[PATCH AUTOSEL 6.11 01/16] s390/pci: Sort PCI functions prior to creating virtual busses

[PATCH AUTOSEL 6.11 01/16] s390/pci: Sort PCI functions prior to creating virtual busses

[Sasha Levin]

From: Niklas Schnelle <schnelle@linux.ibm.com>

[ Upstream commit 0467cdde8c4320bbfdb31a8cff1277b202f677fc ]

Instead of relying on the observed but not architected firmware behavior
that PCI functions from the same card are listed in ascending RID order
in clp_list_pci() ensure this by sorting. To allow for sorting separate
the initial clp_list_pci() and creation of the virtual PCI busses.

Note that fundamentally in our per-PCI function hotplug design non RID
order of discovery is still possible. For example when the two PFs of
a two port NIC are hotplugged after initial boot and in descending RID
order. In this case the virtual PCI bus would be created by the second
PF using that PF's UID as domain number instead of that of the first PF.
Thus the domain number would then change from the UID of the second PF
to that of the first PF on reboot but there is really nothing we can do
about that since changing domain numbers at runtime seems even worse.
This only impacts the domain number as the RIDs are consistent and thus
even with just the second PF visible it will show up in the correct
position on the virtual bus.

Reviewed-by: Gerd Bayer <gbayer@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/s390/include/asm/pci.h |  5 ++-
 arch/s390/pci/pci.c         | 69 ++++++++++++++++++++++++++++++++-----
 arch/s390/pci/pci_clp.c     | 12 ++++---
 arch/s390/pci/pci_event.c   | 13 ++++---
 4 files changed, 82 insertions(+), 17 deletions(-)

diff --git a/arch/s390/include/asm/pci.h b/arch/s390/include/asm/pci.h
index 30820a649e6e7..fdec455892486 100644
--- a/arch/s390/include/asm/pci.h
+++ b/arch/s390/include/asm/pci.h
@@ -130,6 +130,7 @@ struct zpci_dev {
 	u16		vfn;		/* virtual function number */
 	u16		pchid;		/* physical channel ID */
 	u16		maxstbl;	/* Maximum store block size */
+	u16		rid;		/* RID as supplied by firmware */
 	u8		pfgid;		/* function group ID */
 	u8		pft;		/* pci function type */
 	u8		port;
@@ -203,12 +204,14 @@ extern struct airq_iv *zpci_aif_sbv;
 ----------------------------------------------------------------------------- */
 /* Base stuff */
 struct zpci_dev *zpci_create_device(u32 fid, u32 fh, enum zpci_state state);
+int zpci_add_device(struct zpci_dev *zdev);
 int zpci_enable_device(struct zpci_dev *);
 int zpci_disable_device(struct zpci_dev *);
 int zpci_scan_configured_device(struct zpci_dev *zdev, u32 fh);
 int zpci_deconfigure_device(struct zpci_dev *zdev);
 void zpci_device_reserved(struct zpci_dev *zdev);
 bool zpci_is_device_configured(struct zpci_dev *zdev);
+int zpci_scan_devices(void);
 
 int zpci_hot_reset_device(struct zpci_dev *zdev);
 int zpci_register_ioat(struct zpci_dev *, u8, u64, u64, u64, u8 *);
@@ -218,7 +221,7 @@ void zpci_update_fh(struct zpci_dev *zdev, u32 fh);
 
 /* CLP */
 int clp_setup_writeback_mio(void);
-int clp_scan_pci_devices(void);
+int clp_scan_pci_devices(struct list_head *scan_list);
 int clp_query_pci_fn(struct zpci_dev *zdev);
 int clp_enable_fh(struct zpci_dev *zdev, u32 *fh, u8 nr_dma_as);
 int clp_disable_fh(struct zpci_dev *zdev, u32 *fh);
diff --git a/arch/s390/pci/pci.c b/arch/s390/pci/pci.c
index cff4838fad216..e70318fba275a 100644
--- a/arch/s390/pci/pci.c
+++ b/arch/s390/pci/pci.c
@@ -29,6 +29,7 @@
 #include <linux/pci.h>
 #include <linux/printk.h>
 #include <linux/lockdep.h>
+#include <linux/list_sort.h>
 
 #include <asm/isc.h>
 #include <asm/airq.h>
@@ -786,7 +787,6 @@ struct zpci_dev *zpci_create_device(u32 fid, u32 fh, enum zpci_state state)
 	struct zpci_dev *zdev;
 	int rc;
 
-	zpci_dbg(1, "add fid:%x, fh:%x, c:%d\n", fid, fh, state);
 	zdev = kzalloc(sizeof(*zdev), GFP_KERNEL);
 	if (!zdev)
 		return ERR_PTR(-ENOMEM);
@@ -806,6 +806,19 @@ struct zpci_dev *zpci_create_device(u32 fid, u32 fh, enum zpci_state state)
 	mutex_init(&zdev->fmb_lock);
 	mutex_init(&zdev->kzdev_lock);
 
+	return zdev;
+
+error:
+	zpci_dbg(0, "crt fid:%x, rc:%d\n", fid, rc);
+	kfree(zdev);
+	return ERR_PTR(rc);
+}
+
+int zpci_add_device(struct zpci_dev *zdev)
+{
+	int rc;
+
+	zpci_dbg(1, "add fid:%x, fh:%x, c:%d\n", zdev->fid, zdev->fh, zdev->state);
 	rc = zpci_init_iommu(zdev);
 	if (rc)
 		goto error;
@@ -817,15 +830,13 @@ struct zpci_dev *zpci_create_device(u32 fid, u32 fh, enum zpci_state state)
 	spin_lock(&zpci_list_lock);
 	list_add_tail(&zdev->entry, &zpci_list);
 	spin_unlock(&zpci_list_lock);
-
-	return zdev;
+	return 0;
 
 error_destroy_iommu:
 	zpci_destroy_iommu(zdev);
 error:
-	zpci_dbg(0, "add fid:%x, rc:%d\n", fid, rc);
-	kfree(zdev);
-	return ERR_PTR(rc);
+	zpci_dbg(0, "add fid:%x, rc:%d\n", zdev->fid, rc);
+	return rc;
 }
 
 bool zpci_is_device_configured(struct zpci_dev *zdev)
@@ -1083,6 +1094,49 @@ bool zpci_is_enabled(void)
 	return s390_pci_initialized;
 }
 
+static int zpci_cmp_rid(void *priv, const struct list_head *a,
+			const struct list_head *b)
+{
+	struct zpci_dev *za = container_of(a, struct zpci_dev, entry);
+	struct zpci_dev *zb = container_of(b, struct zpci_dev, entry);
+
+	/*
+	 * PCI functions without RID available maintain original order
+	 * between themselves but sort before those with RID.
+	 */
+	if (za->rid == zb->rid)
+		return za->rid_available > zb->rid_available;
+	/*
+	 * PCI functions with RID sort by RID ascending.
+	 */
+	return za->rid > zb->rid;
+}
+
+static void zpci_add_devices(struct list_head *scan_list)
+{
+	struct zpci_dev *zdev, *tmp;
+
+	list_sort(NULL, scan_list, &zpci_cmp_rid);
+	list_for_each_entry_safe(zdev, tmp, scan_list, entry) {
+		list_del_init(&zdev->entry);
+		zpci_add_device(zdev);
+	}
+}
+
+int zpci_scan_devices(void)
+{
+	LIST_HEAD(scan_list);
+	int rc;
+
+	rc = clp_scan_pci_devices(&scan_list);
+	if (rc)
+		return rc;
+
+	zpci_add_devices(&scan_list);
+	zpci_bus_scan_busses();
+	return 0;
+}
+
 static int __init pci_base_init(void)
 {
 	int rc;
@@ -1112,10 +1166,9 @@ static int __init pci_base_init(void)
 	if (rc)
 		goto out_irq;
 
-	rc = clp_scan_pci_devices();
+	rc = zpci_scan_devices();
 	if (rc)
 		goto out_find;
-	zpci_bus_scan_busses();
 
 	s390_pci_initialized = 1;
 	return 0;
diff --git a/arch/s390/pci/pci_clp.c b/arch/s390/pci/pci_clp.c
index ee90a91ed8881..3049e4eb01fe7 100644
--- a/arch/s390/pci/pci_clp.c
+++ b/arch/s390/pci/pci_clp.c
@@ -164,8 +164,10 @@ static int clp_store_query_pci_fn(struct zpci_dev *zdev,
 	zdev->port = response->port;
 	zdev->uid = response->uid;
 	zdev->fmb_length = sizeof(u32) * response->fmb_len;
-	zdev->rid_available = response->rid_avail;
 	zdev->is_physfn = response->is_physfn;
+	zdev->rid_available = response->rid_avail;
+	if (zdev->rid_available)
+		zdev->rid = response->rid;
 	if (!s390_pci_no_rid && zdev->rid_available)
 		zdev->devfn = response->rid & ZPCI_RID_MASK_DEVFN;
 
@@ -407,6 +409,7 @@ static int clp_find_pci(struct clp_req_rsp_list_pci *rrb, u32 fid,
 
 static void __clp_add(struct clp_fh_list_entry *entry, void *data)
 {
+	struct list_head *scan_list = data;
 	struct zpci_dev *zdev;
 
 	if (!entry->vendor_id)
@@ -417,10 +420,11 @@ static void __clp_add(struct clp_fh_list_entry *entry, void *data)
 		zpci_zdev_put(zdev);
 		return;
 	}
-	zpci_create_device(entry->fid, entry->fh, entry->config_state);
+	zdev = zpci_create_device(entry->fid, entry->fh, entry->config_state);
+	list_add_tail(&zdev->entry, scan_list);
 }
 
-int clp_scan_pci_devices(void)
+int clp_scan_pci_devices(struct list_head *scan_list)
 {
 	struct clp_req_rsp_list_pci *rrb;
 	int rc;
@@ -429,7 +433,7 @@ int clp_scan_pci_devices(void)
 	if (!rrb)
 		return -ENOMEM;
 
-	rc = clp_list_pci(rrb, NULL, __clp_add);
+	rc = clp_list_pci(rrb, scan_list, __clp_add);
 
 	clp_free_block(rrb);
 	return rc;
diff --git a/arch/s390/pci/pci_event.c b/arch/s390/pci/pci_event.c
index d4f19d33914cb..47f934f4e828e 100644
--- a/arch/s390/pci/pci_event.c
+++ b/arch/s390/pci/pci_event.c
@@ -340,6 +340,7 @@ static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
 			zdev = zpci_create_device(ccdf->fid, ccdf->fh, ZPCI_FN_STATE_CONFIGURED);
 			if (IS_ERR(zdev))
 				break;
+			zpci_add_device(zdev);
 		} else {
 			/* the configuration request may be stale */
 			if (zdev->state != ZPCI_FN_STATE_STANDBY)
@@ -349,10 +350,14 @@ static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
 		zpci_scan_configured_device(zdev, ccdf->fh);
 		break;
 	case 0x0302: /* Reserved -> Standby */
-		if (!zdev)
-			zpci_create_device(ccdf->fid, ccdf->fh, ZPCI_FN_STATE_STANDBY);
-		else
+		if (!zdev) {
+			zdev = zpci_create_device(ccdf->fid, ccdf->fh, ZPCI_FN_STATE_STANDBY);
+			if (IS_ERR(zdev))
+				break;
+			zpci_add_device(zdev);
+		} else {
 			zpci_update_fh(zdev, ccdf->fh);
+		}
 		break;
 	case 0x0303: /* Deconfiguration requested */
 		if (zdev) {
@@ -381,7 +386,7 @@ static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
 		break;
 	case 0x0306: /* 0x308 or 0x302 for multiple devices */
 		zpci_remove_reserved_devices();
-		clp_scan_pci_devices();
+		zpci_scan_devices();
 		break;
 	case 0x0308: /* Standby -> Reserved */
 		if (!zdev)
-- 
2.43.0

[PATCH AUTOSEL 6.11 02/16] s390/pci: Use topology ID for multi-function devices

[Sasha Levin]

From: Niklas Schnelle <schnelle@linux.ibm.com>

[ Upstream commit 126034faaac5f356822c4a9bebfa75664da11056 ]

The newly introduced topology ID (TID) field in the CLP Query PCI
Function explicitly identifies groups of PCI functions whose RIDs belong
to the same (sub-)topology. When available use the TID instead of the
PCHID to match zPCI busses/domains for multi-function devices. Note that
currently only a single PCI bus per TID is supported. This change is
required because in future machines the PCHID will not identify a PCI
card but a specific port in the case of some multi-port NICs while from
a PCI point of view the entire card is a subtopology.

Reviewed-by: Gerd Bayer <gbayer@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/s390/include/asm/pci.h     |  9 ++++++---
 arch/s390/include/asm/pci_clp.h |  8 +++++---
 arch/s390/pci/pci_bus.c         | 17 ++++++++++-------
 arch/s390/pci/pci_clp.c         |  3 +++
 4 files changed, 24 insertions(+), 13 deletions(-)

diff --git a/arch/s390/include/asm/pci.h b/arch/s390/include/asm/pci.h
index fdec455892486..94b5ce797c63b 100644
--- a/arch/s390/include/asm/pci.h
+++ b/arch/s390/include/asm/pci.h
@@ -107,9 +107,10 @@ struct zpci_bus {
 	struct list_head	resources;
 	struct list_head	bus_next;
 	struct resource		bus_resource;
-	int			pchid;
+	int			topo;		/* TID if topo_is_tid, PCHID otherwise */
 	int			domain_nr;
-	bool			multifunction;
+	u8			multifunction	: 1;
+	u8			topo_is_tid	: 1;
 	enum pci_bus_speed	max_bus_speed;
 };
 
@@ -131,6 +132,7 @@ struct zpci_dev {
 	u16		pchid;		/* physical channel ID */
 	u16		maxstbl;	/* Maximum store block size */
 	u16		rid;		/* RID as supplied by firmware */
+	u16		tid;		/* Topology for which RID is valid */
 	u8		pfgid;		/* function group ID */
 	u8		pft;		/* pci function type */
 	u8		port;
@@ -141,7 +143,8 @@ struct zpci_dev {
 	u8		is_physfn	: 1;
 	u8		util_str_avail	: 1;
 	u8		irqs_registered	: 1;
-	u8		reserved	: 2;
+	u8		tid_avail	: 1;
+	u8		reserved	: 1;
 	unsigned int	devfn;		/* DEVFN part of the RID*/
 
 	u8 pfip[CLP_PFIP_NR_SEGMENTS];	/* pci function internal path */
diff --git a/arch/s390/include/asm/pci_clp.h b/arch/s390/include/asm/pci_clp.h
index f0c677ddd2706..14afb9ce91f3b 100644
--- a/arch/s390/include/asm/pci_clp.h
+++ b/arch/s390/include/asm/pci_clp.h
@@ -110,7 +110,8 @@ struct clp_req_query_pci {
 struct clp_rsp_query_pci {
 	struct clp_rsp_hdr hdr;
 	u16 vfn;			/* virtual fn number */
-	u16			:  3;
+	u16			:  2;
+	u16 tid_avail		:  1;
 	u16 rid_avail		:  1;
 	u16 is_physfn		:  1;
 	u16 reserved1		:  1;
@@ -130,8 +131,9 @@ struct clp_rsp_query_pci {
 	u64 edma;			/* end dma as */
 #define ZPCI_RID_MASK_DEVFN 0x00ff
 	u16 rid;			/* BUS/DEVFN PCI address */
-	u16 reserved0;
-	u32 reserved[10];
+	u32 reserved0;
+	u16 tid;
+	u32 reserved[9];
 	u32 uid;			/* user defined id */
 	u8 util_str[CLP_UTIL_STR_LEN];	/* utility string */
 	u32 reserved2[16];
diff --git a/arch/s390/pci/pci_bus.c b/arch/s390/pci/pci_bus.c
index daa5d7450c7d3..54879e773e4a3 100644
--- a/arch/s390/pci/pci_bus.c
+++ b/arch/s390/pci/pci_bus.c
@@ -232,13 +232,13 @@ static void zpci_bus_put(struct zpci_bus *zbus)
 	kref_put(&zbus->kref, zpci_bus_release);
 }
 
-static struct zpci_bus *zpci_bus_get(int pchid)
+static struct zpci_bus *zpci_bus_get(int topo, bool topo_is_tid)
 {
 	struct zpci_bus *zbus;
 
 	mutex_lock(&zbus_list_lock);
 	list_for_each_entry(zbus, &zbus_list, bus_next) {
-		if (pchid == zbus->pchid) {
+		if (topo_is_tid == zbus->topo_is_tid && topo == zbus->topo) {
 			kref_get(&zbus->kref);
 			goto out_unlock;
 		}
@@ -249,7 +249,7 @@ static struct zpci_bus *zpci_bus_get(int pchid)
 	return zbus;
 }
 
-static struct zpci_bus *zpci_bus_alloc(int pchid)
+static struct zpci_bus *zpci_bus_alloc(int topo, bool topo_is_tid)
 {
 	struct zpci_bus *zbus;
 
@@ -257,7 +257,8 @@ static struct zpci_bus *zpci_bus_alloc(int pchid)
 	if (!zbus)
 		return NULL;
 
-	zbus->pchid = pchid;
+	zbus->topo = topo;
+	zbus->topo_is_tid = topo_is_tid;
 	INIT_LIST_HEAD(&zbus->bus_next);
 	mutex_lock(&zbus_list_lock);
 	list_add_tail(&zbus->bus_next, &zbus_list);
@@ -321,8 +322,9 @@ static int zpci_bus_add_device(struct zpci_bus *zbus, struct zpci_dev *zdev)
 
 int zpci_bus_device_register(struct zpci_dev *zdev, struct pci_ops *ops)
 {
+	bool topo_is_tid = zdev->tid_avail;
 	struct zpci_bus *zbus = NULL;
-	int rc = -EBADF;
+	int topo, rc = -EBADF;
 
 	if (zpci_nb_devices == ZPCI_NR_DEVICES) {
 		pr_warn("Adding PCI function %08x failed because the configured limit of %d is reached\n",
@@ -333,11 +335,12 @@ int zpci_bus_device_register(struct zpci_dev *zdev, struct pci_ops *ops)
 	if (zdev->devfn >= ZPCI_FUNCTIONS_PER_BUS)
 		return -EINVAL;
 
+	topo = topo_is_tid ? zdev->tid : zdev->pchid;
 	if (!s390_pci_no_rid && zdev->rid_available)
-		zbus = zpci_bus_get(zdev->pchid);
+		zbus = zpci_bus_get(topo, topo_is_tid);
 
 	if (!zbus) {
-		zbus = zpci_bus_alloc(zdev->pchid);
+		zbus = zpci_bus_alloc(topo, topo_is_tid);
 		if (!zbus)
 			return -ENOMEM;
 	}
diff --git a/arch/s390/pci/pci_clp.c b/arch/s390/pci/pci_clp.c
index 3049e4eb01fe7..658802b82f9eb 100644
--- a/arch/s390/pci/pci_clp.c
+++ b/arch/s390/pci/pci_clp.c
@@ -170,6 +170,9 @@ static int clp_store_query_pci_fn(struct zpci_dev *zdev,
 		zdev->rid = response->rid;
 	if (!s390_pci_no_rid && zdev->rid_available)
 		zdev->devfn = response->rid & ZPCI_RID_MASK_DEVFN;
+	zdev->tid_avail = response->tid_avail;
+	if (zdev->tid_avail)
+		zdev->tid = response->tid;
 
 	memcpy(zdev->pfip, response->pfip, sizeof(zdev->pfip));
 	if (response->util_str_avail) {
-- 
2.43.0

[PATCH AUTOSEL 6.11 03/16] s390/pci: Ignore RID for isolated VFs

[Sasha Levin]

From: Niklas Schnelle <schnelle@linux.ibm.com>

[ Upstream commit 25f39d3dcb48bbc824a77d16b3d977f0f3713cfe ]

Ensure that VFs used in isolation, that is with their parent PF not
visible to the configuration but with their RID exposed, are treated
compatibly with existing isolated VF use cases without exposed RID
including RoCE Express VFs. This allows creating configurations where
one LPAR manages PFs while their child VFs are used by other LPARs. This
gives the LPAR managing the PFs a role analogous to that of the
hypervisor in a typical use case of passing child VFs to guests.

Instead of creating a multifunction struct zpci_bus whenever a PCI
function with RID exposed is discovered only create such a bus for
configured physical functions and only consider multifunction busses
when searching for an existing bus. Additionally only set zdev->devfn to
the devfn part of the RID once the function is added to a multifunction
bus.

This also fixes probing of more than 7 such isolated VFs from the same
physical bus. This is because common PCI code in pci_scan_slot() only
looks for more functions when pdev->multifunction is set which somewhat
counter intutively is not the case for VFs.

Note that PFs are looked at before their child VFs is guaranteed because
we sort the zpci_list by RID ascending.

Reviewed-by: Gerd Bayer <gbayer@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/s390/pci/pci_bus.c | 33 ++++++++++++++++++++-------------
 arch/s390/pci/pci_clp.c |  2 --
 2 files changed, 20 insertions(+), 15 deletions(-)

diff --git a/arch/s390/pci/pci_bus.c b/arch/s390/pci/pci_bus.c
index 54879e773e4a3..1b74a000ff645 100644
--- a/arch/s390/pci/pci_bus.c
+++ b/arch/s390/pci/pci_bus.c
@@ -168,9 +168,16 @@ void zpci_bus_scan_busses(void)
 	mutex_unlock(&zbus_list_lock);
 }
 
+static bool zpci_bus_is_multifunction_root(struct zpci_dev *zdev)
+{
+	return !s390_pci_no_rid && zdev->rid_available &&
+		zpci_is_device_configured(zdev) &&
+		!zdev->vfn;
+}
+
 /* zpci_bus_create_pci_bus - Create the PCI bus associated with this zbus
  * @zbus: the zbus holding the zdevices
- * @fr: PCI root function that will determine the bus's domain, and bus speeed
+ * @fr: PCI root function that will determine the bus's domain, and bus speed
  * @ops: the pci operations
  *
  * The PCI function @fr determines the domain (its UID), multifunction property
@@ -188,7 +195,7 @@ static int zpci_bus_create_pci_bus(struct zpci_bus *zbus, struct zpci_dev *fr, s
 		return domain;
 
 	zbus->domain_nr = domain;
-	zbus->multifunction = fr->rid_available;
+	zbus->multifunction = zpci_bus_is_multifunction_root(fr);
 	zbus->max_bus_speed = fr->max_bus_speed;
 
 	/*
@@ -238,6 +245,8 @@ static struct zpci_bus *zpci_bus_get(int topo, bool topo_is_tid)
 
 	mutex_lock(&zbus_list_lock);
 	list_for_each_entry(zbus, &zbus_list, bus_next) {
+		if (!zbus->multifunction)
+			continue;
 		if (topo_is_tid == zbus->topo_is_tid && topo == zbus->topo) {
 			kref_get(&zbus->kref);
 			goto out_unlock;
@@ -293,19 +302,22 @@ static int zpci_bus_add_device(struct zpci_bus *zbus, struct zpci_dev *zdev)
 {
 	int rc = -EINVAL;
 
+	if (zbus->multifunction) {
+		if (!zdev->rid_available) {
+			WARN_ONCE(1, "rid_available not set for multifunction\n");
+			return rc;
+		}
+		zdev->devfn = zdev->rid & ZPCI_RID_MASK_DEVFN;
+	}
+
 	if (zbus->function[zdev->devfn]) {
 		pr_err("devfn %04x is already assigned\n", zdev->devfn);
 		return rc;
 	}
-
 	zdev->zbus = zbus;
 	zbus->function[zdev->devfn] = zdev;
 	zpci_nb_devices++;
 
-	if (zbus->multifunction && !zdev->rid_available) {
-		WARN_ONCE(1, "rid_available not set for multifunction\n");
-		goto error;
-	}
 	rc = zpci_init_slot(zdev);
 	if (rc)
 		goto error;
@@ -332,13 +344,8 @@ int zpci_bus_device_register(struct zpci_dev *zdev, struct pci_ops *ops)
 		return -ENOSPC;
 	}
 
-	if (zdev->devfn >= ZPCI_FUNCTIONS_PER_BUS)
-		return -EINVAL;
-
 	topo = topo_is_tid ? zdev->tid : zdev->pchid;
-	if (!s390_pci_no_rid && zdev->rid_available)
-		zbus = zpci_bus_get(topo, topo_is_tid);
-
+	zbus = zpci_bus_get(topo, topo_is_tid);
 	if (!zbus) {
 		zbus = zpci_bus_alloc(topo, topo_is_tid);
 		if (!zbus)
diff --git a/arch/s390/pci/pci_clp.c b/arch/s390/pci/pci_clp.c
index 658802b82f9eb..421c95e1841ef 100644
--- a/arch/s390/pci/pci_clp.c
+++ b/arch/s390/pci/pci_clp.c
@@ -168,8 +168,6 @@ static int clp_store_query_pci_fn(struct zpci_dev *zdev,
 	zdev->rid_available = response->rid_avail;
 	if (zdev->rid_available)
 		zdev->rid = response->rid;
-	if (!s390_pci_no_rid && zdev->rid_available)
-		zdev->devfn = response->rid & ZPCI_RID_MASK_DEVFN;
 	zdev->tid_avail = response->tid_avail;
 	if (zdev->tid_avail)
 		zdev->tid = response->tid;
-- 
2.43.0

[PATCH AUTOSEL 6.11 06/16] s390/cpum_sf: Handle CPU hotplug remove during sampling

[Sasha Levin]

From: Thomas Richter <tmricht@linux.ibm.com>

[ Upstream commit a0bd7dacbd51c632b8e2c0500b479af564afadf3 ]

CPU hotplug remove handling triggers the following function
call sequence:

   CPUHP_AP_PERF_S390_SF_ONLINE  --> s390_pmu_sf_offline_cpu()
   ...
   CPUHP_AP_PERF_ONLINE          --> perf_event_exit_cpu()

The s390 CPUMF sampling CPU hotplug handler invokes:

 s390_pmu_sf_offline_cpu()
 +-->  cpusf_pmu_setup()
       +--> setup_pmc_cpu()
            +--> deallocate_buffers()

This function de-allocates all sampling data buffers (SDBs) allocated
for that CPU at event initialization. It also clears the
PMU_F_RESERVED bit. The CPU is gone and can not be sampled.

With the event still being active on the removed CPU, the CPU event
hotplug support in kernel performance subsystem triggers the
following function calls on the removed CPU:

  perf_event_exit_cpu()
  +--> perf_event_exit_cpu_context()
       +--> __perf_event_exit_context()
	    +--> __perf_remove_from_context()
	         +--> event_sched_out()
	              +--> cpumsf_pmu_del()
	                   +--> cpumsf_pmu_stop()
                                +--> hw_perf_event_update()

to stop and remove the event. During removal of the event, the
sampling device driver tries to read out the remaining samples from
the sample data buffers (SDBs). But they have already been freed
(and may have been re-assigned). This may lead to a use after free
situation in which case the samples are most likely invalid. In the
best case the memory has not been reassigned and still contains
valid data.

Remedy this situation and check if the CPU is still in reserved
state (bit PMU_F_RESERVED set). In this case the SDBs have not been
released an contain valid data. This is always the case when
the event is removed (and no CPU hotplug off occured).
If the PMU_F_RESERVED bit is not set, the SDB buffers are gone.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Hendrik Brueckner <brueckner@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/s390/kernel/perf_cpum_sf.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/s390/kernel/perf_cpum_sf.c b/arch/s390/kernel/perf_cpum_sf.c
index 48fa9471660a8..cf9a65cdbedb6 100644
--- a/arch/s390/kernel/perf_cpum_sf.c
+++ b/arch/s390/kernel/perf_cpum_sf.c
@@ -1922,7 +1922,9 @@ static void cpumsf_pmu_stop(struct perf_event *event, int flags)
 	event->hw.state |= PERF_HES_STOPPED;
 
 	if ((flags & PERF_EF_UPDATE) && !(event->hw.state & PERF_HES_UPTODATE)) {
-		hw_perf_event_update(event, 1);
+		/* CPU hotplug off removes SDBs. No samples to extract. */
+		if (cpuhw->flags & PMU_F_RESERVED)
+			hw_perf_event_update(event, 1);
 		event->hw.state |= PERF_HES_UPTODATE;
 	}
 	perf_pmu_enable(event->pmu);
-- 
2.43.0