2017-06-26 14:06:36

by Michal Suchánek

Subject: [PATCH 1/2] powerpc/fadump: return 0 on re-registration

When fadump is already registered return success.

Currently EEXIST is returned which is difficult to handle race-free in
userspace when shell scripts are used. If multiple writers are trying to
write '1' there is no difference in whichever succeeds so just return 0
to all.

Signed-off-by: Michal Suchanek <[email protected]>
---
arch/powerpc/kernel/fadump.c | 1 -
1 file changed, 1 deletion(-)

diff --git a/arch/powerpc/kernel/fadump.c b/arch/powerpc/kernel/fadump.c
index 436aedf195ab..5a7355381dac 100644
--- a/arch/powerpc/kernel/fadump.c
+++ b/arch/powerpc/kernel/fadump.c
@@ -1214,7 +1214,6 @@ static ssize_t fadump_register_store(struct kobject *kobj,
break;
case '1':
if (fw_dump.dump_registered == 1) {
- ret = -EEXIST;
goto unlock_out;
}
/* Register Firmware-assisted dump */
--
2.10.2

2017-06-26 14:06:46

by Michal Suchánek

Subject: [PATCH 2/2] powerpc/fadump: use kstrtoint to handle sysfs store

Currently sysfs store handlers in fadump use if buf[0] == 'char'.

This means input "100foo" is interpreted as '1' and "01" as '0'.

Change to kstrtoint so leading zeroes and the like is handled in
expected way.

Signed-off-by: Michal Suchanek <[email protected]>
---
arch/powerpc/kernel/fadump.c | 17 +++++++++++++----
1 file changed, 13 insertions(+), 4 deletions(-)

diff --git a/arch/powerpc/kernel/fadump.c b/arch/powerpc/kernel/fadump.c
index 5a7355381dac..241eff0b5f76 100644
--- a/arch/powerpc/kernel/fadump.c
+++ b/arch/powerpc/kernel/fadump.c
@@ -1161,10 +1161,15 @@ static ssize_t fadump_release_memory_store(struct kobject *kobj,
struct kobj_attribute *attr,
const char *buf, size_t count)
{
+ int input = -1;
+
if (!fw_dump.dump_active)
return -EPERM;

- if (buf[0] == '1') {
+ if (kstrtoint(buf, 0, &input))
+ return -EINVAL;
+
+ if (input == 1) {
/*
* Take away the '/proc/vmcore'. We are releasing the dump
* memory, hence it will not be valid anymore.
@@ -1198,21 +1203,25 @@ static ssize_t fadump_register_store(struct kobject *kobj,
const char *buf, size_t count)
{
int ret = 0;
+ int input = -1;

if (!fw_dump.fadump_enabled || fdm_active)
return -EPERM;

+ if (kstrtoint(buf, 0, &input))
+ return -EINVAL;
+
mutex_lock(&fadump_mutex);

- switch (buf[0]) {
- case '0':
+ switch (input) {
+ case 0:
if (fw_dump.dump_registered == 0) {
goto unlock_out;
}
/* Un-register Firmware-assisted dump */
fadump_unregister_dump(&fdm);
break;
- case '1':
+ case 1:
if (fw_dump.dump_registered == 1) {
goto unlock_out;
}
--
2.10.2

2017-06-27 09:31:14

by Michal Suchánek

Subject: Re: [PATCH 1/2] powerpc/fadump: return 0 on re-registration

On Mon, 26 Jun 2017 16:06:00 +0200
Michal Suchanek <[email protected]> wrote:

> When fadump is already registered return success.
>
> Currently EEXIST is returned which is difficult to handle race-free in
> userspace when shell scripts are used. If multiple writers are trying
> to write '1' there is no difference in whichever succeeds so just
> return 0 to all.
>
> Signed-off-by: Michal Suchanek <[email protected]>
> ---
> arch/powerpc/kernel/fadump.c | 1 -
> 1 file changed, 1 deletion(-)
>
> diff --git a/arch/powerpc/kernel/fadump.c
> b/arch/powerpc/kernel/fadump.c index 436aedf195ab..5a7355381dac 100644
> --- a/arch/powerpc/kernel/fadump.c
> +++ b/arch/powerpc/kernel/fadump.c
> @@ -1214,7 +1214,6 @@ static ssize_t fadump_register_store(struct
> kobject *kobj, break;
> case '1':
> if (fw_dump.dump_registered == 1) {
> - ret = -EEXIST;
> goto unlock_out;
> }
> /* Register Firmware-assisted dump */

Forget about this one.

It breaks another case when fadump is registered and you need to
re-register to account for change in system configuration.

Thanks

Michal

2017-11-14 12:01:23

by Michael Ellerman

Subject: Re: [2/2] powerpc/fadump: use kstrtoint to handle sysfs store

On Mon, 2017-06-26 at 14:06:01 UTC, Michal Suchanek wrote:
> Currently sysfs store handlers in fadump use if buf[0] == 'char'.
>
> This means input "100foo" is interpreted as '1' and "01" as '0'.
>
> Change to kstrtoint so leading zeroes and the like is handled in
> expected way.
>
> Signed-off-by: Michal Suchanek <[email protected]>
> Acked-by: Hari Bathini <[email protected]>
> Signed-off-by: Michal Suchanek <a class="moz-txt-link-rfc2396E" href="mailto:[email protected]">&lt;[email protected]&gt;</a></pre>

Applied to powerpc next, thanks.

https://git.kernel.org/powerpc/c/dcdc46794b7bb76733d9792cca2f45

cheers

From 1571276522868990095@xxx Mon Jun 26 14:07:43 +0000 2017
X-GM-THRID: 1571276522868990095
X-Gmail-Labels: Inbox,Category Forums