| Message ID | 1471505714-30049-1-git-send-email-andrew.donnellan@au1.ibm.com |
|---|---|
| State | Changes Requested |
| Headers | show |
> On Aug 18, 2016, at 2:35 AM, Andrew Donnellan <andrew.donnellan@au1.ibm.com> wrote: > > When cxl removes a vPHB, it's possible that the pci_controller may be freed > before all references to the devices on the vPHB have been released. This > in turn causes an invalid memory access when the devices are eventually > released, as pcibios_release_device() attempts to call the phb's > release_device hook. > > In cxl_pci_vphb_remove(), remove the existing call to > pcibios_free_controller(). Instead, use > pcibios_free_controller_deferred() to free the pci_controller after all > devices have been released. Export pci_set_host_bridge_release() so we can > do this. > > Cc: stable@vger.kernel.org > Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com> Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com> -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi Andrew,
[auto build test ERROR on char-misc/char-misc-testing]
[also build test ERROR on v4.8-rc3 next-20160822]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
[Suggest to use git(>=2.9.0) format-patch --base=<commit> (or --base=auto for convenience) to record what (public, well-known) commit your patch series was built on]
[Check https://git-scm.com/docs/git-format-patch for more information]
url: https://github.com/0day-ci/linux/commits/Andrew-Donnellan/cxl-use-pcibios_free_controller_deferred-when-removing-vPHBs/20160818-154040
config: powerpc-allmodconfig (attached as .config)
compiler: powerpc64-linux-gnu-gcc (Debian 5.4.0-6) 5.4.0 20160609
reproduce:
wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=powerpc
All errors (new ones prefixed by >>):
drivers/misc/cxl/vphb.c: In function 'cxl_pci_vphb_add':
>> drivers/misc/cxl/vphb.c:235:9: error: 'pcibios_free_controller_deferred' undeclared (first use in this function)
pcibios_free_controller_deferred,
^
drivers/misc/cxl/vphb.c:235:9: note: each undeclared identifier is reported only once for each function it appears in
vim +/pcibios_free_controller_deferred +235 drivers/misc/cxl/vphb.c
229 pcibios_scan_phb(phb);
230 if (phb->bus == NULL)
231 return -ENXIO;
232
233 /* Set release hook on root bus */
234 pci_set_host_bridge_release(to_pci_host_bridge(phb->bus->bridge),
> 235 pcibios_free_controller_deferred,
236 (void *) phb);
237
238 /* Claim resources. This might need some rework as well depending
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
Hi stable team, The following patch, which ended up upstream as 6f38a8b9a45833495dc878c335c5431cd98a16ed: On 18/08/16 17:35, Andrew Donnellan wrote: > When cxl removes a vPHB, it's possible that the pci_controller may be freed > before all references to the devices on the vPHB have been released. This > in turn causes an invalid memory access when the devices are eventually > released, as pcibios_release_device() attempts to call the phb's > release_device hook. > > In cxl_pci_vphb_remove(), remove the existing call to > pcibios_free_controller(). Instead, use > pcibios_free_controller_deferred() to free the pci_controller after all > devices have been released. Export pci_set_host_bridge_release() so we can > do this. > > Cc: stable@vger.kernel.org > Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com> > > --- > > This patch requires http://patchwork.ozlabs.org/patch/658324/. It should go > through the powerpc tree. This depends on 2dd9c11b9d4dfbd6c070eab7b81197f65e82f1a0 which didn't end up being tagged as cc: stable. It also ended up being applied in the wrong order in the powerpc/fixes tree... Thanks,
On Tue, 2016-08-30 at 11:58 +1000, Andrew Donnellan wrote: > Hi stable team, > > The following patch, which ended up upstream as > 6f38a8b9a45833495dc878c335c5431cd98a16ed: > > On 18/08/16 17:35, Andrew Donnellan wrote: > > > > When cxl removes a vPHB, it's possible that the pci_controller may be freed > > before all references to the devices on the vPHB have been released. This > > in turn causes an invalid memory access when the devices are eventually > > released, as pcibios_release_device() attempts to call the phb's > > release_device hook. > > > > In cxl_pci_vphb_remove(), remove the existing call to > > pcibios_free_controller(). Instead, use > > pcibios_free_controller_deferred() to free the pci_controller after all > > devices have been released. Export pci_set_host_bridge_release() so we can > > do this. > > > > Cc: stable@vger.kernel.org > > > > Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com> > > > > --- > > > > > > This patch requires http://patchwork.ozlabs.org/patch/658324/. It should go > > through the powerpc tree. > > This depends on 2dd9c11b9d4dfbd6c070eab7b81197f65e82f1a0 which didn't > end up being tagged as cc: stable. It also ended up being applied in the > wrong order in the powerpc/fixes tree... My fault. Best at this point is to also apply 2dd9c11b9d4dfbd6c070eab7b81197f65e82f1a0 to stable. Cheers, Ben. -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/misc/cxl/vphb.c b/drivers/misc/cxl/vphb.c index 7ada5f1..3519ace 100644 --- a/drivers/misc/cxl/vphb.c +++ b/drivers/misc/cxl/vphb.c @@ -230,6 +230,11 @@ int cxl_pci_vphb_add(struct cxl_afu *afu) if (phb->bus == NULL) return -ENXIO; + /* Set release hook on root bus */ + pci_set_host_bridge_release(to_pci_host_bridge(phb->bus->bridge), + pcibios_free_controller_deferred, + (void *) phb); + /* Claim resources. This might need some rework as well depending * whether we are doing probe-only or not, like assigning unassigned * resources etc... @@ -256,7 +261,10 @@ void cxl_pci_vphb_remove(struct cxl_afu *afu) afu->phb = NULL; pci_remove_root_bus(phb->bus); - pcibios_free_controller(phb); + /* + * We don't free phb here - that's handled by + * pcibios_free_controller_deferred() + */ } static bool _cxl_pci_is_vphb_device(struct pci_controller *phb) diff --git a/drivers/pci/host-bridge.c b/drivers/pci/host-bridge.c index 5f4a2e0..add6623 100644 --- a/drivers/pci/host-bridge.c +++ b/drivers/pci/host-bridge.c @@ -44,6 +44,7 @@ void pci_set_host_bridge_release(struct pci_host_bridge *bridge, bridge->release_fn = release_fn; bridge->release_data = release_data; } +EXPORT_SYMBOL_GPL(pci_set_host_bridge_release); void pcibios_resource_to_bus(struct pci_bus *bus, struct pci_bus_region *region, struct resource *res)
When cxl removes a vPHB, it's possible that the pci_controller may be freed before all references to the devices on the vPHB have been released. This in turn causes an invalid memory access when the devices are eventually released, as pcibios_release_device() attempts to call the phb's release_device hook. In cxl_pci_vphb_remove(), remove the existing call to pcibios_free_controller(). Instead, use pcibios_free_controller_deferred() to free the pci_controller after all devices have been released. Export pci_set_host_bridge_release() so we can do this. Cc: stable@vger.kernel.org Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com> --- This patch requires http://patchwork.ozlabs.org/patch/658324/. It should go through the powerpc tree. --- drivers/misc/cxl/vphb.c | 10 +++++++++- drivers/pci/host-bridge.c | 1 + 2 files changed, 10 insertions(+), 1 deletion(-)