Re: [PATCH 2/3] powperc/mm: read TLB Block Invalidate Characteristics

[Laurent Dufour <ldufour@linux.ibm.com>]

Le 13/09/2019 à 04:00, Aneesh Kumar K.V a écrit :
> On 9/13/19 12:56 AM, Laurent Dufour wrote:
>> Le 12/09/2019 à 16:44, Aneesh Kumar K.V a écrit :
>>> Laurent Dufour <ldufour@linux.ibm.com> writes:
> 
>>>> +
>>>> +    idx = 2;
>>>> +    while (idx < len) {
>>>> +        unsigned int block_size = local_buffer[idx++];
>>>> +        unsigned int npsize;
>>>> +
>>>> +        if (!block_size)
>>>> +            break;
>>>> +
>>>> +        block_size = 1 << block_size;
>>>> +        if (block_size != 8)
>>>> +            /* We only support 8 bytes size TLB invalidate buffer */
>>>> +            pr_warn("Unsupported H_BLOCK_REMOVE block size : %d\n",
>>>> +                block_size);
>>>
>>> Should we skip setting block size if we find block_size != 8? Also can
>>> we avoid doing that pr_warn in loop and only warn if we don't find
>>> block_size 8 in the invalidate characteristics array?
>>
>> My idea here is to fully read and process the data returned by the hcall, 
>> and to put the limitation to 8 when checking before calling H_BLOCK_REMOVE.
>> The warning is there because I want it to be displayed once at boot.
>>
> 
> 
> Can we have two block size reported for the same base page size/actual page 
> size combination? If so we will overwrite the hblk[actual_psize] ?

In check_lp_set_hblk() I'm only keeping the bigger one.

> 
>>>
>>>> +
>>>> +        for (npsize = local_buffer[idx++];  npsize > 0; npsize--)
>>>> +            check_lp_set_hblk((unsigned int) local_buffer[idx++],
>>>> +                      block_size);
>>>> +    }
>>>> +
>>>> +    for (bpsize = 0; bpsize < MMU_PAGE_COUNT; bpsize++)
>>>> +        for (idx = 0; idx < MMU_PAGE_COUNT; idx++)
>>>> +            if (mmu_psize_defs[bpsize].hblk[idx])
>>>> +                pr_info("H_BLOCK_REMOVE supports base psize:%d 
>>>> psize:%d block size:%d",
>>>> +                    bpsize, idx,
>>>> +                    mmu_psize_defs[bpsize].hblk[idx]);
>>>> +
>>>> +    return 0;
>>>> +}
>>>> +machine_arch_initcall(pseries, read_tlbbi_characteristics);
>>>> +
>>>>   /*
>>>>    * Take a spinlock around flushes to avoid bouncing the hypervisor tlbie
>>>>    * lock.
> 
> -aneesh