mbox series

[bpf-next,v4,0/2] bpf, arm64: fix bpf line info

Message ID 20220226121906.5709-1-houtao1@huawei.com (mailing list archive)
Headers show
Series bpf, arm64: fix bpf line info | expand

Message

Hou Tao Feb. 26, 2022, 12:19 p.m. UTC
Hi,

The patchset addresses two issues in bpf line info for arm64:

(1) insn_to_jit_off only considers the body itself and ignores
    prologue before the body. Fixed in patch #1.

(2) insn_to_jit_off passed to bpf_prog_fill_jited_linfo() is
    calculated in instruction granularity instead of bytes
    granularity. Fixed in patch #2.

Comments are always welcome.

Regards,
Tao

Change Log:
v4:
 * patch #2: convert ctx.offset into byte offset before call
   bpf_prog_fill_jited_linfo() instead of converting it back and forth.

v3: https://lore.kernel.org/bpf/20220208012539.491753-1-houtao1@huawei.com
 * patch #2: explain why bpf2a64_offset() needs update
 * add Fixes tags in both patches

v2: https://lore.kernel.org/bpf/20220125105707.292449-1-houtao1@huawei.com
 * split into two independent patches (from Daniel)
 * use AARCH64_INSN_SIZE instead of defining INSN_SIZE

v1: https://lore.kernel.org/bpf/20220104014236.1512639-1-houtao1@huawei.com

Hou Tao (2):
  bpf, arm64: call build_prologue() first in first JIT pass
  bpf, arm64: feed byte-offset into bpf line info

 arch/arm64/net/bpf_jit_comp.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)

Comments

Hou Tao Feb. 28, 2022, 1:15 a.m. UTC | #1
Hi,

On 2/26/2022 8:19 PM, Hou Tao wrote:
> Hi,
>
> The patchset addresses two issues in bpf line info for arm64:
>
> (1) insn_to_jit_off only considers the body itself and ignores
>     prologue before the body. Fixed in patch #1.
>
> (2) insn_to_jit_off passed to bpf_prog_fill_jited_linfo() is
>     calculated in instruction granularity instead of bytes
>     granularity. Fixed in patch #2.
>
> Comments are always welcome.
>
> Regards,
> Tao
>
> Change Log:
> v4:
>  * patch #2: convert ctx.offset into byte offset before call
>    bpf_prog_fill_jited_linfo() instead of converting it back and forth.
Oops, forget to add Acked-by tag from Martin
>
> v3: https://lore.kernel.org/bpf/20220208012539.491753-1-houtao1@huawei.com
>  * patch #2: explain why bpf2a64_offset() needs update
>  * add Fixes tags in both patches
>
> v2: https://lore.kernel.org/bpf/20220125105707.292449-1-houtao1@huawei.com
>  * split into two independent patches (from Daniel)
>  * use AARCH64_INSN_SIZE instead of defining INSN_SIZE
>
> v1: https://lore.kernel.org/bpf/20220104014236.1512639-1-houtao1@huawei.com
>
> Hou Tao (2):
>   bpf, arm64: call build_prologue() first in first JIT pass
>   bpf, arm64: feed byte-offset into bpf line info
>
>  arch/arm64/net/bpf_jit_comp.c | 18 +++++++++++++-----
>  1 file changed, 13 insertions(+), 5 deletions(-)
>
Daniel Borkmann Feb. 28, 2022, 12:59 p.m. UTC | #2
On 2/26/22 1:19 PM, Hou Tao wrote:
> Hi,
> 
> The patchset addresses two issues in bpf line info for arm64:
> 
> (1) insn_to_jit_off only considers the body itself and ignores
>      prologue before the body. Fixed in patch #1.
> 
> (2) insn_to_jit_off passed to bpf_prog_fill_jited_linfo() is
>      calculated in instruction granularity instead of bytes
>      granularity. Fixed in patch #2.
> 
> Comments are always welcome.
> 
> Regards,
> Tao

LGTM, applied, thanks Hou!