2018-06-07 10:42:02

by Andreas Schwab

Subject: [PATCH] RISC-V: Handle R_RISCV_32 in modules

With CONFIG_MODVERSIONS=y the R_RISCV_32 relocation is used by the
__kcrctab section.

Signed-off-by: Andreas Schwab <[email protected]>
---
arch/riscv/kernel/module.c | 12 ++++++++++++
1 file changed, 12 insertions(+)

diff --git a/arch/riscv/kernel/module.c b/arch/riscv/kernel/module.c
index 5dddba301d..1d5e9b934b 100644
--- a/arch/riscv/kernel/module.c
+++ b/arch/riscv/kernel/module.c
@@ -17,6 +17,17 @@
#include <linux/errno.h>
#include <linux/moduleloader.h>

+static int apply_r_riscv_32_rela(struct module *me, u32 *location, Elf_Addr v)
+{
+ if (v != (u32)v) {
+ pr_err("%s: value %016llx out of range for 32-bit field\n",
+ me->name, v);
+ return -EINVAL;
+ }
+ *location = v;
+ return 0;
+}
+
static int apply_r_riscv_64_rela(struct module *me, u32 *location, Elf_Addr v)
{
*(u64 *)location = v;
@@ -265,6 +276,7 @@ static int apply_r_riscv_sub32_rela(struct module *me, u32 *location,

static int (*reloc_handlers_rela[]) (struct module *me, u32 *location,
Elf_Addr v) = {
+ [R_RISCV_32] = apply_r_riscv_32_rela,
[R_RISCV_64] = apply_r_riscv_64_rela,
[R_RISCV_BRANCH] = apply_r_riscv_branch_rela,
[R_RISCV_JAL] = apply_r_riscv_jal_rela,
--
2.17.1


--
Andreas Schwab, SUSE Labs, [email protected]
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

2018-06-08 22:36:25

by Palmer Dabbelt

Subject: Re: [PATCH] RISC-V: Handle R_RISCV_32 in modules

On Thu, 07 Jun 2018 03:27:27 PDT (-0700), [email protected] wrote:
> With CONFIG_MODVERSIONS=y the R_RISCV_32 relocation is used by the
> __kcrctab section.
>
> Signed-off-by: Andreas Schwab <[email protected]>
> ---
> arch/riscv/kernel/module.c | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/arch/riscv/kernel/module.c b/arch/riscv/kernel/module.c
> index 5dddba301d..1d5e9b934b 100644
> --- a/arch/riscv/kernel/module.c
> +++ b/arch/riscv/kernel/module.c
> @@ -17,6 +17,17 @@
> #include <linux/errno.h>
> #include <linux/moduleloader.h>
>
> +static int apply_r_riscv_32_rela(struct module *me, u32 *location, Elf_Addr v)
> +{
> + if (v != (u32)v) {

My worry with this kind of check is that it relies on some sort of undefined
behavior in C and that at some point in the future GCC will just go decide the
check can never fail. I checked and GCC doesn't elide these checks now, so I
might be wrong.

Is this defined to do what it looks like it's doing?

> + pr_err("%s: value %016llx out of range for 32-bit field\n",
> + me->name, v);
> + return -EINVAL;
> + }
> + *location = v;
> + return 0;
> +}
> +
> static int apply_r_riscv_64_rela(struct module *me, u32 *location, Elf_Addr v)
> {
> *(u64 *)location = v;
> @@ -265,6 +276,7 @@ static int apply_r_riscv_sub32_rela(struct module *me, u32 *location,
>
> static int (*reloc_handlers_rela[]) (struct module *me, u32 *location,
> Elf_Addr v) = {
> + [R_RISCV_32] = apply_r_riscv_32_rela,
> [R_RISCV_64] = apply_r_riscv_64_rela,
> [R_RISCV_BRANCH] = apply_r_riscv_branch_rela,
> [R_RISCV_JAL] = apply_r_riscv_jal_rela,

Thanks!

2018-06-11 07:54:43

by Andreas Schwab

Subject: Re: [PATCH] RISC-V: Handle R_RISCV_32 in modules

On Jun 08 2018, Palmer Dabbelt <[email protected]> wrote:

> On Thu, 07 Jun 2018 03:27:27 PDT (-0700), [email protected] wrote:
>> With CONFIG_MODVERSIONS=y the R_RISCV_32 relocation is used by the
>> __kcrctab section.
>>
>> Signed-off-by: Andreas Schwab <[email protected]>
>> ---
>> arch/riscv/kernel/module.c | 12 ++++++++++++
>> 1 file changed, 12 insertions(+)
>>
>> diff --git a/arch/riscv/kernel/module.c b/arch/riscv/kernel/module.c
>> index 5dddba301d..1d5e9b934b 100644
>> --- a/arch/riscv/kernel/module.c
>> +++ b/arch/riscv/kernel/module.c
>> @@ -17,6 +17,17 @@
>> #include <linux/errno.h>
>> #include <linux/moduleloader.h>
>>
>> +static int apply_r_riscv_32_rela(struct module *me, u32 *location, Elf_Addr v)
>> +{
>> + if (v != (u32)v) {
>
> My worry with this kind of check is that it relies on some sort of
> undefined behavior in C and that at some point in the future GCC will just
> go decide the check can never fail. I checked and GCC doesn't elide these
> checks now, so I might be wrong.
>
> Is this defined to do what it looks like it's doing?

This is unsigned arithmetic, thus fully defined.

Andreas.

--
Andreas Schwab, SUSE Labs, [email protected]
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

2018-06-11 20:58:52

by Palmer Dabbelt

Subject: Re: [PATCH] RISC-V: Handle R_RISCV_32 in modules

On Mon, 11 Jun 2018 00:54:00 PDT (-0700), [email protected] wrote:
> On Jun 08 2018, Palmer Dabbelt <[email protected]> wrote:
>
>> On Thu, 07 Jun 2018 03:27:27 PDT (-0700), [email protected] wrote:
>>> With CONFIG_MODVERSIONS=y the R_RISCV_32 relocation is used by the
>>> __kcrctab section.
>>>
>>> Signed-off-by: Andreas Schwab <[email protected]>
>>> ---
>>> arch/riscv/kernel/module.c | 12 ++++++++++++
>>> 1 file changed, 12 insertions(+)
>>>
>>> diff --git a/arch/riscv/kernel/module.c b/arch/riscv/kernel/module.c
>>> index 5dddba301d..1d5e9b934b 100644
>>> --- a/arch/riscv/kernel/module.c
>>> +++ b/arch/riscv/kernel/module.c
>>> @@ -17,6 +17,17 @@
>>> #include <linux/errno.h>
>>> #include <linux/moduleloader.h>
>>>
>>> +static int apply_r_riscv_32_rela(struct module *me, u32 *location, Elf_Addr v)
>>> +{
>>> + if (v != (u32)v) {
>>
>> My worry with this kind of check is that it relies on some sort of
>> undefined behavior in C and that at some point in the future GCC will just
>> go decide the check can never fail. I checked and GCC doesn't elide these
>> checks now, so I might be wrong.
>>
>> Is this defined to do what it looks like it's doing?
>
> This is unsigned arithmetic, thus fully defined.

Great, thanks! I guess I'm just a bit paranoid :)