All of entries are freed in a loop, however, the freed entry is accessed
by list_del() after the loop.
When epf driver that includes pci-epf-test unload, the pci_epf_remove_cfs()
is called, and occurred the use after free. Therefore, kernel panics
randomly after or while the module unloading.
I tested this patch with r8a77951-Salvator-xs boards.
Fixes: ef1433f ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry")
Signed-off-by: Shunsuke Mie <[email protected]>
---
drivers/pci/endpoint/pci-epf-core.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/pci/endpoint/pci-epf-core.c b/drivers/pci/endpoint/pci-epf-core.c
index e9289d10f822..538e902b0ba6 100644
--- a/drivers/pci/endpoint/pci-epf-core.c
+++ b/drivers/pci/endpoint/pci-epf-core.c
@@ -202,8 +202,10 @@ static void pci_epf_remove_cfs(struct pci_epf_driver *driver)
return;
mutex_lock(&pci_epf_mutex);
- list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry)
+ list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry) {
+ list_del(&group->group_entry);
pci_ep_cfs_remove_epf_group(group);
+ }
list_del(&driver->epf_group);
mutex_unlock(&pci_epf_mutex);
}
--
2.17.1
On Mon, Jun 21, 2021 at 04:00:58PM +0900, Shunsuke Mie wrote:
> All of entries are freed in a loop, however, the freed entry is accessed
> by list_del() after the loop.
>
> When epf driver that includes pci-epf-test unload, the pci_epf_remove_cfs()
> is called, and occurred the use after free. Therefore, kernel panics
> randomly after or while the module unloading.
>
> I tested this patch with r8a77951-Salvator-xs boards.
>
> Fixes: ef1433f ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry")
> Signed-off-by: Shunsuke Mie <[email protected]>
> ---
> drivers/pci/endpoint/pci-epf-core.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
<formletter>
This is not the correct way to submit patches for inclusion in the
stable kernel tree. Please read:
https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
for how to do this properly.
</formletter>
[dropped stable, erroneously added to the CC list]
On Mon, Jun 21, 2021 at 04:00:58PM +0900, Shunsuke Mie wrote:
> All of entries are freed in a loop, however, the freed entry is accessed
> by list_del() after the loop.
>
> When epf driver that includes pci-epf-test unload, the pci_epf_remove_cfs()
> is called, and occurred the use after free. Therefore, kernel panics
> randomly after or while the module unloading.
>
> I tested this patch with r8a77951-Salvator-xs boards.
>
> Fixes: ef1433f ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry")
> Signed-off-by: Shunsuke Mie <[email protected]>
> ---
> drivers/pci/endpoint/pci-epf-core.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
Kishon, please review this patch, thanks.
Lorenzo
>
> diff --git a/drivers/pci/endpoint/pci-epf-core.c b/drivers/pci/endpoint/pci-epf-core.c
> index e9289d10f822..538e902b0ba6 100644
> --- a/drivers/pci/endpoint/pci-epf-core.c
> +++ b/drivers/pci/endpoint/pci-epf-core.c
> @@ -202,8 +202,10 @@ static void pci_epf_remove_cfs(struct pci_epf_driver *driver)
> return;
>
> mutex_lock(&pci_epf_mutex);
> - list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry)
> + list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry) {
> + list_del(&group->group_entry);
> pci_ep_cfs_remove_epf_group(group);
> + }
> list_del(&driver->epf_group);
> mutex_unlock(&pci_epf_mutex);
> }
> --
> 2.17.1
>
Hi Shunsuke,
On 21/06/21 12:30 pm, Shunsuke Mie wrote:
> All of entries are freed in a loop, however, the freed entry is accessed
> by list_del() after the loop.
>
> When epf driver that includes pci-epf-test unload, the pci_epf_remove_cfs()
> is called, and occurred the use after free. Therefore, kernel panics
> randomly after or while the module unloading.
>
> I tested this patch with r8a77951-Salvator-xs boards.
Can you provide the crash dump?
>
> Fixes: ef1433f ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry")
> Signed-off-by: Shunsuke Mie <[email protected]>
> ---
> drivers/pci/endpoint/pci-epf-core.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/pci/endpoint/pci-epf-core.c b/drivers/pci/endpoint/pci-epf-core.c
> index e9289d10f822..538e902b0ba6 100644
> --- a/drivers/pci/endpoint/pci-epf-core.c
> +++ b/drivers/pci/endpoint/pci-epf-core.c
> @@ -202,8 +202,10 @@ static void pci_epf_remove_cfs(struct pci_epf_driver *driver)
> return;
>
> mutex_lock(&pci_epf_mutex);
> - list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry)
> + list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry) {
> + list_del(&group->group_entry);
Need full crash dump to see if this is really required or not. Ideally this
should be handled by configfs (or a configfs API could be used).
Thanks,
Kishon
Hi Kishon,
2021年12月7日(火) 16:26 Kishon Vijay Abraham I <[email protected]>:
>
> Hi Shunsuke,
>
> On 21/06/21 12:30 pm, Shunsuke Mie wrote:
> > All of entries are freed in a loop, however, the freed entry is accessed
> > by list_del() after the loop.
> >
> > When epf driver that includes pci-epf-test unload, the pci_epf_remove_cfs()
> > is called, and occurred the use after free. Therefore, kernel panics
> > randomly after or while the module unloading.
> >
> > I tested this patch with r8a77951-Salvator-xs boards.
>
> Can you provide the crash dump?
Ok. However, that is use-after-free bug. so, the crash log is changed randomly.
I'll send some crash dumps.
> >
> > Fixes: ef1433f ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry")
> > Signed-off-by: Shunsuke Mie <[email protected]>
> > ---
> > drivers/pci/endpoint/pci-epf-core.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/pci/endpoint/pci-epf-core.c b/drivers/pci/endpoint/pci-epf-core.c
> > index e9289d10f822..538e902b0ba6 100644
> > --- a/drivers/pci/endpoint/pci-epf-core.c
> > +++ b/drivers/pci/endpoint/pci-epf-core.c
> > @@ -202,8 +202,10 @@ static void pci_epf_remove_cfs(struct pci_epf_driver *driver)
> > return;
> >
> > mutex_lock(&pci_epf_mutex);
> > - list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry)
> > + list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry) {
> > + list_del(&group->group_entry);
>
> Need full crash dump to see if this is really required or not. Ideally this
> should be handled by configfs (or a configfs API could be used).
>
> Thanks,
> Kishon
Hi Kishon,
This is the crash log. I think you can reproduce this problem by loading and
unloading the pci-epf-test modules some times, but I don't think you'll see
similar logs in your environment.
The pci_ep_cfs_remove_epf_group function calls kfree for the `group`, however
a member of group, `group_entry``, is accessed by a list_del() after the loop.
The patch changes to remove list entries to prevent the accessing after freed
object.
Thanks,
Shunsuke
2021年12月7日(火) 16:44 Shunsuke Mie <[email protected]>:
>
> Hi Kishon,
>
> 2021年12月7日(火) 16:26 Kishon Vijay Abraham I <[email protected]>:
> >
> > Hi Shunsuke,
> >
> > On 21/06/21 12:30 pm, Shunsuke Mie wrote:
> > > All of entries are freed in a loop, however, the freed entry is accessed
> > > by list_del() after the loop.
> > >
> > > When epf driver that includes pci-epf-test unload, the pci_epf_remove_cfs()
> > > is called, and occurred the use after free. Therefore, kernel panics
> > > randomly after or while the module unloading.
> > >
> > > I tested this patch with r8a77951-Salvator-xs boards.
> >
> > Can you provide the crash dump?
> Ok. However, that is use-after-free bug. so, the crash log is changed randomly.
> I'll send some crash dumps.
> > >
> > > Fixes: ef1433f ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry")
> > > Signed-off-by: Shunsuke Mie <[email protected]>
> > > ---
> > > drivers/pci/endpoint/pci-epf-core.c | 4 +++-
> > > 1 file changed, 3 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/pci/endpoint/pci-epf-core.c b/drivers/pci/endpoint/pci-epf-core.c
> > > index e9289d10f822..538e902b0ba6 100644
> > > --- a/drivers/pci/endpoint/pci-epf-core.c
> > > +++ b/drivers/pci/endpoint/pci-epf-core.c
> > > @@ -202,8 +202,10 @@ static void pci_epf_remove_cfs(struct pci_epf_driver *driver)
> > > return;
> > >
> > > mutex_lock(&pci_epf_mutex);
> > > - list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry)
> > > + list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry) {
> > > + list_del(&group->group_entry);
> >
> > Need full crash dump to see if this is really required or not. Ideally this
> > should be handled by configfs (or a configfs API could be used).
> >
> > Thanks,
> > Kishon