1 line
pdev struct doesn't exits for the devices whose status are disabled
s/exits/exist/
ok
Quoted 16 lines
Quoted 16 lines
from DT node, in such cases NULL is returned from '
of_find_device_by_node' Later when we try to get drvdata from pdev struct NULL pointer dereference is triggered.Add a NULL check for return values to fix the issue.
We were hitting this issue when one of QUP is disabled.
Fixes 048eb908a1f2 ("soc: qcom-geni-se: Add interconnect support to fix earlycon crash")Reported-by Sai Prakash Ranjan <[email protected]>Signed-off-by Akash Asthana <[email protected]>--- drivers/soc/qcom/qcom-geni-se.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)
Quoted 13 lines
drivers/soc/qcom/qcom-geni-se.c void geni_remove_earlycon_icc_vote(void) · line 823
Quoted 9 lines
drivers/soc/qcom/qcom-geni-se.c +1
-0@@ -820,6 +820,7 @@ void geni_remove_earlycon_icc_vote(void)struct geni_wrapper *wrapper;struct device_node *parent;struct device_node *child; +struct platform_device *wrapper_pdev;nit: since there is no other 'pdev' in this function you could just name it 'pdev', which is less clunky. The variable is only used immediately after it is assigned, so it's clear from the context that it refers to the 'wrapper'.
ok
Thankyou for review!.
regards,
Akash
Quoted 2 lines
Reviewed-by Matthias Kaehlcke <[email protected]>
Signature · 1 line
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,\na Linux Foundation Collaborative Project
Quoted 5 lines · Akash Asthana
pdev struct doesn't exits for the devices whose status are disabled from DT node, in such cases NULL is returned from '
of_find_device_by_node' Later when we try to get drvdata from pdev struct NULL pointer dereference is triggered.
Quoted 7 lines
Add a NULL check for return values to fix the issue.
We were hitting this issue when one of QUP is disabled.
Fixes 048eb908a1f2 ("soc: qcom-geni-se: Add interconnect support to fix earlycon crash")Reported-by Sai Prakash Ranjan <[email protected]>Signed-off-by Akash Asthana <[email protected]>
And applied with Matthias suggested rename of wrapper_pdev to pdev.
PS. Please include [email protected] among your recipients for future patches.
Regards,
Bjorn
Quoted 33 lines
--- drivers/soc/qcom/qcom-geni-se.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)drivers/soc/qcom/qcom-geni-se.c +7
-1@@ -820,6 +820,7 @@ void geni_remove_earlycon_icc_vote(void)struct geni_wrapper *wrapper;struct device_node *parent;struct device_node *child; +struct platform_device *wrapper_pdev;if (!earlycon_wrapper)return ;@@ -829,7 +830,12 @@ void geni_remove_earlycon_icc_vote(void)for_each_child_of_node (parent, child) {if (!of_device_is_compatible (child,"qcom,geni-se-qup" ))continue ; - wrapper =platform_get_drvdata (of_find_device_by_node (child)); + + wrapper_pdev =of_find_device_by_node (child); +if (!wrapper_pdev) +continue ; + + wrapper =platform_get_drvdata (wrapper_pdev);icc_put (wrapper->to_core.path); wrapper->to_core.path =NULL ;Signature · 1 line
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,\na Linux Foundation Collaborative Project