Merge tag 'scmi-fixes-6.8' of git://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux into arm/fixes

Arm SCMI fixes for v6.8

Few fixes addressing the below issues:

1. A spurious IRQ related to the late reply can get wrongly associated
   with the new enqueued request resulting in misinterpretation of data
   in shared memory. This race-condition can be detected by looking at
   the channel status bits which the platform must set to the channel
   free before triggering the completion IRQ. Adding a consistency check
   to validate such condition will fix the issue.
2. Incorrect use of asm-generic/bug.h instead of generic linux/bUg.h
3. xa_store() can't check for possible duplication insertion, use
   xa_insert() instead
4. Fix the SCMI clock protocol version in the v3.2 SCMI specification
5. Incorrect upgrade of highest supported clock protocol version from
   v2.0 to v3.0

* tag 'scmi-fixes-6.8' of git://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux:
  firmware: arm_scmi: Fix the clock protocol supported version
  firmware: arm_scmi: Fix the clock protocol version for v3.2
  firmware: arm_scmi: Use xa_insert() when saving raw queues
  firmware: arm_scmi: Use xa_insert() to store opps
  firmware: arm_scmi: Replace asm-generic/bug.h with linux/bug.h
  firmware: arm_scmi: Check mailbox/SMT channel for consistency

Link: https://lore.kernel.org/r/20240122161640.3551085-1-sudeep.holla@arm.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

6 files changed, 50 insertions, 13 deletions

diff --git a/drivers/firmware/arm_scmi/clock.c b/drivers/firmware/arm_scmi/clock.c
index c0644558042a06..e2050adbf85c6a 100644
--- a/drivers/firmware/arm_scmi/clock.c
+++ b/drivers/firmware/arm_scmi/clock.c
@@ -13,7 +13,7 @@
 #include "notify.h"
 
 /* Updated only after ALL the mandatory features for that version are merged */
-#define SCMI_PROTOCOL_SUPPORTED_VERSION		0x20001
+#define SCMI_PROTOCOL_SUPPORTED_VERSION		0x20000
 
 enum scmi_clock_protocol_cmd {
 	CLOCK_ATTRIBUTES = 0x3,
@@ -954,8 +954,7 @@ static int scmi_clock_protocol_init(const struct scmi_protocol_handle *ph)
 			scmi_clock_describe_rates_get(ph, clkid, clk);
 	}
 
-	if (PROTOCOL_REV_MAJOR(version) >= 0x2 &&
-	    PROTOCOL_REV_MINOR(version) >= 0x1) {
+	if (PROTOCOL_REV_MAJOR(version) >= 0x3) {
 		cinfo->clock_config_set = scmi_clock_config_set_v2;
 		cinfo->clock_config_get = scmi_clock_config_get_v2;
 	} else {
diff --git a/drivers/firmware/arm_scmi/common.h b/drivers/firmware/arm_scmi/common.h
index c46dc5215af7a7..00b165d1f502df 100644
--- a/drivers/firmware/arm_scmi/common.h
+++ b/drivers/firmware/arm_scmi/common.h
@@ -314,6 +314,7 @@ void shmem_fetch_notification(struct scmi_shared_mem __iomem *shmem,
 void shmem_clear_channel(struct scmi_shared_mem __iomem *shmem);
 bool shmem_poll_done(struct scmi_shared_mem __iomem *shmem,
 		     struct scmi_xfer *xfer);
+bool shmem_channel_free(struct scmi_shared_mem __iomem *shmem);
 
 /* declarations for message passing transports */
 struct scmi_msg_payld;
diff --git a/drivers/firmware/arm_scmi/mailbox.c b/drivers/firmware/arm_scmi/mailbox.c
index 19246ed1f01ff7..b8d470417e8f99 100644
--- a/drivers/firmware/arm_scmi/mailbox.c
+++ b/drivers/firmware/arm_scmi/mailbox.c
@@ -45,6 +45,20 @@ static void rx_callback(struct mbox_client *cl, void *m)
 {
 	struct scmi_mailbox *smbox = client_to_scmi_mailbox(cl);
 
+	/*
+	 * An A2P IRQ is NOT valid when received while the platform still has
+	 * the ownership of the channel, because the platform at first releases
+	 * the SMT channel and then sends the completion interrupt.
+	 *
+	 * This addresses a possible race condition in which a spurious IRQ from
+	 * a previous timed-out reply which arrived late could be wrongly
+	 * associated with the next pending transaction.
+	 */
+	if (cl->knows_txdone && !shmem_channel_free(smbox->shmem)) {
+		dev_warn(smbox->cinfo->dev, "Ignoring spurious A2P IRQ !\n");
+		return;
+	}
+
 	scmi_rx_callback(smbox->cinfo, shmem_read_header(smbox->shmem), NULL);
 }
 
diff --git a/drivers/firmware/arm_scmi/perf.c b/drivers/firmware/arm_scmi/perf.c
index 8ea2a7b3d35d20..211e8e0aef2c2b 100644
--- a/drivers/firmware/arm_scmi/perf.c
+++ b/drivers/firmware/arm_scmi/perf.c
@@ -350,8 +350,8 @@ process_response_opp(struct scmi_opp *opp, unsigned int loop_idx,
 }
 
 static inline void
-process_response_opp_v4(struct perf_dom_info *dom, struct scmi_opp *opp,
-			unsigned int loop_idx,
+process_response_opp_v4(struct device *dev, struct perf_dom_info *dom,
+			struct scmi_opp *opp, unsigned int loop_idx,
 			const struct scmi_msg_resp_perf_describe_levels_v4 *r)
 {
 	opp->perf = le32_to_cpu(r->opp[loop_idx].perf_val);
@@ -362,10 +362,23 @@ process_response_opp_v4(struct perf_dom_info *dom, struct scmi_opp *opp,
 	/* Note that PERF v4 reports always five 32-bit words */
 	opp->indicative_freq = le32_to_cpu(r->opp[loop_idx].indicative_freq);
 	if (dom->level_indexing_mode) {
+		int ret;
+
 		opp->level_index = le32_to_cpu(r->opp[loop_idx].level_index);
 
-		xa_store(&dom->opps_by_idx, opp->level_index, opp, GFP_KERNEL);
-		xa_store(&dom->opps_by_lvl, opp->perf, opp, GFP_KERNEL);
+		ret = xa_insert(&dom->opps_by_idx, opp->level_index, opp,
+				GFP_KERNEL);
+		if (ret)
+			dev_warn(dev,
+				 "Failed to add opps_by_idx at %d - ret:%d\n",
+				 opp->level_index, ret);
+
+		ret = xa_insert(&dom->opps_by_lvl, opp->perf, opp, GFP_KERNEL);
+		if (ret)
+			dev_warn(dev,
+				 "Failed to add opps_by_lvl at %d - ret:%d\n",
+				 opp->perf, ret);
+
 		hash_add(dom->opps_by_freq, &opp->hash, opp->indicative_freq);
 	}
 }
@@ -382,7 +395,7 @@ iter_perf_levels_process_response(const struct scmi_protocol_handle *ph,
 	if (PROTOCOL_REV_MAJOR(p->version) <= 0x3)
 		process_response_opp(opp, st->loop_idx, response);
 	else
-		process_response_opp_v4(p->perf_dom, opp, st->loop_idx,
+		process_response_opp_v4(ph->dev, p->perf_dom, opp, st->loop_idx,
 					response);
 	p->perf_dom->opp_count++;
 
diff --git a/drivers/firmware/arm_scmi/raw_mode.c b/drivers/firmware/arm_scmi/raw_mode.c
index 0493aa3c12bf53..35057351850335 100644
--- a/drivers/firmware/arm_scmi/raw_mode.c
+++ b/drivers/firmware/arm_scmi/raw_mode.c
@@ -1111,7 +1111,6 @@ static int scmi_raw_mode_setup(struct scmi_raw_mode_info *raw,
 		int i;
 
 		for (i = 0; i < num_chans; i++) {
-			void *xret;
 			struct scmi_raw_queue *q;
 
 			q = scmi_raw_queue_init(raw);
@@ -1120,13 +1119,12 @@ static int scmi_raw_mode_setup(struct scmi_raw_mode_info *raw,
 				goto err_xa;
 			}
 
-			xret = xa_store(&raw->chans_q, channels[i], q,
+			ret = xa_insert(&raw->chans_q, channels[i], q,
 					GFP_KERNEL);
-			if (xa_err(xret)) {
+			if (ret) {
 				dev_err(dev,
 					"Fail to allocate Raw queue 0x%02X\n",
 					channels[i]);
-				ret = xa_err(xret);
 				goto err_xa;
 			}
 		}
@@ -1322,6 +1320,12 @@ void scmi_raw_message_report(void *r, struct scmi_xfer *xfer,
 	dev = raw->handle->dev;
 	q = scmi_raw_queue_select(raw, idx,
 				  SCMI_XFER_IS_CHAN_SET(xfer) ? chan_id : 0);
+	if (!q) {
+		dev_warn(dev,
+			 "RAW[%d] - NO queue for chan 0x%X. Dropping report.\n",
+			 idx, chan_id);
+		return;
+	}
 
 	/*
 	 * Grab the msg_q_lock upfront to avoid a possible race between
diff --git a/drivers/firmware/arm_scmi/shmem.c b/drivers/firmware/arm_scmi/shmem.c
index 87b4f4d35f0623..8bf495bcad09b7 100644
--- a/drivers/firmware/arm_scmi/shmem.c
+++ b/drivers/firmware/arm_scmi/shmem.c
@@ -10,7 +10,7 @@
 #include <linux/processor.h>
 #include <linux/types.h>
 
-#include <asm-generic/bug.h>
+#include <linux/bug.h>
 
 #include "common.h"
 
@@ -122,3 +122,9 @@ bool shmem_poll_done(struct scmi_shared_mem __iomem *shmem,
 		(SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR |
 		 SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE);
 }
+
+bool shmem_channel_free(struct scmi_shared_mem __iomem *shmem)
+{
+	return (ioread32(&shmem->channel_status) &
+			SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE);
+}