| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <57e94a65-d9df-4114-bcee-998addb6e60f@huawei.com>
Date: Mon, 9 Dec 2024 15:49:37 +0800
From: "zhenglifeng (A)" <zhenglifeng1@...wei.com>
To: Pierre Gondois <pierre.gondois@....com>, <rafael@...nel.org>,
<lenb@...nel.org>, <robert.moore@...el.com>, <viresh.kumar@...aro.org>
CC: <acpica-devel@...ts.linux.dev>, <linux-acpi@...r.kernel.org>,
<linux-kernel@...r.kernel.org>, <linux-pm@...r.kernel.org>,
<zhanjie9@...ilicon.com>, <lihuisong@...wei.com>, <fanghao11@...wei.com>,
"zhenglifeng (A)" <zhenglifeng1@...wei.com>
Subject: Re: [PATCH 1/3] ACPI: CPPC: Refactor register get and set ABIs
Hello Pierre，
On 2024/12/6 22:23, Pierre Gondois wrote:
> Hello Lifeng,
>
> On 11/14/24 09:48, Lifeng Zheng wrote:
>> Refactor register get and set ABIs using cppc_get_reg() and cppc_set_reg().
>>
>> Rename cppc_get_perf() to cppc_get_reg() as a generic function to read cppc
>> registers, with two changes:
>>
>> 1. Change the error kind to "no such device" when pcc_ss_id < 0, which
>> means that this cpu cannot get a valid pcc_ss_id.
>>
>> 2. Add a check to verify if the register is a cpc supported one before
>> using it.
>>
>> Add cppc_set_reg() as a generic function for setting cppc registers. Unlike
>> other set reg ABIs, this function checks CPC_SUPPORTED right after getting
>> the register, because the rest of the operations are meaningless if this
>> register is not a cpc supported one.
>>
>> Signed-off-by: Lifeng Zheng <zhenglifeng1@...wei.com>
>> ---
>> drivers/acpi/cppc_acpi.c | 191 +++++++++++++++------------------------
>> 1 file changed, 72 insertions(+), 119 deletions(-)
>>
>> diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c
>> index c1f3568d0c50..306ced9c3376 100644
>> --- a/drivers/acpi/cppc_acpi.c
>> +++ b/drivers/acpi/cppc_acpi.c
>> @@ -1179,10 +1179,13 @@ static int cpc_write(int cpu, struct cpc_register_resource *reg_res, u64 val)
>> return ret_val;
>> }
>> -static int cppc_get_perf(int cpunum, enum cppc_regs reg_idx, u64 *perf)
>> +static int cppc_get_reg(int cpunum, enum cppc_regs reg_idx, u64 *val)
>> {
>> struct cpc_desc *cpc_desc = per_cpu(cpc_desc_ptr, cpunum);
>> + struct cppc_pcc_data *pcc_ss_data = NULL;
>> struct cpc_register_resource *reg;
>> + int pcc_ss_id;
>> + int ret = 0;
>
> NIT: Might not be necessary if we save the value returned by cpc_read(),
> cf. other comment below.
>
>> if (!cpc_desc) {
>> pr_debug("No CPC descriptor for CPU:%d\n", cpunum);
>> @@ -1191,20 +1194,23 @@ static int cppc_get_perf(int cpunum, enum cppc_regs reg_idx, u64 *perf)
>> reg = &cpc_desc->cpc_regs[reg_idx];
>> + if (!CPC_SUPPORTED(reg)) {
>> + pr_debug("CPC register (reg_idx=%u) is not supported\n", reg_idx);
>> + return -EOPNOTSUPP;
>> + }
>> +
>> if (CPC_IN_PCC(reg)) {
>> - int pcc_ss_id = per_cpu(cpu_pcc_subspace_idx, cpunum);
>> - struct cppc_pcc_data *pcc_ss_data = NULL;
>> - int ret = 0;
>> + pcc_ss_id = per_cpu(cpu_pcc_subspace_idx, cpunum);
>> if (pcc_ss_id < 0)
>> - return -EIO;
>> + return -ENODEV;
>
> NIT: Could add here:
> pr_debug("Invalid pcc_ss_id\n");
> just as you did in cppc_set_reg()
Will add it in next version, Thanks.
>
>> pcc_ss_data = pcc_data[pcc_ss_id];
>> down_write(&pcc_ss_data->pcc_lock);
>> if (send_pcc_cmd(pcc_ss_id, CMD_READ) >= 0)
>> - cpc_read(cpunum, reg, perf);
>> + cpc_read(cpunum, reg, val);
>
> This was not introduced by your patch, but cpc_read() return a value.
> Shouldn't we return it instead of 0 ?
Indeed. Will optimize it, Thanks.
>
>> else
>> ret = -EIO;
>> @@ -1213,21 +1219,65 @@ static int cppc_get_perf(int cpunum, enum cppc_regs reg_idx, u64 *perf)
>> return ret;
>> }
>> - cpc_read(cpunum, reg, perf);
>> + cpc_read(cpunum, reg, val);
>
> Same comment as above
>
>> return 0;
>> }
>> +static int cppc_set_reg(int cpu, enum cppc_regs reg_idx, u64 val)
>
> Just to have similar functions, maybe 'cpu' should be renamed to 'cpunum' ?
> Or the other way around.
I prefer 'cpu', 'cpunum' looks like the number of cpus to me.
Will rename 'cpunum' to 'cpu' in cppc_get_reg(). Thanks.
>
>> +{
>> + struct cpc_desc *cpc_desc = per_cpu(cpc_desc_ptr, cpu);
>> + struct cppc_pcc_data *pcc_ss_data = NULL;
>> + struct cpc_register_resource *reg;
>> + int pcc_ss_id;
>> + int ret;
>> +
>> + if (!cpc_desc) {
>> + pr_debug("No CPC descriptor for CPU:%d\n", cpu);
>> + return -ENODEV;
>> + }
>> +
>> + reg = &cpc_desc->cpc_regs[reg_idx];
>> +
>> + if (!CPC_SUPPORTED(reg)) {
>> + pr_debug("CPC register (reg_idx=%u) is not supported\n", reg_idx);
>> + return -EOPNOTSUPP;
>> + }
>> +
>> + if (CPC_IN_PCC(reg)) {
>> + pcc_ss_id = per_cpu(cpu_pcc_subspace_idx, cpu);
>> +
>> + if (pcc_ss_id < 0) {
>> + pr_debug("Invalid pcc_ss_id\n");
>> + return -ENODEV;
>> + }
>> +
>> + ret = cpc_write(cpu, reg, val);
>> + if (ret)
>> + return ret;
>> +
>> + pcc_ss_data = pcc_data[pcc_ss_id];
>> +
>> + down_write(&pcc_ss_data->pcc_lock);
>> + /* after writing CPC, transfer the ownership of PCC to platform */
>> + ret = send_pcc_cmd(pcc_ss_id, CMD_WRITE);
>> + up_write(&pcc_ss_data->pcc_lock);
>> + return ret;
>> + }
>> +
>> + return cpc_write(cpu, reg, val);
>> +}
>> +
>
> [snip]
>
Powered by blists - more mailing lists