2021-10-03 18:43:55

by Eric Badger

Subject: [PATCH] EDAC/mc_sysfs: Print MC-scope sysfs counters unsigned

This is cosmetically nicer for counts > INT32_MAX, and aligns the
MC-scope format with that of the lower layer sysfs counter files.

Signed-off-by: Eric Badger <[email protected]>
---
drivers/edac/edac_mc_sysfs.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c
index 2f9f1e7..0a638c9 100644
--- a/drivers/edac/edac_mc_sysfs.c
+++ b/drivers/edac/edac_mc_sysfs.c
@@ -744,7 +744,7 @@ static ssize_t mci_ue_count_show(struct device *dev,
{
struct mem_ctl_info *mci = to_mci(dev);

- return sprintf(data, "%d\n", mci->ue_mc);
+ return sprintf(data, "%u\n", mci->ue_mc);
}

static ssize_t mci_ce_count_show(struct device *dev,
@@ -753,7 +753,7 @@ static ssize_t mci_ce_count_show(struct device *dev,
{
struct mem_ctl_info *mci = to_mci(dev);

- return sprintf(data, "%d\n", mci->ce_mc);
+ return sprintf(data, "%u\n", mci->ce_mc);
}

static ssize_t mci_ce_noinfo_show(struct device *dev,
@@ -762,7 +762,7 @@ static ssize_t mci_ce_noinfo_show(struct device *dev,
{
struct mem_ctl_info *mci = to_mci(dev);

- return sprintf(data, "%d\n", mci->ce_noinfo_count);
+ return sprintf(data, "%u\n", mci->ce_noinfo_count);
}

static ssize_t mci_ue_noinfo_show(struct device *dev,
@@ -771,7 +771,7 @@ static ssize_t mci_ue_noinfo_show(struct device *dev,
{
struct mem_ctl_info *mci = to_mci(dev);

- return sprintf(data, "%d\n", mci->ue_noinfo_count);
+ return sprintf(data, "%u\n", mci->ue_noinfo_count);
}

static ssize_t mci_seconds_show(struct device *dev,
--
1.9.1

2021-10-04 22:26:24

by Luck, Tony

Subject: RE: [PATCH] EDAC/mc_sysfs: Print MC-scope sysfs counters unsigned

> This is cosmetically nicer for counts > INT32_MAX, and aligns the
> MC-scope format with that of the lower layer sysfs counter files.

While this is technically the right thing to do, I pity the system administrator that
is looking at a system with more than 2147483647 corrected or uncorrected errors!

So:

Acked-by: Tony Luck <[email protected]>

but maybe this is just churn and not really useful in practice?

-Tony

2021-10-04 22:45:01

by Eric Badger

Subject: Re: [PATCH] EDAC/mc_sysfs: Print MC-scope sysfs counters unsigned

On Mon, Oct 04, 2021 at 03:21:13PM +0000, Luck, Tony wrote:
> > This is cosmetically nicer for counts > INT32_MAX, and aligns the
> > MC-scope format with that of the lower layer sysfs counter files.
>
> While this is technically the right thing to do, I pity the system administrator that
> is looking at a system with more than 2147483647 corrected or uncorrected errors!
>
> So:
>
> Acked-by: Tony Luck <[email protected]>
>
> but maybe this is just churn and not really useful in practice?

Pity accepted :). I only noticed the sign mismatch after seeing a
negative value on a server in the wild. But it's cosmetic really; if
you've reached INT32_MAX you'll probably reach UINT32_MAX and can't rely
on the counter.

Cheers,
Eric

2021-10-07 10:26:47

by Borislav Petkov

Subject: Re: [PATCH] EDAC/mc_sysfs: Print MC-scope sysfs counters unsigned

On Sun, Oct 03, 2021 at 11:16:53AM -0700, Eric Badger wrote:
> This is cosmetically nicer for counts > INT32_MAX, and aligns the
> MC-scope format with that of the lower layer sysfs counter files.
>
> Signed-off-by: Eric Badger <[email protected]>
> ---
> drivers/edac/edac_mc_sysfs.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)

Applied, thanks.

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette