2022-05-11 09:47:23

by Shreyas K K

Subject: [PATCH] arm64: Enable repeat tlbi workaround on KRYO4XX gold CPUs

Add KRYO4XX gold/big cores to the list of CPUs that need the
repeat TLBI workaround. Apply this to the affected
KRYO4XX cores (rcpe to rdpe).

The variant and revision bits are implementation defined and are
different from the their Cortex CPU counterparts on which they are
based on, i.e., (r0p0 to r1p0) is equivalent to (rcpe to rdpe).

Signed-off-by: Shreyas K K <[email protected]>
---
Documentation/arm64/silicon-errata.rst | 3 +++
arch/arm64/kernel/cpu_errata.c | 2 ++
2 files changed, 5 insertions(+)

diff --git a/Documentation/arm64/silicon-errata.rst b/Documentation/arm64/silicon-errata.rst
index 466cb9e89047..d27db84d585e 100644
--- a/Documentation/arm64/silicon-errata.rst
+++ b/Documentation/arm64/silicon-errata.rst
@@ -189,6 +189,9 @@ stable kernels.
+----------------+-----------------+-----------------+-----------------------------+
| Qualcomm Tech. | Kryo4xx Silver | N/A | ARM64_ERRATUM_1024718 |
+----------------+-----------------+-----------------+-----------------------------+
+| Qualcomm Tech. | Kryo4xx Gold | N/A | ARM64_ERRATUM_1286807 |
++----------------+-----------------+-----------------+-----------------------------+
+
+----------------+-----------------+-----------------+-----------------------------+
| Fujitsu | A64FX | E#010001 | FUJITSU_ERRATUM_010001 |
+----------------+-----------------+-----------------+-----------------------------+
diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index 4c9b5b4b7a0b..2518657e6de1 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -208,6 +208,8 @@ static const struct arm64_cpu_capabilities arm64_repeat_tlbi_list[] = {
#ifdef CONFIG_ARM64_ERRATUM_1286807
{
ERRATA_MIDR_RANGE(MIDR_CORTEX_A76, 0, 0, 3, 0),
+ /* Kryo4xx Gold (rcpe to rdpe) => (r0p0 to r1p0) */
+ ERRATA_MIDR_RANGE(QCOM_CPU_PART_KRYO_4XX_GOLD, 0xc, 0xe, 0xd, 0xe),
},
#endif
{},
--
2.17.1

2022-05-11 18:33:21

by Sai Prakash Ranjan

Subject: Re: [PATCH] arm64: Enable repeat tlbi workaround on KRYO4XX gold CPUs

Hi Shreyas,

On 5/11/2022 1:32 PM, Shreyas K K wrote:
> Add KRYO4XX gold/big cores to the list of CPUs that need the
> repeat TLBI workaround. Apply this to the affected
> KRYO4XX cores (rcpe to rdpe).
>
> The variant and revision bits are implementation defined and are
> different from the their Cortex CPU counterparts on which they are
> based on, i.e., (r0p0 to r1p0) is equivalent to (rcpe to rdpe).
>
> Signed-off-by: Shreyas K K <[email protected]>
> ---
> Documentation/arm64/silicon-errata.rst | 3 +++
> arch/arm64/kernel/cpu_errata.c | 2 ++
> 2 files changed, 5 insertions(+)
>
> diff --git a/Documentation/arm64/silicon-errata.rst b/Documentation/arm64/silicon-errata.rst
> index 466cb9e89047..d27db84d585e 100644
> --- a/Documentation/arm64/silicon-errata.rst
> +++ b/Documentation/arm64/silicon-errata.rst
> @@ -189,6 +189,9 @@ stable kernels.
> +----------------+-----------------+-----------------+-----------------------------+
> | Qualcomm Tech. | Kryo4xx Silver | N/A | ARM64_ERRATUM_1024718 |
> +----------------+-----------------+-----------------+-----------------------------+
> +| Qualcomm Tech. | Kryo4xx Gold | N/A | ARM64_ERRATUM_1286807 |
> ++----------------+-----------------+-----------------+-----------------------------+
> +
> +----------------+-----------------+-----------------+-----------------------------+
> | Fujitsu | A64FX | E#010001 | FUJITSU_ERRATUM_010001 |
> +----------------+-----------------+-----------------+-----------------------------+
> diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
> index 4c9b5b4b7a0b..2518657e6de1 100644
> --- a/arch/arm64/kernel/cpu_errata.c
> +++ b/arch/arm64/kernel/cpu_errata.c
> @@ -208,6 +208,8 @@ static const struct arm64_cpu_capabilities arm64_repeat_tlbi_list[] = {
> #ifdef CONFIG_ARM64_ERRATUM_1286807
> {
> ERRATA_MIDR_RANGE(MIDR_CORTEX_A76, 0, 0, 3, 0),
> + /* Kryo4xx Gold (rcpe to rdpe) => (r0p0 to r1p0) */
> + ERRATA_MIDR_RANGE(QCOM_CPU_PART_KRYO_4XX_GOLD, 0xc, 0xe, 0xd, 0xe),
> },
> #endif
> {},

Why not include r2p0 and r3p0 which are affected by this erratum? I see these revisions are present
in our SoCs as per the document.

Thanks,
Sai

2022-05-12 15:06:05

by Shreyas K K

Subject: Re: [PATCH] arm64: Enable repeat tlbi workaround on KRYO4XX gold CPUs

Hi Sai,

On 5/11/2022 4:43 PM, Sai Prakash Ranjan wrote:
> Hi Shreyas,
>
> On 5/11/2022 1:32 PM, Shreyas K K wrote:
>> Add KRYO4XX gold/big cores to the list of CPUs that need the
>> repeat TLBI workaround. Apply this to the affected
>> KRYO4XX cores (rcpe to rdpe).
>>
>> The variant and revision bits are implementation defined and are
>> different from the their Cortex CPU counterparts on which they are
>> based on, i.e., (r0p0 to r1p0) is equivalent to (rcpe to rdpe).
>>
>> Signed-off-by: Shreyas K K <[email protected]>
>> ---
>>   Documentation/arm64/silicon-errata.rst | 3 +++
>>   arch/arm64/kernel/cpu_errata.c         | 2 ++
>>   2 files changed, 5 insertions(+)
>>
>> diff --git a/Documentation/arm64/silicon-errata.rst
>> b/Documentation/arm64/silicon-errata.rst
>> index 466cb9e89047..d27db84d585e 100644
>> --- a/Documentation/arm64/silicon-errata.rst
>> +++ b/Documentation/arm64/silicon-errata.rst
>> @@ -189,6 +189,9 @@ stable kernels.
>> +----------------+-----------------+-----------------+-----------------------------+
>>   | Qualcomm Tech. | Kryo4xx Silver  | N/A             |
>> ARM64_ERRATUM_1024718       |
>> +----------------+-----------------+-----------------+-----------------------------+
>> +| Qualcomm Tech. | Kryo4xx Gold    | N/A             |
>> ARM64_ERRATUM_1286807       |
>> ++----------------+-----------------+-----------------+-----------------------------+
>>
>> +
>> +----------------+-----------------+-----------------+-----------------------------+
>>   | Fujitsu        | A64FX           | E#010001        |
>> FUJITSU_ERRATUM_010001      |
>> +----------------+-----------------+-----------------+-----------------------------+
>> diff --git a/arch/arm64/kernel/cpu_errata.c
>> b/arch/arm64/kernel/cpu_errata.c
>> index 4c9b5b4b7a0b..2518657e6de1 100644
>> --- a/arch/arm64/kernel/cpu_errata.c
>> +++ b/arch/arm64/kernel/cpu_errata.c
>> @@ -208,6 +208,8 @@ static const struct arm64_cpu_capabilities
>> arm64_repeat_tlbi_list[] = {
>>   #ifdef CONFIG_ARM64_ERRATUM_1286807
>>       {
>>           ERRATA_MIDR_RANGE(MIDR_CORTEX_A76, 0, 0, 3, 0),
>> +        /* Kryo4xx Gold (rcpe to rdpe) => (r0p0 to r1p0) */
>> +        ERRATA_MIDR_RANGE(QCOM_CPU_PART_KRYO_4XX_GOLD, 0xc, 0xe,
>> 0xd, 0xe),
>>       },
>>   #endif
>>       {},
>
> Why not include r2p0 and r3p0 which are affected by this erratum? I
> see these revisions are present
> in our SoCs as per the document.
>
> Thanks,
> Sai

Acknowledged. Changing in the next version.

Thanks,
Shreyas