[PATCH v4 3/5] RISC-V: fix jal addresses in patched alternatives

Jisheng Zhang jszhang at kernel.org
Wed Jan 11 09:15:27 PST 2023


On Tue, Jan 10, 2023 at 10:28:09AM +0100, Andrew Jones wrote:
> On Mon, Jan 09, 2023 at 07:17:53PM +0100, Heiko Stuebner wrote:
> > From: Heiko Stuebner <heiko.stuebner at vrull.eu>
> > 
> > Alternatives live in a different section, so addresses used by jal
> > instructions will point to wrong locations after the patch got applied.
> > 
> > Similar to arm64, adjust the location to consider that offset.
> > 
> > Signed-off-by: Heiko Stuebner <heiko.stuebner at vrull.eu>
> > ---
> >  arch/riscv/kernel/alternative.c | 27 +++++++++++++++++++++++++++
> >  1 file changed, 27 insertions(+)
> > 
> > diff --git a/arch/riscv/kernel/alternative.c b/arch/riscv/kernel/alternative.c
> > index 6212ea0eed72..985c284fe9f4 100644
> > --- a/arch/riscv/kernel/alternative.c
> > +++ b/arch/riscv/kernel/alternative.c
> > @@ -79,6 +79,21 @@ static void riscv_alternative_fix_auipc_jalr(void *ptr, u32 auipc_insn,
> >  	patch_text_nosync(ptr, call, sizeof(u32) * 2);
> >  }
> >  
> > +static void riscv_alternative_fix_jal(void *ptr, u32 jal_insn, int patch_offset)
> > +{
> > +	s32 imm;
> > +
> > +	/* get and adjust new target address */
> > +	imm = riscv_insn_extract_jtype_imm(jal_insn);
> > +	imm -= patch_offset;
> > +
> > +	/* update instructions */
> 
> instruction
> 
> > +	riscv_insn_insert_jtype_imm(&jal_insn, imm);
> > +
> > +	/* patch the call place again */
> > +	patch_text_nosync(ptr, &jal_insn, sizeof(u32));
> > +}
> > +
> >  void riscv_alternative_fix_offsets(void *alt_ptr, unsigned int len,
> >  				      int patch_offset)
> >  {
> > @@ -106,6 +121,18 @@ void riscv_alternative_fix_offsets(void *alt_ptr, unsigned int len,
> >  			riscv_alternative_fix_auipc_jalr(alt_ptr + i * sizeof(u32),
> >  							 insn, insn2, patch_offset);
> 
> I think we should add an i++ here. There's not a problem now, since we're
> only adding a fixup for jal, not jalr, but we should future-proof this and
> there's no reason to revisit an already fixed-up instruction anyway.

Hi Andrew,

IMHO, adding the i++ here belongs to the aupic jalr imm patching
commit, since the patch has been merged to riscv-next, I think
it's better to add a seperate patch to do this improvement, thus
I didn't update as commented here in my v3 series.

Thanks

> 
> >  		}
> > +
> > +		if (riscv_insn_is_jal(insn)) {
> > +			s32 imm = riscv_insn_extract_jtype_imm(insn);
> > +
> > +			/* don't modify jumps inside the alternative block */
> 
> Don't
> 
> > +			if ((alt_ptr + i * sizeof(u32) + imm) >= alt_ptr &&
> > +			    (alt_ptr + i * sizeof(u32) + imm) < (alt_ptr + len))
> > +				continue;
> > +
> > +			riscv_alternative_fix_jal(alt_ptr + i * sizeof(u32),
> > +						  insn, patch_offset);
> > +		}
> >  	}
> >  }
> >  
> > -- 
> > 2.35.1
> >
> 
> Otherwise
> 
> Reviewed-by: Andrew Jones <ajones at ventanamicro.com>
> 
> Thanks,
> drew



More information about the linux-riscv mailing list