Re: [PATCH v3 1/3] powerpc: make fadump resilient with memory add/remove events

[Sourabh Jain <sourabhjain@linux.ibm.com>]

Hi Aditya,

On 24/10/23 11:05, Aditya Gupta wrote:
> Hi sourabh, found a typo in comment.
>
> On Mon, Oct 09, 2023 at 09:49:51AM +0530, Sourabh Jain wrote:
>> +/*
>> + * Process an active dump in four steps. First, verify the crash info header
>> + * signature/magic number for integrity and accuracy. Second, if the fadump
>> + * version is greater than 0, prepare the elfcorehdr; for fadump version 0,
>> + * it's already created in the first kernel as part of the fadump reserved
>> + * area. Third, let the platform update CPU notes in elfcorehdr. Finally,
>> + * set elfcorehdr_addr so that the vmcore module can export the elfcore
>> + * header through '/proc/vmcore'.
>> + */
>> +static void process_fadump(void)
>> +{
>>
>> ...
>>
>> +     /*
>> +      * fadump version zero indicates that fadump crash info header
>> +      * is corrupted.
>> +      */
>> +     if (fadump_version < 0) {
>> +             pr_err("Crash info header is not valid.\n");
>> +             goto err_out;
>> +     }
>> +
> Here, comment says 0 indicates header is corrupted. But as per my understanding,
> version 0 means the earlier case (when MAGIC number was FADUMPINF).
> Maybe it should be:
>
> /*
>   * fadump version less than zero indicates that fadump crash info header
>   * is corrupted.
>   */
>
> Or:
>
> /*
>   * Negative fadump version indicates that fadump crash info header
>   * is corrupted.
>   */

My mistake, I'll address that comment in the next release.

I appreciate you taking a look.

By the way, version 4 is now available, and it includes a similar comment.

- Sourabh