Tiezhu Yang (2):
bpf: Add some description about BPF_JIT_ALWAYS_ON in Kconfig
bpf: Make BPF_JIT_DEFAULT_ON selectable in Kconfig
kernel/bpf/Kconfig | 16 ++++++++++++++--
1 file changed, 14 insertions(+), 2 deletions(-)
--
2.1.0
When CONFIG_BPF_JIT_ALWAYS_ON is enabled, bpf_jit_enable is permanently
set to 1 and setting any other value than that will return in failure.
Add the above description in the help text of BPF_JIT_ALWAYS_ON, and then
we can distinguish between BPF_JIT_ALWAYS_ON and BPF_JIT_DEFAULT_ON.
Signed-off-by: Tiezhu Yang <[email protected]>
---
kernel/bpf/Kconfig | 3 +++
1 file changed, 3 insertions(+)
diff --git a/kernel/bpf/Kconfig b/kernel/bpf/Kconfig
index d24d518..88409f8 100644
--- a/kernel/bpf/Kconfig
+++ b/kernel/bpf/Kconfig
@@ -58,6 +58,9 @@ config BPF_JIT_ALWAYS_ON
Enables BPF JIT and removes BPF interpreter to avoid speculative
execution of BPF instructions by the interpreter.
+ When CONFIG_BPF_JIT_ALWAYS_ON is enabled, bpf_jit_enable is permanently
+ set to 1 and setting any other value than that will return in failure.
+
config BPF_JIT_DEFAULT_ON
def_bool ARCH_WANT_DEFAULT_BPF_JIT || BPF_JIT_ALWAYS_ON
depends on HAVE_EBPF_JIT && BPF_JIT
--
2.1.0
On 2/11/22 9:57 AM, Tiezhu Yang wrote:
> When CONFIG_BPF_JIT_ALWAYS_ON is enabled, bpf_jit_enable is permanently
> set to 1 and setting any other value than that will return in failure.
>
> Add the above description in the help text of BPF_JIT_ALWAYS_ON, and then
> we can distinguish between BPF_JIT_ALWAYS_ON and BPF_JIT_DEFAULT_ON.
>
> Signed-off-by: Tiezhu Yang <[email protected]>
> ---
> kernel/bpf/Kconfig | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/kernel/bpf/Kconfig b/kernel/bpf/Kconfig
> index d24d518..88409f8 100644
> --- a/kernel/bpf/Kconfig
> +++ b/kernel/bpf/Kconfig
> @@ -58,6 +58,9 @@ config BPF_JIT_ALWAYS_ON
> Enables BPF JIT and removes BPF interpreter to avoid speculative
> execution of BPF instructions by the interpreter.
>
> + When CONFIG_BPF_JIT_ALWAYS_ON is enabled, bpf_jit_enable is permanently
> + set to 1 and setting any other value than that will return in failure.
Small nit here: lets use the full path (/proc/sys/net/core/bpf_jit_enable) in order
to be consistent with the other descriptions in this Kconfig.
> config BPF_JIT_DEFAULT_ON
> def_bool ARCH_WANT_DEFAULT_BPF_JIT || BPF_JIT_ALWAYS_ON
> depends on HAVE_EBPF_JIT && BPF_JIT
>
Currently, it is not possible to set bpf_jit_enable to 1 by default
and the users can change it to 0 or 2, it seems bad for some users,
make BPF_JIT_DEFAULT_ON selectable to give them a chance.
Signed-off-by: Tiezhu Yang <[email protected]>
---
kernel/bpf/Kconfig | 13 +++++++++++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/kernel/bpf/Kconfig b/kernel/bpf/Kconfig
index 88409f8..28b7d71 100644
--- a/kernel/bpf/Kconfig
+++ b/kernel/bpf/Kconfig
@@ -54,6 +54,7 @@ config BPF_JIT
config BPF_JIT_ALWAYS_ON
bool "Permanently enable BPF JIT and remove BPF interpreter"
depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
+ select BPF_JIT_DEFAULT_ON
help
Enables BPF JIT and removes BPF interpreter to avoid speculative
execution of BPF instructions by the interpreter.
@@ -62,8 +63,16 @@ config BPF_JIT_ALWAYS_ON
set to 1 and setting any other value than that will return in failure.
config BPF_JIT_DEFAULT_ON
- def_bool ARCH_WANT_DEFAULT_BPF_JIT || BPF_JIT_ALWAYS_ON
- depends on HAVE_EBPF_JIT && BPF_JIT
+ bool "Defaultly enable BPF JIT and remove BPF interpreter"
+ default y if ARCH_WANT_DEFAULT_BPF_JIT
+ depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
+ help
+ Enables BPF JIT and removes BPF interpreter to avoid speculative
+ execution of BPF instructions by the interpreter.
+
+ When CONFIG_BPF_JIT_DEFAULT_ON is enabled but CONFIG_BPF_JIT_ALWAYS_ON
+ is disabled, bpf_jit_enable is set to 1 by default and can be changed
+ to 0 or 2.
config BPF_UNPRIV_DEFAULT_OFF
bool "Disable unprivileged BPF by default"
--
2.1.0
On 2/11/22 9:57 AM, Tiezhu Yang wrote:
> Currently, it is not possible to set bpf_jit_enable to 1 by default
> and the users can change it to 0 or 2, it seems bad for some users,
> make BPF_JIT_DEFAULT_ON selectable to give them a chance.
I'm not fully sure I follow the above, so you are saying that a kconfig of
!BPF_JIT_ALWAYS_ON and ARCH_WANT_DEFAULT_BPF_JIT, enables BPF_JIT_DEFAULT_ON
however in such setting you are not able to reset bpf_jit_enable back to 0 at
runtime?
Thanks,
Daniel
On 02/11/2022 06:23 PM, Daniel Borkmann wrote:
> On 2/11/22 9:57 AM, Tiezhu Yang wrote:
>> Currently, it is not possible to set bpf_jit_enable to 1 by default
>> and the users can change it to 0 or 2, it seems bad for some users,
>> make BPF_JIT_DEFAULT_ON selectable to give them a chance.
>
> I'm not fully sure I follow the above, so you are saying that a kconfig of
> !BPF_JIT_ALWAYS_ON and ARCH_WANT_DEFAULT_BPF_JIT, enables
> BPF_JIT_DEFAULT_ON
> however in such setting you are not able to reset bpf_jit_enable back to
> 0 at
> runtime?
Oh, no. Sorry for the unclear description.
currently, only x86, arm64 and s390 select ARCH_WANT_DEFAULT_BPF_JIT,
the other archs do not select ARCH_WANT_DEFAULT_BPF_JIT. On the archs
without ARCH_WANT_DEFAULT_BPF_JIT, if we want to set bpf_jit_enable to
1 by default, the only way is to enable CONFIG_BPF_JIT_ALWAYS_ON, then
the users can not change it to 0 or 2, it seems bad for some users, we
can select ARCH_WANT_DEFAULT_BPF_JIT for those archs if it is proper,
but at least for now, make BPF_JIT_DEFAULT_ON selectable can give them
a chance.
Additionaly, with this patch, under !BPF_JIT_ALWAYS_ON, we can disable
BPF_JIT_DEFAULT_ON on the archs with ARCH_WANT_DEFAULT_BPF_JIT when
make menuconfig, it seems flexible for some developers.
If you are OK, I will update the commit message and then send v2.
Thanks,
Tiezhu
>
> Thanks,
> Daniel