[PATCH bpf] riscv: remove BPF JIT for nommu builds

[PATCH bpf] riscv: remove BPF JIT for nommu builds

[Björn Töpel]

The BPF JIT fails to build for kernels configured to !MMU. Without an
MMU, the BPF JIT does not make much sense, therefore this patch
disables the JIT for nommu builds.

This was reported by the kbuild test robot:

   All errors (new ones prefixed by >>):

      arch/riscv/net/bpf_jit_comp64.c: In function 'bpf_jit_alloc_exec':
   >> arch/riscv/net/bpf_jit_comp64.c:1094:47: error: 'BPF_JIT_REGION_START' undeclared (first use in this function)
       1094 |  return __vmalloc_node_range(size, PAGE_SIZE, BPF_JIT_REGION_START,
            |                                               ^~~~~~~~~~~~~~~~~~~~
      arch/riscv/net/bpf_jit_comp64.c:1094:47: note: each undeclared identifier is reported only once for each function it appears in
   >> arch/riscv/net/bpf_jit_comp64.c:1095:9: error: 'BPF_JIT_REGION_END' undeclared (first use in this function)
       1095 |         BPF_JIT_REGION_END, GFP_KERNEL,
            |         ^~~~~~~~~~~~~~~~~~
      arch/riscv/net/bpf_jit_comp64.c:1098:1: warning: control reaches end of non-void function [-Wreturn-type]
       1098 | }
            | ^

Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
---
 arch/riscv/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
index 8672e77a5b7a..bd35ac72fe24 100644
--- a/arch/riscv/Kconfig
+++ b/arch/riscv/Kconfig
@@ -55,7 +55,7 @@ config RISCV
 	select ARCH_HAS_PTE_SPECIAL
 	select ARCH_HAS_MMIOWB
 	select ARCH_HAS_DEBUG_VIRTUAL
-	select HAVE_EBPF_JIT
+	select HAVE_EBPF_JIT if MMU
 	select EDAC_SUPPORT
 	select ARCH_HAS_GIGANTIC_PAGE
 	select ARCH_WANT_HUGE_PMD_SHARE if 64BIT
-- 
2.20.1

Re: [PATCH bpf] riscv: remove BPF JIT for nommu builds

[Luke Nelson]

Thanks for looking into this!

Acked-by: Luke Nelson <luke.r.nels@gmail.com>

Re: [PATCH bpf] riscv: remove BPF JIT for nommu builds

[Daniel Borkmann]

On 3/31/20 12:10 PM, Björn Töpel wrote:
> The BPF JIT fails to build for kernels configured to !MMU. Without an
> MMU, the BPF JIT does not make much sense, therefore this patch
> disables the JIT for nommu builds.
> 
> This was reported by the kbuild test robot:
> 
>     All errors (new ones prefixed by >>):
> 
>        arch/riscv/net/bpf_jit_comp64.c: In function 'bpf_jit_alloc_exec':
>     >> arch/riscv/net/bpf_jit_comp64.c:1094:47: error: 'BPF_JIT_REGION_START' undeclared (first use in this function)
>         1094 |  return __vmalloc_node_range(size, PAGE_SIZE, BPF_JIT_REGION_START,
>              |                                               ^~~~~~~~~~~~~~~~~~~~
>        arch/riscv/net/bpf_jit_comp64.c:1094:47: note: each undeclared identifier is reported only once for each function it appears in
>     >> arch/riscv/net/bpf_jit_comp64.c:1095:9: error: 'BPF_JIT_REGION_END' undeclared (first use in this function)
>         1095 |         BPF_JIT_REGION_END, GFP_KERNEL,
>              |         ^~~~~~~~~~~~~~~~~~
>        arch/riscv/net/bpf_jit_comp64.c:1098:1: warning: control reaches end of non-void function [-Wreturn-type]
>         1098 | }
>              | ^
> 
> Reported-by: kbuild test robot <lkp@intel.com>
> Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>

Applied, thanks!