[External] Re: [PATCH v3 2/3] riscv: cacheinfo: initialize cacheinfo's level and type from ACPI PPTT

[Jeremy Linton]

Hi,

On 4/16/24 22:15, yunhui cui wrote:
> Hi Jeremy,
> 
> On Wed, Apr 17, 2024 at 4:04 AM Jeremy Linton <jeremy.linton at arm.com> wrote:
>>
>> Hi,
>>
>>
>> On 4/15/24 22:14, Yunhui Cui wrote:
>>> Before cacheinfo can be built correctly, we need to initialize level
>>> and type. Since RSIC-V currently does not have a register group that
>>> describes cache-related attributes like ARM64, we cannot obtain them
>>> directly, so now we obtain cache leaves from the ACPI PPTT table
>>> (acpi_get_cache_info()) and set the cache type through split_levels.
>>>
>>> Suggested-by: Jeremy Linton <jeremy.linton at arm.com>
>>> Suggested-by: Sudeep Holla <sudeep.holla at arm.com>
>>> Signed-off-by: Yunhui Cui <cuiyunhui at bytedance.com>
>>> ---
>>>    arch/riscv/kernel/cacheinfo.c | 20 ++++++++++++++++++++
>>>    1 file changed, 20 insertions(+)
>>>
>>> diff --git a/arch/riscv/kernel/cacheinfo.c b/arch/riscv/kernel/cacheinfo.c
>>> index 30a6878287ad..dc5fb70362f1 100644
>>> --- a/arch/riscv/kernel/cacheinfo.c
>>> +++ b/arch/riscv/kernel/cacheinfo.c
>>> @@ -6,6 +6,7 @@
>>>    #include <linux/cpu.h>
>>>    #include <linux/of.h>
>>>    #include <asm/cacheinfo.h>
>>> +#include <linux/acpi.h>
>>>
>>>    static struct riscv_cacheinfo_ops *rv_cache_ops;
>>>
>>> @@ -78,6 +79,25 @@ int populate_cache_leaves(unsigned int cpu)
>>>        struct device_node *prev = NULL;
>>>        int levels = 1, level = 1;
>>>
>>> +     if (!acpi_disabled) {
>>> +             int ret, idx, fw_levels, split_levels;
>>> +
>>> +             ret = acpi_get_cache_info(cpu, &fw_levels, &split_levels);
>>> +             if (ret)
>>> +                     return ret;
>>> +
>>> +             for (idx = 0; level <= this_cpu_ci->num_levels &&
>>> +                  idx < this_cpu_ci->num_leaves; idx++, level++) {
>>
>> AFAIK the purpose of idx here it to assure that the number of cache
>> leaves is not overflowing. But right below we are utilizing two of them
>> at once, so this check isn't correct. OTOH, since its allocated as
>> levels + split_levels I don't think its actually possible for this to
>> cause a problem. Might be worthwhile to just hoist it before the loop
>> and revalidate the total leaves about to be utilized.
>>

I think I was suggesting something along the lines of:

BUG_ON((split_levels > fw_levels) || (split_levels + fw_levels > 
this_cpu_ci->num_leaves));

Then removing idx entirely. ex:

for (; level <= this_cpu_ci->num_levels; level++)
...
> 
> Do you mean to modify the logic as follows to make it more complete?
Sure that is one way to do it, but then you need to probably repeat the 
idx check:
> for (idx = 0; level <= this_cpu_ci->num_levels &&
>        idx < this_cpu_ci->num_leaves; level++) {
>          if (level <= split_levels) {
>                 ci_leaf_init(this_leaf++, CACHE_TYPE_DATA, level);
>                 idx++;
if (idx >= this_cpu_ci->num_leaves) break;
>                 ci_leaf_init(this_leaf++, CACHE_TYPE_INST, level);
>                 idx++;
>         } else {
>                 ci_leaf_init(this_leaf++, CACHE_TYPE_UNIFIED, level);
>                 idx++;
>        }
> }