PCI: s390: Fix user-after-free and clean up
Niklas Schnelle
maintainerv2 cover PCI: s390: Fix user-after-free and clean up
Hi Bjorn, Hi Lukas,
This is v2 of my fix for a use-after-free of PCI MMIO resources in the s390 PCI layer. As Bjorn found some redundant pci_bus_add_device() uses and function zero special treatment I've added 3 cleanup patches in addition to the fix itself. These are independent but in my opinion make it easier to see the struct zpci_dev interactions with the common PCI code. If you prefer I can of course split them off. As discussed this version still uses the pci_rescan_remove lock to allow backporting but I'll be looking into adding a more specific resource lock as a follow up.
Thanks,
Niklas
Changes since v1:
- Added clean up patches inspired by Bjorn's questions
- Removed return at end of function returning void
Niklas Schnelle (4):
PCI: s390: Fix use-after-free of PCI resources with per-function
hotplug
s390/pci: only add specific device in zpci_bus_scan_device() s390/pci: remove redundant pci_bus_add_devices() on new bus s390/pci: clean up left over special treatment for function zero
arch/s390/pci/pci.c | 39 +++++++++++++-------------------------- arch/s390/pci/pci_bus.c | 16 ++++++---------- arch/s390/pci/pci_bus.h | 3 +-- drivers/pci/bus.c | 21 +++++++++++++++++++++ include/linux/pci.h | 1 + 5 files changed, 42 insertions(+), 38 deletions(-)
Signature · 1 line
2.37.2
Niklas Schnelle
maintainerPATCH v2 1/4 PCI: s390: Fix use-after-free of PCI resources with per-function hotplug
On s390 PCI functions may be hotplugged individually even when they belong to a multi-function device. In particular on an SR-IOV device VFs may be removed and later re-added.
In commit a50297cf8235 ("s390/pci: separate zbus creation from scanning") it was missed however that struct pci_bus and struct zpci_bus's resource list retained a reference to the PCI functions MMIO resources even though those resources are released and freed on hot-unplug. These stale resources may subsequently be claimed when the PCI function re-appears resulting in use-after-free.
One idea of fixing this use-after-free in s390 specific code that was investigated was to simply keep resources around from the moment a PCI function first appeared until the whole virtual PCI bus created for a multi-function device disappears. The problem with this however is that due to the requirement of artificial MMIO addreesses (address cookies) extra logic is then needed to keep the address cookies compatible on re-plug. At the same time the MMIO resources semantically belong to the PCI function so tying their lifecycle to the function seems more logical.
Instead a simpler approach is to remove the resources of an individually hot-unplugged PCI function from the PCI bus's resource list while keeping the resources of other PCI functions on the PCI bus untouched.
This is done by introducing pci_bus_remove_resource() to remove an individual resource. Similarly the resource also needs to be removed from the struct zpci_bus's resource list. It turns out however, that there is really no need to add the MMIO resources to the struct zpci_bus's resource list at all and instead we can simply use the zpci_bar_struct's resource pointer directly.
---
v1 -> v2:
- Remove return at the end of function returning void
arch/s390/pci/pci.c | 16 ++++++++++------ arch/s390/pci/pci_bus.c | 12 +++++------- arch/s390/pci/pci_bus.h | 3 +-- drivers/pci/bus.c | 21 +++++++++++++++++++++ include/linux/pci.h | 1 + 5 files changed, 38 insertions(+), 15 deletions(-)
arch/s390/pci/pci.c +10 -6
@@ -544,8 +544,7 @@ static struct resource *__alloc_res(struct zpci_dev *zdev, unsigned long start,return r; } -int zpci_setup_bus_resources (struct zpci_dev *zdev, -struct list_head *resources) +int zpci_setup_bus_resources (struct zpci_dev *zdev) {unsigned long addr, size, flags;struct resource *res;
@@ -581,7 +580,6 @@ int zpci_setup_bus_resources(struct zpci_dev *zdev,return -ENOMEM ; } zdev->bars[i].res = res; -pci_add_resource (resources, res); } zdev->has_resources =1 ;
@@ -590,17 +588,23 @@ int zpci_setup_bus_resources(struct zpci_dev *zdev,static void zpci_cleanup_bus_resources (struct zpci_dev *zdev) { +struct resource *res;int i; +pci_lock_rescan_remove ();for (i =0 ; i <PCI_STD_NUM_BARS ; i++) { -if (!zdev->bars[i].size || !zdev->bars[i].res) + res = zdev->bars[i].res; +if (!res)continue ; +release_resource (res); +pci_bus_remove_resource (zdev->zbus->bus, res);zpci_free_iomap (zdev, zdev->bars[i].map_idx); -release_resource (zdev->bars[i].res); -kfree (zdev->bars[i].res); + zdev->bars[i].res =NULL ; +kfree (res); } zdev->has_resources =0 ; +pci_unlock_rescan_remove (); }int pcibios_device_add (struct pci_dev *pdev)
arch/s390/pci/pci_bus.c +5 -7
@@ -41,9 +41,7 @@ static int zpci_nb_devices;*/ static int zpci_bus_prepare_device (struct zpci_dev *zdev) { -struct resource_entry *window, *n; -struct resource *res; -int rc; +int rc, i;if (!zdev_enabled (zdev)) { rc =zpci_enable_device (zdev);
@@ -57,10 +55,10 @@ static int zpci_bus_prepare_device(struct zpci_dev *zdev) }if (!zdev->has_resources) { -zpci_setup_bus_resources (zdev, &zdev->zbus->resources); -resource_list_for_each_entry_safe (window, n, &zdev->zbus->resources) { - res = window->res; -pci_bus_add_resource (zdev->zbus->bus, res,0 ); +zpci_setup_bus_resources (zdev); +for (i =0 ; i <PCI_STD_NUM_BARS ; i++) { +if (zdev->bars[i].res) +pci_bus_add_resource (zdev->zbus->bus, zdev->bars[i].res,0 ); } }
arch/s390/pci/pci_bus.h +1 -2
@@ -30,8 +30,7 @@ static inline void zpci_zdev_get(struct zpci_dev *zdev)int zpci_alloc_domain (int domain);void zpci_free_domain (int domain); -int zpci_setup_bus_resources (struct zpci_dev *zdev, -struct list_head *resources); +int zpci_setup_bus_resources (struct zpci_dev *zdev);static inline struct zpci_dev *zdev_from_bus (struct pci_bus *bus,unsigned int devfn)
drivers/pci/bus.c +21 -0
@@ -76,6 +76,27 @@ struct resource *pci_bus_resource_n(const struct pci_bus *bus, int n) }EXPORT_SYMBOL_GPL (pci_bus_resource_n); +void pci_bus_remove_resource (struct pci_bus *bus,struct resource *res) +{ +struct pci_bus_resource *bus_res, *tmp; +int i; + +for (i =0 ; i <PCI_BRIDGE_RESOURCE_NUM ; i++) { +if (bus->resource[i] == res) { + bus->resource[i] =NULL ; +return ; + } + } + +list_for_each_entry_safe (bus_res, tmp, &bus->resources, list) { +if (bus_res->res == res) { +list_del (&bus_res->list); +kfree (bus_res); +return ; + } + } +} +void pci_bus_remove_resources (struct pci_bus *bus) {int i;
include/linux/pci.h +1 -0
@@ -1438,6 +1438,7 @@ void pci_bus_add_resource(struct pci_bus *bus, struct resource *res, unsigned int flags);struct resource *pci_bus_resource_n (const struct pci_bus *bus,int n);void pci_bus_remove_resources (struct pci_bus *bus); +void pci_bus_remove_resource (struct pci_bus *bus,struct resource *res);int devm_request_pci_bus_resources (struct device *dev,struct list_head *resources);
Signature · 1 line
2.37.2
Re: [PATCH v2 1/4] PCI: s390: Fix use-after-free of PCI resources with per-function hotplug
Quoted 32 lines · Niklas Schnelle
On s390 PCI functions may be hotplugged individually even when they belong to a multi-function device. In particular on an SR-IOV device VFs may be removed and later re-added.
In commit
a50297cf8235("s390/pci: separate zbus creation from scanning") it was missed however that structpci_busand struct zpci_bus's resource list retained a reference to the PCI functions MMIO resources even though those resources are released and freed on hot-unplug. These stale resources may subsequently be claimed when the PCI function re-appears resulting in use-after-free.One idea of fixing this use-after-free in s390 specific code that was investigated was to simply keep resources around from the moment a PCI function first appeared until the whole virtual PCI bus created for a multi-function device disappears. The problem with this however is that due to the requirement of artificial MMIO addreesses (address cookies) extra logic is then needed to keep the address cookies compatible on re-plug. At the same time the MMIO resources semantically belong to the PCI function so tying their lifecycle to the function seems more logical.
Instead a simpler approach is to remove the resources of an individually hot-unplugged PCI function from the PCI bus's resource list while keeping the resources of other PCI functions on the PCI bus untouched.
This is done by introducing
pci_bus_remove_resource()to remove an individual resource. Similarly the resource also needs to be removed from the struct zpci_bus's resource list. It turns out however, that there is really no need to add the MMIO resources to the struct zpci_bus's resource list at all and instead we can simply use the zpci_bar_struct's resource pointer directly.
Quoted 2 lines
Fixes a50297cf8235 ("s390/pci: separate zbus creation from scanning")Signed-off-by Niklas Schnelle <[email protected]>
The meat of this is mostly in s390, so I think it makes more sense to merge via that tree. But let me know if you'd rather that I take it.
Quoted 151 lines
---
v1 -> v2:
- Remove return at the end of function returning voidarch/s390/pci/pci.c | 16 ++++++++++------ arch/s390/pci/pci_bus.c | 12 +++++------- arch/s390/pci/pci_bus.h | 3 +-- drivers/pci/bus.c | 21 +++++++++++++++++++++ include/linux/pci.h | 1 + 5 files changed, 38 insertions(+), 15 deletions(-)arch/s390/pci/pci.c +10
-6@@ -544,8 +544,7 @@ static struct resource *__alloc_res(struct zpci_dev *zdev, unsigned long start,return r; } -int zpci_setup_bus_resources (struct zpci_dev *zdev, -struct list_head *resources) +int zpci_setup_bus_resources (struct zpci_dev *zdev) {unsigned long addr, size, flags;struct resource *res;@@ -581,7 +580,6 @@ int zpci_setup_bus_resources(struct zpci_dev *zdev,return -ENOMEM ; } zdev->bars[i].res = res; -pci_add_resource (resources, res); } zdev->has_resources =1 ;@@ -590,17 +588,23 @@ int zpci_setup_bus_resources(struct zpci_dev *zdev,static void zpci_cleanup_bus_resources (struct zpci_dev *zdev) { +struct resource *res;int i; +pci_lock_rescan_remove ();for (i =0 ; i <PCI_STD_NUM_BARS ; i++) { -if (!zdev->bars[i].size || !zdev->bars[i].res) + res = zdev->bars[i].res; +if (!res)continue ; +release_resource (res); +pci_bus_remove_resource (zdev->zbus->bus, res);zpci_free_iomap (zdev, zdev->bars[i].map_idx); -release_resource (zdev->bars[i].res); -kfree (zdev->bars[i].res); + zdev->bars[i].res =NULL ; +kfree (res); } zdev->has_resources =0 ; +pci_unlock_rescan_remove (); }int pcibios_device_add (struct pci_dev *pdev)arch/s390/pci/pci_bus.c +5
-7@@ -41,9 +41,7 @@ static int zpci_nb_devices;*/ static int zpci_bus_prepare_device (struct zpci_dev *zdev) { -struct resource_entry *window, *n; -struct resource *res; -int rc; +int rc, i;if (!zdev_enabled (zdev)) { rc =zpci_enable_device (zdev);@@ -57,10 +55,10 @@ static int zpci_bus_prepare_device(struct zpci_dev *zdev) }if (!zdev->has_resources) { -zpci_setup_bus_resources (zdev, &zdev->zbus->resources); -resource_list_for_each_entry_safe (window, n, &zdev->zbus->resources) { - res = window->res; -pci_bus_add_resource (zdev->zbus->bus, res,0 ); +zpci_setup_bus_resources (zdev); +for (i =0 ; i <PCI_STD_NUM_BARS ; i++) { +if (zdev->bars[i].res) +pci_bus_add_resource (zdev->zbus->bus, zdev->bars[i].res,0 ); } }arch/s390/pci/pci_bus.h +1
-2@@ -30,8 +30,7 @@ static inline void zpci_zdev_get(struct zpci_dev *zdev)int zpci_alloc_domain (int domain);void zpci_free_domain (int domain); -int zpci_setup_bus_resources (struct zpci_dev *zdev, -struct list_head *resources); +int zpci_setup_bus_resources (struct zpci_dev *zdev);static inline struct zpci_dev *zdev_from_bus (struct pci_bus *bus,unsigned int devfn)drivers/pci/bus.c +21
-0@@ -76,6 +76,27 @@ struct resource *pci_bus_resource_n(const struct pci_bus *bus, int n) }EXPORT_SYMBOL_GPL (pci_bus_resource_n); +void pci_bus_remove_resource (struct pci_bus *bus,struct resource *res) +{ +struct pci_bus_resource *bus_res, *tmp; +int i; + +for (i =0 ; i <PCI_BRIDGE_RESOURCE_NUM ; i++) { +if (bus->resource[i] == res) { + bus->resource[i] =NULL ; +return ; + } + } + +list_for_each_entry_safe (bus_res, tmp, &bus->resources, list) { +if (bus_res->res == res) { +list_del (&bus_res->list); +kfree (bus_res); +return ; + } + } +} +void pci_bus_remove_resources (struct pci_bus *bus) {int i;include/linux/pci.h +1
-0@@ -1438,6 +1438,7 @@ void pci_bus_add_resource(struct pci_bus *bus, struct resource *res, unsigned int flags);struct resource *pci_bus_resource_n (const struct pci_bus *bus,int n);void pci_bus_remove_resources (struct pci_bus *bus); +void pci_bus_remove_resource (struct pci_bus *bus,struct resource *res);int devm_request_pci_bus_resources (struct device *dev,struct list_head *resources);Signature · 1 line
2.37.2
Niklas Schnelle
maintainerRe: [PATCH v2 1/4] PCI: s390: Fix use-after-free of PCI resources with per-function hotplug
Quoted 33 lines · Bjorn Helgaas
Quoted 32 lines · Niklas Schnelle
On s390 PCI functions may be hotplugged individually even when they belong to a multi-function device. In particular on an SR-IOV device VFs may be removed and later re-added.
In commit
a50297cf8235("s390/pci: separate zbus creation from scanning") it was missed however that structpci_busand struct zpci_bus's resource list retained a reference to the PCI functions MMIO resources even though those resources are released and freed on hot-unplug. These stale resources may subsequently be claimed when the PCI function re-appears resulting in use-after-free.One idea of fixing this use-after-free in s390 specific code that was investigated was to simply keep resources around from the moment a PCI function first appeared until the whole virtual PCI bus created for a multi-function device disappears. The problem with this however is that due to the requirement of artificial MMIO addreesses (address cookies) extra logic is then needed to keep the address cookies compatible on re-plug. At the same time the MMIO resources semantically belong to the PCI function so tying their lifecycle to the function seems more logical.
Instead a simpler approach is to remove the resources of an individually hot-unplugged PCI function from the PCI bus's resource list while keeping the resources of other PCI functions on the PCI bus untouched.
This is done by introducing
pci_bus_remove_resource()to remove an individual resource. Similarly the resource also needs to be removed from the struct zpci_bus's resource list. It turns out however, that there is really no need to add the MMIO resources to the struct zpci_bus's resource list at all and instead we can simply use the zpci_bar_struct's resource pointer directly.
Quoted 9 lines
Quoted 2 lines
Fixes a50297cf8235 ("s390/pci: separate zbus creation from scanning")Signed-off-by Niklas Schnelle <[email protected]>Acked-by Bjorn Helgaas <[email protected]>The meat of this is mostly in s390, so I think it makes more sense to merge via that tree. But let me know if you'd rather that I take it.
Thanks for taking a look and the valuable suggestions. I'll coordinate with Vasily to take this via the s390 tree. As for the locking I agree it is out of scope for this series. Meant more that the resource handling might be a good place to start splitting up the pci_rescan_remove_lock and that I might take a look at that if I find the time which of course we're all lacking.
Regards,
Niklas
Vasily Gorbik
maintainerRe: [PATCH v2 1/4] PCI: s390: Fix use-after-free of PCI resources with per-function hotplug
In reply to Bjorn HelgaasQuoted 33 lines · Bjorn Helgaas
Quoted 32 lines · Niklas Schnelle
On s390 PCI functions may be hotplugged individually even when they belong to a multi-function device. In particular on an SR-IOV device VFs may be removed and later re-added.
In commit
a50297cf8235("s390/pci: separate zbus creation from scanning") it was missed however that structpci_busand struct zpci_bus's resource list retained a reference to the PCI functions MMIO resources even though those resources are released and freed on hot-unplug. These stale resources may subsequently be claimed when the PCI function re-appears resulting in use-after-free.One idea of fixing this use-after-free in s390 specific code that was investigated was to simply keep resources around from the moment a PCI function first appeared until the whole virtual PCI bus created for a multi-function device disappears. The problem with this however is that due to the requirement of artificial MMIO addreesses (address cookies) extra logic is then needed to keep the address cookies compatible on re-plug. At the same time the MMIO resources semantically belong to the PCI function so tying their lifecycle to the function seems more logical.
Instead a simpler approach is to remove the resources of an individually hot-unplugged PCI function from the PCI bus's resource list while keeping the resources of other PCI functions on the PCI bus untouched.
This is done by introducing
pci_bus_remove_resource()to remove an individual resource. Similarly the resource also needs to be removed from the struct zpci_bus's resource list. It turns out however, that there is really no need to add the MMIO resources to the struct zpci_bus's resource list at all and instead we can simply use the zpci_bar_struct's resource pointer directly.
Quoted 7 lines
Quoted 2 lines
Fixes a50297cf8235 ("s390/pci: separate zbus creation from scanning")Signed-off-by Niklas Schnelle <[email protected]>Acked-by Bjorn Helgaas <[email protected]>The meat of this is mostly in s390, so I think it makes more sense to merge via that tree. But let me know if you'd rather that I take it.
I'll take it via s390 tree. Thanks
Re: [PATCH v2 1/4] PCI: s390: Fix use-after-free of PCI resources with per-function hotplug
In reply to Niklas SchnelleQuoted 32 lines · Niklas Schnelle
On s390 PCI functions may be hotplugged individually even when they belong to a multi-function device. In particular on an SR-IOV device VFs may be removed and later re-added.
In commit
a50297cf8235("s390/pci: separate zbus creation from scanning") it was missed however that structpci_busand struct zpci_bus's resource list retained a reference to the PCI functions MMIO resources even though those resources are released and freed on hot-unplug. These stale resources may subsequently be claimed when the PCI function re-appears resulting in use-after-free.One idea of fixing this use-after-free in s390 specific code that was investigated was to simply keep resources around from the moment a PCI function first appeared until the whole virtual PCI bus created for a multi-function device disappears. The problem with this however is that due to the requirement of artificial MMIO addreesses (address cookies) extra logic is then needed to keep the address cookies compatible on re-plug. At the same time the MMIO resources semantically belong to the PCI function so tying their lifecycle to the function seems more logical.
Instead a simpler approach is to remove the resources of an individually hot-unplugged PCI function from the PCI bus's resource list while keeping the resources of other PCI functions on the PCI bus untouched.
This is done by introducing
pci_bus_remove_resource()to remove an individual resource. Similarly the resource also needs to be removed from the struct zpci_bus's resource list. It turns out however, that there is really no need to add the MMIO resources to the struct zpci_bus's resource list at all and instead we can simply use the zpci_bar_struct's resource pointer directly.
Quoted 2 lines
Fixes a50297cf8235 ("s390/pci: separate zbus creation from scanning")Signed-off-by Niklas Schnelle <[email protected]>
Re: [PATCH v2 1/4] PCI: s390: Fix use-after-free of PCI resources with per-function hotplug
In reply to Niklas SchnelleQuoted 16 lines · Niklas Schnelle
drivers/pci/bus.c +10
-0@@ -76,6 +76,27 @@ struct resource *pci_bus_resource_n(const struct pci_bus *bus, int n) }EXPORT_SYMBOL_GPL (pci_bus_resource_n); +void pci_bus_remove_resource (struct pci_bus *bus,struct resource *res) +{ +struct pci_bus_resource *bus_res, *tmp; +int i; + +for (i =0 ; i <PCI_BRIDGE_RESOURCE_NUM ; i++) { +if (bus->resource[i] == res) { + bus->resource[i] =NULL ; +return ; + }
drivers/pci/bus.c struct resource *pci_bus_resource_n(const struct pci_bus *bus, int n) · line 98
Quoted 10 lines
drivers/pci/bus.c +10
-0+ } + +list_for_each_entry_safe (bus_res, tmp, &bus->resources, list) { +if (bus_res->res == res) { +list_del (&bus_res->list); +kfree (bus_res); +return ; + } + } +}
I realize this has already been applied so s390.git/master, but nevertheless would like to point out there's a handy pci_bus_for_each_resource() helper which could have been used here instead of the for-loop.
Sorry for chiming in late, I just spotted this while flushing out my inbox.
Adding Andy to cc: who's been active in this area recently.
Thanks,
Lukas
Re: [PATCH v2 1/4] PCI: s390: Fix use-after-free of PCI resources with per-function hotplug
Quoted 1 line · Lukas Wunner
On Mon, Mar 06, 2023 at 04:10:11PM +0100, Niklas Schnelle wrote:
...
Quoted 17 lines
Quoted 17 lines
drivers/pci/bus.c +17
-0+void pci_bus_remove_resource (struct pci_bus *bus,struct resource *res) +{ +struct pci_bus_resource *bus_res, *tmp; +int i; + +for (i =0 ; i <PCI_BRIDGE_RESOURCE_NUM ; i++) { +if (bus->resource[i] == res) { + bus->resource[i] =NULL ; +return ; + } + } + +list_for_each_entry_safe (bus_res, tmp, &bus->resources, list) { +if (bus_res->res == res) { +list_del (&bus_res->list); +kfree (bus_res); +return ;
Quoted 8 lines
drivers/pci/bus.c struct resource *pci_bus_resource_n(const struct pci_bus *bus, int n) · line 96
Quoted 3 lines
drivers/pci/bus.c +3
-0+ } + } +}I realize this has already been applied so s390.git/master, but nevertheless would like to point out there's a handy
pci_bus_for_each_resource()helper which could have been used here instead of the for-loop.
Actually in this case it's not possible. The above code nullifies the matched resource or removes it from the list. We don't have any good iterator inside pci_bus_for_each_resource() to do the latter.
There might be other places in the kernel code where something like above is used and we can split a separate helper exactly for matching cases, but with the above context I dunno we need to do anything right now.
P.S. Thanks for Cc'ing me.
Signature · 2 lines
With Best Regards, Andy Shevchenko
Niklas Schnelle
maintainerPATCH v2 2/4 s390/pci: only add specific device in zpci_bus_scan_device()
In reply to Niklas SchnelleAs the name suggests zpci_bus_scan_device() is used to scan a specific device and thus pci_bus_add_device() for that device is sufficient. Furthermore move this call inside the rescan/remove locking.
--- arch/s390/pci/pci_bus.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
arch/s390/pci/pci_bus.c +1 -2
@@ -85,9 +85,8 @@ int zpci_bus_scan_device(struct zpci_dev *zdev)if (!pdev)return -ENODEV ; -pci_bus_add_device (pdev);pci_lock_rescan_remove (); -pci_bus_add_devices (zdev->zbus->bus); +pci_bus_add_device (pdev);pci_unlock_rescan_remove ();return 0 ;
Signature · 1 line
2.37.2
Re: [PATCH v2 2/4] s390/pci: only add specific device in zpci_bus_scan_device()
Quoted 5 lines · Niklas Schnelle
As the name suggests
zpci_bus_scan_device()is used to scan a specific device and thuspci_bus_add_device()for that device is sufficient. Furthermore move this call inside the rescan/remove locking.Signed-off-by Niklas Schnelle <[email protected]>
I think a suggested by tag from Bjorn is appropriate here
Otherwise:
Quoted 19 lines
--- arch/s390/pci/pci_bus.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)arch/s390/pci/pci_bus.c +1
-2@@ -85,9 +85,8 @@ int zpci_bus_scan_device(struct zpci_dev *zdev)if (!pdev)return -ENODEV ; -pci_bus_add_device (pdev);pci_lock_rescan_remove (); -pci_bus_add_devices (zdev->zbus->bus); +pci_bus_add_device (pdev);pci_unlock_rescan_remove ();return 0 ;
Niklas Schnelle
maintainerPATCH v2 3/4 s390/pci: remove redundant pci_bus_add_devices() on new bus
In reply to Niklas SchnelleThe pci_bus_add_devices() call in zpci_bus_create_pci_bus() is without function since at this point no device could have been added to the freshly created PCI bus.
--- arch/s390/pci/pci_bus.c | 1 - 1 file changed, 1 deletion(-)
arch/s390/pci/pci_bus.c +0 -1
@@ -210,7 +210,6 @@ static int zpci_bus_create_pci_bus(struct zpci_bus *zbus, struct zpci_dev *fr, s } zbus->bus = bus; -pci_bus_add_devices (bus);return 0 ; }
Signature · 1 line
2.37.2
Re: [PATCH v2 3/4] s390/pci: remove redundant pci_bus_add_devices() on new bus
Quoted 5 lines · Niklas Schnelle
The
pci_bus_add_devices()call inzpci_bus_create_pci_bus()is without function since at this point no device could have been added to the freshly created PCI bus.
Niklas Schnelle
maintainerPATCH v2 4/4 s390/pci: clean up left over special treatment for function zero
In reply to Niklas SchnellePrior to commit 960ac3626487 ("s390/pci: allow zPCI zbus without a function zero") enabling and scanning a PCI function had to potentially be postponed until the function with devfn zero on that bus was plugged. While the commit removed the waiting itself extra code to scan all functions on the PCI bus once function zero appeared was missed. Remove that code and the outdated comments about waiting for function zero.
--- arch/s390/pci/pci.c | 23 +++-------------------- 1 file changed, 3 insertions(+), 20 deletions(-)
arch/s390/pci/pci.c +3 -20
@@ -874,32 +874,15 @@ bool zpci_is_device_configured(struct zpci_dev *zdev)* @fh: The general function handle supplied by the platform * * Given a device in the configuration state Configured, enables, scans and -* adds it to the common code PCI subsystem if possible. If the PCI device is -* parked because we can not yet create a PCI bus because we have not seen -* function 0, it is ignored but will be scanned once function 0 appears. -* If any failure occurs, the zpci_dev is left disabled. +* adds it to the common code PCI subsystem if possible.If any failure occurs, +* the zpci_dev is left disabled. * * Return: 0 on success, or an error code otherwise */ int zpci_scan_configured_device (struct zpci_dev *zdev,u32 fh) { -int rc; -zpci_update_fh (zdev, fh); -/* the PCI function will be scanned once function 0 appears */ -if (!zdev->zbus->bus) -return 0 ; - -/* For function 0 on a multi-function bus scan whole bus as we might -* have to pick up existing functions waiting for it to allow creating -* the PCI bus -*/ -if (zdev->devfn ==0 && zdev->zbus->multifunction) - rc =zpci_bus_scan_bus (zdev->zbus); -else - rc =zpci_bus_scan_device (zdev); - -return rc; +return zpci_bus_scan_device (zdev); }/**
Signature · 1 line
2.37.2
Re: [PATCH v2 4/4] s390/pci: clean up left over special treatment for function zero
Quoted 18 lines · Niklas Schnelle
Prior to commit
960ac3626487("s390/pci: allow zPCI zbus without a function zero") enabling and scanning a PCI function had to potentially be postponed until the function with devfn zero on that bus was plugged. While the commit removed the waiting itself extra code to scan all functions on the PCI bus once function zero appeared was missed. Remove that code and the outdated comments about waiting for function zero.Signed-off-by Niklas Schnelle <[email protected]>--- arch/s390/pci/pci.c | 23 +++-------------------- 1 file changed, 3 insertions(+), 20 deletions(-)arch/s390/pci/pci.c +0
-0
arch/s390/pci/pci.c bool zpci_is_device_configured(struct zpci_dev *zdev) · line 877
Quoted 8 lines
arch/s390/pci/pci.c +1
-4@@ -874,32 +874,15 @@ bool zpci_is_device_configured(struct zpci_dev *zdev)* @fh: The general function handle supplied by the platform * * Given a device in the configuration state Configured, enables, scans and -* adds it to the common code PCI subsystem if possible. If the PCI device is -* parked because we can not yet create a PCI bus because we have not seen -* function 0, it is ignored but will be scanned once function 0 appears. -* If any failure occurs, the zpci_dev is left disabled. +* adds it to the common code PCI subsystem if possible.If any failure occurs,
Nit: missing space after period here
Also, while you're at this, commit 960ac3626487 removed the function 0 check from zpci_bus_scan_bus but the comment block for zpci_bus_scan_bus is still outdated. Please fix also with this patch.
With that: