RE: [PATCH v2 2/3] ACPI:RAS2: Add ACPI RAS2 driver

[Shiju Jose]

>-----Original Message-----
>From: Jonathan Cameron <jonathan.cameron@xxxxxxxxxx>
>Sent: 06 March 2025 09:19
>To: Shiju Jose <shiju.jose@xxxxxxxxxx>
>Cc: linux-edac@xxxxxxxxxxxxxxx; linux-acpi@xxxxxxxxxxxxxxx; bp@xxxxxxxxx;
>tony.luck@xxxxxxxxx; rafael@xxxxxxxxxx; lenb@xxxxxxxxxx;
>mchehab@xxxxxxxxxx; leo.duran@xxxxxxx; Yazen.Ghannam@xxxxxxx; linux-
>cxl@xxxxxxxxxxxxxxx; dan.j.williams@xxxxxxxxx; dave@xxxxxxxxxxxx;
>dave.jiang@xxxxxxxxx; alison.schofield@xxxxxxxxx; vishal.l.verma@xxxxxxxxx;
>ira.weiny@xxxxxxxxx; david@xxxxxxxxxx; Vilas.Sridharan@xxxxxxx; linux-
>mm@xxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; rientjes@xxxxxxxxxx;
>jiaqiyan@xxxxxxxxxx; Jon.Grimm@xxxxxxx; dave.hansen@xxxxxxxxxxxxxxx;
>naoya.horiguchi@xxxxxxx; james.morse@xxxxxxx; jthoughton@xxxxxxxxxx;
>somasundaram.a@xxxxxxx; erdemaktas@xxxxxxxxxx; pgonda@xxxxxxxxxx;
>duenwen@xxxxxxxxxx; gthelen@xxxxxxxxxx;
>wschwartz@xxxxxxxxxxxxxxxxxxx; dferguson@xxxxxxxxxxxxxxxxxxx;
>wbs@xxxxxxxxxxxxxxxxxxxxxx; nifan.cxl@xxxxxxxxx; tanxiaofei
><tanxiaofei@xxxxxxxxxx>; Zengtao (B) <prime.zeng@xxxxxxxxxxxxx>; Roberto
>Sassu <roberto.sassu@xxxxxxxxxx>; kangkang.shen@xxxxxxxxxxxxx;
>wanghuiqiang <wanghuiqiang@xxxxxxxxxx>; Linuxarm
><linuxarm@xxxxxxxxxx>
>Subject: Re: [PATCH v2 2/3] ACPI:RAS2: Add ACPI RAS2 driver
>
>On Wed, 5 Mar 2025 18:02:23 +0000
><shiju.jose@xxxxxxxxxx> wrote:
>
>> From: Shiju Jose <shiju.jose@xxxxxxxxxx>
>>
>> Add support for ACPI RAS2 feature table (RAS2) defined in the ACPI 6.5
>> Specification, section 5.2.21.
>> Driver defines RAS2 Init, which extracts the RAS2 table and driver
>> adds auxiliary device for each memory feature which binds to the
>> RAS2 memory driver.
>>
>> Driver uses PCC mailbox to communicate with the ACPI HW and the driver
>> adds OSPM interfaces to send RAS2 commands.
>>
>> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
>> Co-developed-by: A Somasundaram <somasundaram.a@xxxxxxx>
>> Signed-off-by: A Somasundaram <somasundaram.a@xxxxxxx>
>> Co-developed-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
>> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
>> Tested-by: Daniel Ferguson <danielf@xxxxxxxxxxxxxxxxxxxxxx>
>> Signed-off-by: Shiju Jose <shiju.jose@xxxxxxxxxx>
>
>Hi Shiju,
>
>I took another look through as it's been a while and I've pretty much forgotten
>this code :(
>
>Anyhow, a few minor comments inline.

Hi Jonathan,

Thanks for the feedbacks.

Please find reply inline.

Thanks,
Shiju
>
>Thanks,
>
>Jonathan
>
>> diff --git a/drivers/acpi/ras2.c b/drivers/acpi/ras2.c new file mode
>> 100755 index 000000000000..8831a2bd5fab
>> --- /dev/null
>> +++ b/drivers/acpi/ras2.c
>
>
>
>
>> +static int ras2_register_pcc_channel(struct ras2_mem_ctx *ras2_ctx, int
>pcc_id)
>> +{
>> +	struct ras2_pcc_subspace *pcc_subspace;
>> +	struct pcc_mbox_chan *pcc_chan;
>> +	struct mbox_client *mbox_cl;
>> +
>> +	if (pcc_id < 0)
>> +		return -EINVAL;
>> +
>> +	mutex_lock(&ras2_pcc_lock);
>> +	list_for_each_entry(pcc_subspace, &ras2_pcc_subspaces, elem) {
>> +		if (pcc_subspace->pcc_id != pcc_id)
>> +			continue;
>> +		ras2_ctx->pcc_subspace = pcc_subspace;
>> +		pcc_subspace->ref_count++;
>> +		mutex_unlock(&ras2_pcc_lock);
>> +		return 0;
>> +	}
>> +	mutex_unlock(&ras2_pcc_lock);
>> +
>> +	pcc_subspace = kcalloc(1, sizeof(*pcc_subspace), GFP_KERNEL);
>
>if allocating a count of 1, why not kzalloc?

Will use kzalloc.
>
>> +	if (!pcc_subspace)
>> +		return -ENOMEM;
>
>
>
>
>> +static int acpi_ras2_parse(void)
>> +{
>> +	struct acpi_ras2_pcc_desc *pcc_desc_list;
>> +	int pcc_id;
>> +	u8 count = 0;
>> +	int rc, i;
>> +
>> +	if (ras2_tab->header.length  < sizeof(struct acpi_table_ras2)) {
>
>extra space before <

Will fix. 
>
>Maybe sizeof(*ras2_tab) is cleaner.
Sure.
>
>> +		pr_warn(FW_WARN "ACPI RAS2 table present but broken (too
>short #1)\n");
>> +		return -EINVAL;
>> +	}
>> +
>> +	if (!ras2_tab->num_pcc_descs) {
>> +		pr_warn(FW_WARN "No PCC descs in ACPI RAS2 table\n");
>> +		return -EINVAL;
>> +	}
>> +
>> +	pcc_desc_list = (struct acpi_ras2_pcc_desc *)(ras2_tab + 1);
>> +	/* Double scan for the case of only one actual controller */
>> +	pcc_id = -1;
>> +	count = 0;
>
>Already set above, so no need to do it again.  I'd do it just here.  Can
>put it in the loop init though.
Will change.
>
>> +	for (i = 0; i < ras2_tab->num_pcc_descs; i++, pcc_desc_list++) {
>> +		if (pcc_desc_list->feature_type != RAS2_FEAT_TYPE_MEMORY)
>> +			continue;
>> +		if (pcc_id == -1) {
>> +			pcc_id = pcc_desc_list->channel_id;
>> +			count++;
>> +		}
>> +		if (pcc_desc_list->channel_id != pcc_id)
>> +			count++;
>> +	}
>> +
>> +	/*
>> +	 * Workaround for the client platform with multiple scrub devices
>> +	 * but uses single PCC subspace for communication.
>> +	 */
>> +	if (count == 1) {
>> +		/* Add auxiliary device and bind ACPI RAS2 memory driver */
>> +		rc = ras2_add_aux_device(RAS2_MEM_DEV_ID_NAME, pcc_id);
>> +		if (rc)
>> +			return rc;
>> +
>> +		return 0;
>> +	}
>> +
>> +	pcc_desc_list = (struct acpi_ras2_pcc_desc *)(ras2_tab + 1);
>> +	count = 0;
>
>Maybe set in loop init.
Sure.
>
>> +	for (i = 0; i < ras2_tab->num_pcc_descs; i++, pcc_desc_list++) {
>> +		if (pcc_desc_list->feature_type != RAS2_FEAT_TYPE_MEMORY)
>> +			continue;
>> +		pcc_id = pcc_desc_list->channel_id;
>> +		/* Add auxiliary device and bind ACPI RAS2 memory driver */
>obvious enough to drop the comment I think.
Will do.
>
>> +		rc = ras2_add_aux_device(RAS2_MEM_DEV_ID_NAME, pcc_id);
>					 pcc_desc_list->channel_id);
>and no local variable.
Ok.
>
>> +		if (rc)
>> +			return rc;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +void __init acpi_ras2_init(void)
>> +{
>> +	acpi_status status;
>> +	int rc;
>> +
>> +	status = acpi_get_table(ACPI_SIG_RAS2, 0,
>> +				(struct acpi_table_header **)&ras2_tab);
>> +	if (ACPI_FAILURE(status) || !ras2_tab) {
>> +		const char *msg = acpi_format_exception(status);
>> +
>> +		pr_err("Failed to get table, %s\n", msg);
>
>If only going to use it here maybe
>		pr_err("Failed to get table, %s\n",
>		       acpi_format_exception(status));
>and save on the local variable.
Will change.
>
>> +		return;
>> +	}
>> +
>> +	rc = acpi_ras2_parse();
>> +	if (rc) {
>> +		acpi_put_table((struct acpi_table_header *)ras2_tab);
>> +		pr_err("Failed to parse RAS2 table\n");
>> +	}
>> +}
>> diff --git a/include/acpi/ras2.h b/include/acpi/ras2.h
>> new file mode 100644
>> index 000000000000..5b27c1f30096
>> --- /dev/null
>> +++ b/include/acpi/ras2.h
>> @@ -0,0 +1,48 @@
>> +/* SPDX-License-Identifier: GPL-2.0-only */
>> +/*
>> + * ACPI RAS2 driver header file
>> + *
>> + * Copyright (c) 2024-2025 HiSilicon Limited
>> + */
>> +
>> +#ifndef _ACPI_RAS2_H
>> +#define _ACPI_RAS2_H
>> +
>> +#include <linux/acpi.h>
>> +#include <linux/auxiliary_bus.h>
>> +#include <linux/mailbox_client.h>
>> +#include <linux/mutex.h>
>> +#include <linux/types.h>
>> +
>> +#define RAS2_PCC_CMD_COMPLETE	BIT(0)
>> +#define RAS2_PCC_CMD_ERROR	BIT(2)
>> +
>I think these bits are from table 14.11 and
>generic to all PCC status registers? Should these
>have more generic names rather than ras2 specific ones?
Yes.
Instead will use PCC_STATUS_CMD_ COMPLETE and  PCC_STATUS_ ERROR
from include/acpi/pcc.h. 

>
>> +/* RAS2 specific PCC commands */
>> +#define RAS2_PCC_CMD_EXEC 0x01
>Are we mixing commands and field definitions both
>with prefix RAS2_PCC_CMD_ ?  That is somewhat
>confusing.
Will add Table 5.82: .. here in the comment and 
Is rename to PCC_CMD_ EXEC_RAS2  better?
>
>> +
>> +#define RAS2_AUX_DEV_NAME "ras2"
>> +#define RAS2_MEM_DEV_ID_NAME "acpi_ras2_mem"
>> +
>I would add a forwards def
>struct device;
Sure.
>
>whilst it is really unlikely that headers would ever be reorganized
>such that auxiliary_bus.h would not include device.h given the embedded
>device we shouldn't rely on that here.
>
>> +/* Data structure RAS2 table */
>> +struct ras2_mem_ctx {
>> +	struct auxiliary_device adev;
>> +	/* Lock to provide mutually exclusive access to PCC channel */
>> +	struct mutex lock;
>> +	struct device *dev;
>> +	struct acpi_ras2_shmem __iomem *comm_addr;
>> +	void *pcc_subspace;
>> +	int id;
>> +};
>> +
>> +#ifdef CONFIG_ACPI_RAS2
>> +void __init acpi_ras2_init(void);
>> +int ras2_send_pcc_cmd(struct ras2_mem_ctx *ras2_ctx, u16 cmd);
>> +#else
>> +static inline void acpi_ras2_init(void) { }
>> +
>> +static inline int ras2_send_pcc_cmd(struct ras2_mem_ctx *ras2_ctx, u16
>cmd)
>
>Is this stub ever needed?  To me it seems unlikely
>we would have a user that is built without a dependency
>on CONFIG_ACPI_RAS2.  This is different from acpi_ras2_init()
>which makes much more sense to me.
Ok will remove.
>
>> +{
>> +	return -EOPNOTSUPP;
>> +}
>> +#endif
>> +#endif /* _ACPI_RAS2_H */

Thanks,
Shiju