nfp: bpf: fix return address from register-saving subroutine to callee
authorQuentin Monnet <quentin.monnet@netronome.com>
Sun, 7 Oct 2018 11:56:55 +0000 (12:56 +0100)
committerDaniel Borkmann <daniel@iogearbox.net>
Mon, 8 Oct 2018 08:24:13 +0000 (10:24 +0200)
commit2178f3f0dc200557312e783aa683b87794084ae2
tree614c80c50dd9ab3186f6091c5f99937073acfab6
parentbdf4c66faf5fa6fd5ffb0b59c39c7629103d6479
nfp: bpf: fix return address from register-saving subroutine to callee

On performing a BPF-to-BPF call, we first jump to a subroutine that
pushes callee-saved registers (R6~R9) to the stack, and from there we
goes to the start of the callee next. In order to do so, the caller must
pass to the subroutine the address of the NFP instruction to jump to at
the end of that subroutine. This cannot be reliably implemented when
translated the caller, as we do not always know the start offset of the
callee yet.

This patch implement the required fixup step for passing the start
offset in the callee via the register used by the subroutine to hold its
return address.

Signed-off-by: Quentin Monnet <quentin.monnet@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
drivers/net/ethernet/netronome/nfp/bpf/jit.c