Re: [PATCH 2/2] powerpc/uaccess: Move might_fault() into user_access_begin()

Michael Ellerman a écrit : Christophe Leroy writes: Le 08/02/2021 à 14:57, Michael Ellerman a écrit : We have a might_fault() check in __unsafe_put_user_goto(), but that is dangerous as it potentially calls lots of code while user access is enabled. It also triggers the check Alexey added t

Re: [PATCH 2/2] powerpc/uaccess: Move might_fault() into user_access_begin()

Christophe Leroy writes: > Le 08/02/2021 à 14:57, Michael Ellerman a écrit : >> We have a might_fault() check in __unsafe_put_user_goto(), but that is >> dangerous as it potentially calls lots of code while user access is >> enabled. >> >> It also triggers the check Alexey added to the irq restor

Re: [PATCH 2/2] powerpc/uaccess: Move might_fault() into user_access_begin()

Le 08/02/2021 à 14:57, Michael Ellerman a écrit : We have a might_fault() check in __unsafe_put_user_goto(), but that is dangerous as it potentially calls lots of code while user access is enabled. It also triggers the check Alexey added to the irq restore path to catch cases like that: W

[PATCH 2/2] powerpc/uaccess: Move might_fault() into user_access_begin()

We have a might_fault() check in __unsafe_put_user_goto(), but that is dangerous as it potentially calls lots of code while user access is enabled. It also triggers the check Alexey added to the irq restore path to catch cases like that: WARNING: CPU: 30 PID: 1 at arch/powerpc/include/asm/book3