Linux DTrace development list
 help / color / mirror / Atom feed
* [PATCH] Use DT_TRAMP_SP_SLOT() for BPF stack scratch space in trampoline
@ 2025-02-21 20:09 eugene.loh
  2025-02-21 20:20 ` [DTrace-devel] " Kris Van Hees
  0 siblings, 1 reply; 2+ messages in thread
From: eugene.loh @ 2025-02-21 20:09 UTC (permalink / raw)
  To: dtrace, dtrace-devel

From: Eugene Loh <eugene.loh@oracle.com>

We might as well get this code right, even if this "fix" is
arguably irrelevant for two reasons:

*)  The offset just so happens to be -96 before and after the
    change anyhow, just by coincidence.

*)  The fix is on a code path that is not currently in use.

Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 libdtrace/dt_cg.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/libdtrace/dt_cg.c b/libdtrace/dt_cg.c
index 846f0fe2e..cc64a3a13 100644
--- a/libdtrace/dt_cg.c
+++ b/libdtrace/dt_cg.c
@@ -594,11 +594,11 @@ dt_cg_tramp_copy_pc_from_regs(dt_pcb_t *pcb)
 		/* test just a single byte */
 		emit(dlp,  BPF_MOV_IMM(BPF_REG_2, 1));
 
-		/* safe to write to FP+DT_STK_SP_BASE, which becomes the clause stack */
+		/* write to scratch space */
 		emit(dlp,  BPF_MOV_REG(BPF_REG_1, BPF_REG_FP));
-		emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, DT_STK_SP_BASE));
+		emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, DT_TRAMP_SP_SLOT(0)));
 
-		/* bpf_probe_read_kernel(%fp + DT_STK_SP, 1, PC) */
+		/* bpf_probe_read_kernel(%fp + DT_TRAMP_SP_SLOT(0), 1, PC) */
 		dt_regset_xalloc(drp, BPF_REG_0);
 		emit(dlp,  BPF_CALL_HELPER(BPF_FUNC_probe_read_kernel));
 
-- 
2.43.5


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [DTrace-devel] [PATCH] Use DT_TRAMP_SP_SLOT() for BPF stack scratch space in trampoline
  2025-02-21 20:09 [PATCH] Use DT_TRAMP_SP_SLOT() for BPF stack scratch space in trampoline eugene.loh
@ 2025-02-21 20:20 ` Kris Van Hees
  0 siblings, 0 replies; 2+ messages in thread
From: Kris Van Hees @ 2025-02-21 20:20 UTC (permalink / raw)
  To: eugene.loh; +Cc: dtrace, dtrace-devel

On Fri, Feb 21, 2025 at 03:09:24PM -0500, eugene.loh--- via DTrace-devel wrote:
> From: Eugene Loh <eugene.loh@oracle.com>
> 
> We might as well get this code right, even if this "fix" is
> arguably irrelevant for two reasons:
> 
> *)  The offset just so happens to be -96 before and after the
>     change anyhow, just by coincidence.
> 
> *)  The fix is on a code path that is not currently in use.
> 
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

Reviewed-by: Kris Van Hees <kris.van.hees@oracle.com>

> ---
>  libdtrace/dt_cg.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/libdtrace/dt_cg.c b/libdtrace/dt_cg.c
> index 846f0fe2e..cc64a3a13 100644
> --- a/libdtrace/dt_cg.c
> +++ b/libdtrace/dt_cg.c
> @@ -594,11 +594,11 @@ dt_cg_tramp_copy_pc_from_regs(dt_pcb_t *pcb)
>  		/* test just a single byte */
>  		emit(dlp,  BPF_MOV_IMM(BPF_REG_2, 1));
>  
> -		/* safe to write to FP+DT_STK_SP_BASE, which becomes the clause stack */
> +		/* write to scratch space */
>  		emit(dlp,  BPF_MOV_REG(BPF_REG_1, BPF_REG_FP));
> -		emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, DT_STK_SP_BASE));
> +		emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, DT_TRAMP_SP_SLOT(0)));
>  
> -		/* bpf_probe_read_kernel(%fp + DT_STK_SP, 1, PC) */
> +		/* bpf_probe_read_kernel(%fp + DT_TRAMP_SP_SLOT(0), 1, PC) */
>  		dt_regset_xalloc(drp, BPF_REG_0);
>  		emit(dlp,  BPF_CALL_HELPER(BPF_FUNC_probe_read_kernel));
>  
> -- 
> 2.43.5
> 
> 
> _______________________________________________
> DTrace-devel mailing list
> DTrace-devel@oss.oracle.com
> https://oss.oracle.com/mailman/listinfo/dtrace-devel

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2025-02-21 20:20 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-02-21 20:09 [PATCH] Use DT_TRAMP_SP_SLOT() for BPF stack scratch space in trampoline eugene.loh
2025-02-21 20:20 ` [DTrace-devel] " Kris Van Hees

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox