* [PATCH] LoongArch: BPF: Fix the offset of stptr.w and stptr.d stores
@ 2026-11-18 3:02 George Guo
2026-11-18 4:14 ` Tiezhu Yang
0 siblings, 1 reply; 2+ messages in thread
From: George Guo @ 2026-11-18 3:02 UTC (permalink / raw)
To: chenhuacai, yangtiezhu
Cc: hengqi.chen, guodongtai, bpf, loongarch, linux-kernel, stable
From: George Guo <guodongtai@kylinos.cn>
When testing bpf selftests "jit_insns" and "jit_insns_blinded", the
"ldst_offsets" subtest fails for word and doubleword stores:
store_offsets:FAIL:stx size=24 base=r9 src=r0: byte at offset -16384
unexpected ... actual 255 != expected 67
...
#197/11 jit_insns_blinded/ldst_offsets:FAIL
This is because the JIT emits stptr.w or stptr.d for store offsets
that do not fit a signed 12-bit immediate but fit a signed 14-bit
immediate, and passes the byte offset directly to the instruction
encoder. However, these instructions scale their signed 14-bit
immediate by 4 when calculating the memory address. For example, a
store at offset -4096 writes at offset -16384 instead, outside the
memory range checked by the verifier.
Pass the offset in units of 4 bytes when emitting stptr.w and stptr.d
for both immediate and register stores. Use these instructions only
for offsets that are multiples of 4 in the range [-32768, 32764], and
keep the register-offset fallback for other offsets that do not fit
the signed 12-bit immediate of st.w or st.d.
With this patch, both "ldst_offsets" subtests pass:
$ sudo ./test_progs -t jit_insns -t jit_insns_blinded
...
#196/26 jit_insns/ldst_offsets:OK
...
#197/11 jit_insns_blinded/ldst_offsets:OK
Cc: stable@vger.kernel.org
Fixes: 5dc615520c4d ("LoongArch: Add BPF JIT support")
Signed-off-by: George Guo <guodongtai@kylinos.cn>
---
arch/loongarch/net/bpf_jit.c | 16 ++++++++--------
arch/loongarch/net/bpf_jit.h | 9 +++++++++
2 files changed, 17 insertions(+), 8 deletions(-)
diff --git a/arch/loongarch/net/bpf_jit.c b/arch/loongarch/net/bpf_jit.c
index 15af1abb5cb6..16e59722620f 100644
--- a/arch/loongarch/net/bpf_jit.c
+++ b/arch/loongarch/net/bpf_jit.c
@@ -1400,8 +1400,8 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, bool ext
move_imm(ctx, t1, imm);
if (is_signed_imm12(off)) {
emit_insn(ctx, stw, t1, dst, off);
- } else if (is_signed_imm14(off)) {
- emit_insn(ctx, stptrw, t1, dst, off);
+ } else if (is_ptr_off(off)) {
+ emit_insn(ctx, stptrw, t1, dst, off >> 2);
} else {
move_imm(ctx, t2, off);
emit_insn(ctx, stxw, t1, dst, t2);
@@ -1411,8 +1411,8 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, bool ext
move_imm(ctx, t1, imm);
if (is_signed_imm12(off)) {
emit_insn(ctx, std, t1, dst, off);
- } else if (is_signed_imm14(off)) {
- emit_insn(ctx, stptrd, t1, dst, off);
+ } else if (is_ptr_off(off)) {
+ emit_insn(ctx, stptrd, t1, dst, off >> 2);
} else {
move_imm(ctx, t2, off);
emit_insn(ctx, stxd, t1, dst, t2);
@@ -1460,8 +1460,8 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, bool ext
case BPF_W:
if (is_signed_imm12(off)) {
emit_insn(ctx, stw, src, dst, off);
- } else if (is_signed_imm14(off)) {
- emit_insn(ctx, stptrw, src, dst, off);
+ } else if (is_ptr_off(off)) {
+ emit_insn(ctx, stptrw, src, dst, off >> 2);
} else {
move_imm(ctx, t1, off);
emit_insn(ctx, stxw, src, dst, t1);
@@ -1470,8 +1470,8 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, bool ext
case BPF_DW:
if (is_signed_imm12(off)) {
emit_insn(ctx, std, src, dst, off);
- } else if (is_signed_imm14(off)) {
- emit_insn(ctx, stptrd, src, dst, off);
+ } else if (is_ptr_off(off)) {
+ emit_insn(ctx, stptrd, src, dst, off >> 2);
} else {
move_imm(ctx, t1, off);
emit_insn(ctx, stxd, src, dst, t1);
diff --git a/arch/loongarch/net/bpf_jit.h b/arch/loongarch/net/bpf_jit.h
index 2bd0dc5b911c..5207ad30fade 100644
--- a/arch/loongarch/net/bpf_jit.h
+++ b/arch/loongarch/net/bpf_jit.h
@@ -52,6 +52,15 @@ do { \
#define is_signed_imm52(val) signed_imm_check(val, 52)
#define is_unsigned_imm12(val) unsigned_imm_check(val, 12)
+/*
+ * ldptr.{w,d} and stptr.{w,d} take a 14-bit signed immediate in units of
+ * 4 bytes: they reach a multiple of 4 in [-32768, 32764], passed as off >> 2.
+ */
+static inline bool is_ptr_off(int off)
+{
+ return !(off & 3) && is_signed_imm16(off);
+}
+
static inline int bpf2la_offset(int bpf_insn, int off, const struct jit_ctx *ctx)
{
/* BPF JMP offset is relative to the next instruction */
base-commit: aebb7274476403040d582c61478df8f4a7bf9cd3
--
2.53.0
^ permalink raw reply related [flat|nested] 2+ messages in thread* Re: [PATCH] LoongArch: BPF: Fix the offset of stptr.w and stptr.d stores
2026-11-18 3:02 [PATCH] LoongArch: BPF: Fix the offset of stptr.w and stptr.d stores George Guo
@ 2026-11-18 4:14 ` Tiezhu Yang
0 siblings, 0 replies; 2+ messages in thread
From: Tiezhu Yang @ 2026-11-18 4:14 UTC (permalink / raw)
To: George Guo, chenhuacai
Cc: hengqi.chen, guodongtai, bpf, loongarch, linux-kernel, stable
On 11/18/26 11:02 AM, George Guo wrote:
> From: George Guo <guodongtai@kylinos.cn>
>
> When testing bpf selftests "jit_insns" and "jit_insns_blinded", the
> "ldst_offsets" subtest fails for word and doubleword stores:
...
> diff --git a/arch/loongarch/net/bpf_jit.h b/arch/loongarch/net/bpf_jit.h
> index 2bd0dc5b911c..5207ad30fade 100644
> --- a/arch/loongarch/net/bpf_jit.h
> +++ b/arch/loongarch/net/bpf_jit.h
> @@ -52,6 +52,15 @@ do { \
> #define is_signed_imm52(val) signed_imm_check(val, 52)
> #define is_unsigned_imm12(val) unsigned_imm_check(val, 12)
>
> +/*
> + * ldptr.{w,d} and stptr.{w,d} take a 14-bit signed immediate in units of
> + * 4 bytes: they reach a multiple of 4 in [-32768, 32764], passed as off >> 2.
> + */
> +static inline bool is_ptr_off(int off)
> +{
> + return !(off & 3) && is_signed_imm16(off);
> +}
> +
Please use a self-documenting function name and remove the redundant
comment:
static inline bool is_aligned_signed_imm16(int off)
{
return !(off & 3) && is_signed_imm16(off);
}
Moreover, please remove the unused is_signed_imm14 macro in bpf_jit.h.
Thanks,
Tiezhu
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-11-18 4:14 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-18 3:02 [PATCH] LoongArch: BPF: Fix the offset of stptr.w and stptr.d stores George Guo
2026-11-18 4:14 ` Tiezhu Yang
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox