[PATCH] doc: Fix some typo errors in ras.rst

[PATCH] doc: Fix some typo errors in ras.rst

[Waiman Long]

Fix typo errors.

Signed-off-by: Waiman Long <longman@redhat.com>
---
 Documentation/admin-guide/ras.rst | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/Documentation/admin-guide/ras.rst b/Documentation/admin-guide/ras.rst
index 0310db624964..8b1803b2606f 100644
--- a/Documentation/admin-guide/ras.rst
+++ b/Documentation/admin-guide/ras.rst
@@ -156,10 +156,10 @@ the labels provided by the BIOS won't match the real ones.
 ECC memory
 ----------
 
-As mentioned on the previous section, ECC memory has extra bits to be
-used for error correction. So, on 64 bit systems, a memory module
-has 64 bits of *data width*, and 74 bits of *total width*. So, there are
-8 bits extra bits to be used for the error detection and correction
+As mentioned on the previous section, ECC memory has extra bits to
+be used for error correction. So, on 64 bit systems, a memory module
+has 64 bits of *data width*, and 72 bits of *total width*. So, there
+are 8 extra bits to be used for the error detection and correction
 mechanisms. Those extra bits are called *syndrome*\ [#f1]_\ [#f2]_.
 
 So, when the cpu requests the memory controller to write a word with
-- 
2.18.1

Re: [PATCH] doc: Fix some typo errors in ras.rst

[Jonathan Corbet]

On Tue,  5 May 2020 11:10:49 -0400
Waiman Long <longman@redhat.com> wrote:

> Fix typo errors.
> 
> Signed-off-by: Waiman Long <longman@redhat.com>
> ---
>  Documentation/admin-guide/ras.rst | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/Documentation/admin-guide/ras.rst b/Documentation/admin-guide/ras.rst
> index 0310db624964..8b1803b2606f 100644
> --- a/Documentation/admin-guide/ras.rst
> +++ b/Documentation/admin-guide/ras.rst
> @@ -156,10 +156,10 @@ the labels provided by the BIOS won't match the real ones.
>  ECC memory
>  ----------
>  
> -As mentioned on the previous section, ECC memory has extra bits to be
> -used for error correction. So, on 64 bit systems, a memory module
> -has 64 bits of *data width*, and 74 bits of *total width*. So, there are
> -8 bits extra bits to be used for the error detection and correction
> +As mentioned on the previous section, ECC memory has extra bits to
> +be used for error correction. So, on 64 bit systems, a memory module
> +has 64 bits of *data width*, and 72 bits of *total width*. So, there
> +are 8 extra bits to be used for the error detection and correction
>  mechanisms. Those extra bits are called *syndrome*\ [#f1]_\ [#f2]_.

So I had to work to figure out what the change was, since you didn't say
in the changelog and you refilled the paragraph.  But this looks like a
*factual* error, giving the wrong number of bits, right?  It seems like
the changelog should say that.  Do the people who know about this stuff
agree that the change is correct?

Thanks,

jon

Re: [PATCH] doc: Fix some typo errors in ras.rst

[Matthew Wilcox]

On Tue, May 05, 2020 at 11:10:49AM -0400, Waiman Long wrote:
> Fix typo errors.

By reformatting it, you've successfully obscured what typos you've fixed.
As a result I read the whole paragraph, and ...

>  ECC memory
>  ----------
>  
> -As mentioned on the previous section, ECC memory has extra bits to be
> -used for error correction. So, on 64 bit systems, a memory module
> -has 64 bits of *data width*, and 74 bits of *total width*. So, there are
> -8 bits extra bits to be used for the error detection and correction
> +As mentioned on the previous section, ECC memory has extra bits to

s/on/in/

> +be used for error correction. So, on 64 bit systems, a memory module
> +has 64 bits of *data width*, and 72 bits of *total width*.

Usually a 64-bit system refers to the width of a pointer.  Here, it's
referring to the width of the memory system, which is rather confusing.
How about "In the above example" instead of "So, on 64 bit systems".

> So, there
> +are 8 extra bits to be used for the error detection and correction
>  mechanisms. Those extra bits are called *syndrome*\ [#f1]_\ [#f2]_.

This would read better as:

The extra 8 bits which are used for error detection and correction
are referred to as the *syndrome*\ [#f1]_\ [#f2]_.

Re: [PATCH] doc: Fix some typo errors in ras.rst

[Waiman Long]

On 5/5/20 11:32 AM, Jonathan Corbet wrote:
> On Tue,  5 May 2020 11:10:49 -0400
> Waiman Long <longman@redhat.com> wrote:
>
>> Fix typo errors.
>>
>> Signed-off-by: Waiman Long <longman@redhat.com>
>> ---
>>   Documentation/admin-guide/ras.rst | 8 ++++----
>>   1 file changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/Documentation/admin-guide/ras.rst b/Documentation/admin-guide/ras.rst
>> index 0310db624964..8b1803b2606f 100644
>> --- a/Documentation/admin-guide/ras.rst
>> +++ b/Documentation/admin-guide/ras.rst
>> @@ -156,10 +156,10 @@ the labels provided by the BIOS won't match the real ones.
>>   ECC memory
>>   ----------
>>   
>> -As mentioned on the previous section, ECC memory has extra bits to be
>> -used for error correction. So, on 64 bit systems, a memory module
>> -has 64 bits of *data width*, and 74 bits of *total width*. So, there are
>> -8 bits extra bits to be used for the error detection and correction
>> +As mentioned on the previous section, ECC memory has extra bits to
>> +be used for error correction. So, on 64 bit systems, a memory module
>> +has 64 bits of *data width*, and 72 bits of *total width*. So, there
>> +are 8 extra bits to be used for the error detection and correction
>>   mechanisms. Those extra bits are called *syndrome*\ [#f1]_\ [#f2]_.
> So I had to work to figure out what the change was, since you didn't say
> in the changelog and you refilled the paragraph.  But this looks like a
> *factual* error, giving the wrong number of bits, right?  It seems like
> the changelog should say that.  Do the people who know about this stuff
> agree that the change is correct?
>
> Thanks,
>
> jon
>
Yes, I change 74 bits to 72 bits and remove an extra "bits". Will send 
out v2 to clarify that.

Thanks,
Longman

Re: [PATCH] doc: Fix some typo errors in ras.rst

[Borislav Petkov]

On Tue, May 05, 2020 at 08:48:16AM -0700, Matthew Wilcox wrote:
> Usually a 64-bit system refers to the width of a pointer.  Here, it's
> referring to the width of the memory system, which is rather confusing.
> How about "In the above example" instead of "So, on 64 bit systems".

Yes, that should not talk about 64-bit systems but about the length of
the word the memory controller uses to create the ECC check bits out of.

That whole doc needs checking/fixing.

> The extra 8 bits which are used for error detection and correction
> are referred to as the *syndrome*\ [#f1]_\ [#f2]_.

Well, I know it as "The syndrome field uniquely identifies the failing
bit positions of a correctable ECC error." and depending on the error
severity you can have correctable and uncorrectable syndromes.

I don't think the syndrome is the ECC word but it might be, depending
on the ECC algorithm used or it might be that what I pasted above. That
would need deeper digging.

Thx.

-- 
Regards/Gruss,
    Boris.

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

Re: [PATCH] doc: Fix some typo errors in ras.rst

[Waiman Long]

On 5/5/20 11:48 AM, Matthew Wilcox wrote:
> On Tue, May 05, 2020 at 11:10:49AM -0400, Waiman Long wrote:
>> Fix typo errors.
> By reformatting it, you've successfully obscured what typos you've fixed.
> As a result I read the whole paragraph, and ...
>
>>   ECC memory
>>   ----------
>>   
>> -As mentioned on the previous section, ECC memory has extra bits to be
>> -used for error correction. So, on 64 bit systems, a memory module
>> -has 64 bits of *data width*, and 74 bits of *total width*. So, there are
>> -8 bits extra bits to be used for the error detection and correction
>> +As mentioned on the previous section, ECC memory has extra bits to
> s/on/in/
>
>> +be used for error correction. So, on 64 bit systems, a memory module
>> +has 64 bits of *data width*, and 72 bits of *total width*.
> Usually a 64-bit system refers to the width of a pointer.  Here, it's
> referring to the width of the memory system, which is rather confusing.
> How about "In the above example" instead of "So, on 64 bit systems".
>
>> So, there
>> +are 8 extra bits to be used for the error detection and correction
>>   mechanisms. Those extra bits are called *syndrome*\ [#f1]_\ [#f2]_.
> This would read better as:
>
> The extra 8 bits which are used for error detection and correction
> are referred to as the *syndrome*\ [#f1]_\ [#f2]_.
>
Thanks for the suggestion. Will incorporate that in v2.

-Longman

Re: [PATCH] doc: Fix some typo errors in ras.rst

[Waiman Long]

On 5/5/20 12:03 PM, Borislav Petkov wrote:
> On Tue, May 05, 2020 at 08:48:16AM -0700, Matthew Wilcox wrote:
>> Usually a 64-bit system refers to the width of a pointer.  Here, it's
>> referring to the width of the memory system, which is rather confusing.
>> How about "In the above example" instead of "So, on 64 bit systems".
> Yes, that should not talk about 64-bit systems but about the length of
> the word the memory controller uses to create the ECC check bits out of.
>
> That whole doc needs checking/fixing.

I am not knowledgeable enough to check/fix the whole document. This 
patch is motivated primarily by the incorrect "74 bits" total width that 
I saw. I will send a v2 to make the suggested edit. I will leave it to 
you to do the full document check.

Cheers,
Longman