| Message ID | 20210812132144.791268-1-kw@linux.com |
|---|---|
| State | New |
| Headers | show |
| Series | PCI/sysfs: Use correct variable for the legacy_mem sysfs object | expand |
[+to Greg, please update sysfs_defferred_iomem_get_mapping-5.15] On Thu, Aug 12, 2021 at 01:21:44PM +0000, Krzysztof Wilczyński wrote: > Two legacy PCI sysfs objects "legacy_io" and "legacy_mem" were updated > to use an unified address space in the commit 636b21b50152 ("PCI: Revoke > mappings like devmem"). This allows for revocations to be managed from > a single place when drivers want to take over and mmap() a /dev/mem > range. > > Following the update, both of the sysfs objects should leverage the > iomem_get_mapping() function to get an appropriate address range, but > only the "legacy_io" has been correctly updated - the second attribute > seems to be using a wrong variable to pass the iomem_get_mapping() > function to. > > Thus, correct the variable name used so that the "legacy_mem" sysfs > object would also correctly call the iomem_get_mapping() function. > > Fixes: 636b21b50152 ("PCI: Revoke mappings like devmem") > Signed-off-by: Krzysztof Wilczyński <kw@linux.com> Acked-by: Bjorn Helgaas <bhelgaas@google.com> Ouch. This needs to be applied to any -stable trees that contain 636b21b50152, which was merged in v5.12. It *also* needs to be applied with the obvious updates to Greg's sysfs_defferred_iomem_get_mapping-5.15 branch because it was changed by f06aff924f97 ("sysfs: Rename struct bin_attribute member to f_mapping") [1] [1] https://git.kernel.org/cgit/linux/kernel/git/gregkh/driver-core.git/commit/?id=f06aff924f97 > --- > drivers/pci/pci-sysfs.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c > index 5d63df7c1820..7bbf2673c7f2 100644 > --- a/drivers/pci/pci-sysfs.c > +++ b/drivers/pci/pci-sysfs.c > @@ -978,7 +978,7 @@ void pci_create_legacy_files(struct pci_bus *b) > b->legacy_mem->size = 1024*1024; > b->legacy_mem->attr.mode = 0600; > b->legacy_mem->mmap = pci_mmap_legacy_mem; > - b->legacy_io->mapping = iomem_get_mapping(); > + b->legacy_mem->mapping = iomem_get_mapping(); > pci_adjust_legacy_attr(b, pci_mmap_mem); > error = device_create_bin_file(&b->dev, b->legacy_mem); > if (error) > -- > 2.32.0 >
On Thu, Aug 12, 2021 at 11:17:12AM -0500, Bjorn Helgaas wrote: > [+to Greg, please update sysfs_defferred_iomem_get_mapping-5.15] Actually, Greg, totally up to you, but if nobody else is depending on the sysfs_defferred_iomem_get_mapping-5.15 branch, another possibility would be for you to drop that branch and for me to merge the two patches on it + Krzysztof's fix below + (hopefully) Krzysztof's PCI static attribute work. That would make my v5.15 pull request easier because I simple-mindedly base all my branches on -rc1 while your branch is based on -rc3. But again, up to you. I put those two patches on a local branch: 94b34bc04c25 ("sysfs: Rename struct bin_attribute member to f_mapping") 97e9dada53f1 ("sysfs: Invoke iomem_get_mapping() from the sysfs open callback") in case that seems better to you. > On Thu, Aug 12, 2021 at 01:21:44PM +0000, Krzysztof Wilczyński wrote: > > Two legacy PCI sysfs objects "legacy_io" and "legacy_mem" were updated > > to use an unified address space in the commit 636b21b50152 ("PCI: Revoke > > mappings like devmem"). This allows for revocations to be managed from > > a single place when drivers want to take over and mmap() a /dev/mem > > range. > > > > Following the update, both of the sysfs objects should leverage the > > iomem_get_mapping() function to get an appropriate address range, but > > only the "legacy_io" has been correctly updated - the second attribute > > seems to be using a wrong variable to pass the iomem_get_mapping() > > function to. > > > > Thus, correct the variable name used so that the "legacy_mem" sysfs > > object would also correctly call the iomem_get_mapping() function. > > > > Fixes: 636b21b50152 ("PCI: Revoke mappings like devmem") > > Signed-off-by: Krzysztof Wilczyński <kw@linux.com> > > Acked-by: Bjorn Helgaas <bhelgaas@google.com> > > Ouch. This needs to be applied to any -stable trees that contain > 636b21b50152, which was merged in v5.12. > > It *also* needs to be applied with the obvious updates to Greg's > sysfs_defferred_iomem_get_mapping-5.15 branch because it was changed > by f06aff924f97 ("sysfs: Rename struct bin_attribute member to > f_mapping") [1] > > [1] https://git.kernel.org/cgit/linux/kernel/git/gregkh/driver-core.git/commit/?id=f06aff924f97 > > > --- > > drivers/pci/pci-sysfs.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c > > index 5d63df7c1820..7bbf2673c7f2 100644 > > --- a/drivers/pci/pci-sysfs.c > > +++ b/drivers/pci/pci-sysfs.c > > @@ -978,7 +978,7 @@ void pci_create_legacy_files(struct pci_bus *b) > > b->legacy_mem->size = 1024*1024; > > b->legacy_mem->attr.mode = 0600; > > b->legacy_mem->mmap = pci_mmap_legacy_mem; > > - b->legacy_io->mapping = iomem_get_mapping(); > > + b->legacy_mem->mapping = iomem_get_mapping(); > > pci_adjust_legacy_attr(b, pci_mmap_mem); > > error = device_create_bin_file(&b->dev, b->legacy_mem); > > if (error) > > -- > > 2.32.0 > >
On Thu, Aug 12, 2021 at 6:17 PM Bjorn Helgaas <helgaas@kernel.org> wrote: > > [+to Greg, please update sysfs_defferred_iomem_get_mapping-5.15] > > On Thu, Aug 12, 2021 at 01:21:44PM +0000, Krzysztof Wilczyński wrote: > > Two legacy PCI sysfs objects "legacy_io" and "legacy_mem" were updated > > to use an unified address space in the commit 636b21b50152 ("PCI: Revoke > > mappings like devmem"). This allows for revocations to be managed from > > a single place when drivers want to take over and mmap() a /dev/mem > > range. > > > > Following the update, both of the sysfs objects should leverage the > > iomem_get_mapping() function to get an appropriate address range, but > > only the "legacy_io" has been correctly updated - the second attribute > > seems to be using a wrong variable to pass the iomem_get_mapping() > > function to. > > > > Thus, correct the variable name used so that the "legacy_mem" sysfs > > object would also correctly call the iomem_get_mapping() function. > > > > Fixes: 636b21b50152 ("PCI: Revoke mappings like devmem") > > Signed-off-by: Krzysztof Wilczyński <kw@linux.com> > > Acked-by: Bjorn Helgaas <bhelgaas@google.com> > > Ouch. This needs to be applied to any -stable trees that contain > 636b21b50152, which was merged in v5.12. Ouch indeed. Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> > > It *also* needs to be applied with the obvious updates to Greg's > sysfs_defferred_iomem_get_mapping-5.15 branch because it was changed > by f06aff924f97 ("sysfs: Rename struct bin_attribute member to > f_mapping") [1] > > [1] https://git.kernel.org/cgit/linux/kernel/git/gregkh/driver-core.git/commit/?id=f06aff924f97 > > > --- > > drivers/pci/pci-sysfs.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c > > index 5d63df7c1820..7bbf2673c7f2 100644 > > --- a/drivers/pci/pci-sysfs.c > > +++ b/drivers/pci/pci-sysfs.c > > @@ -978,7 +978,7 @@ void pci_create_legacy_files(struct pci_bus *b) > > b->legacy_mem->size = 1024*1024; > > b->legacy_mem->attr.mode = 0600; > > b->legacy_mem->mmap = pci_mmap_legacy_mem; > > - b->legacy_io->mapping = iomem_get_mapping(); > > + b->legacy_mem->mapping = iomem_get_mapping(); > > pci_adjust_legacy_attr(b, pci_mmap_mem); > > error = device_create_bin_file(&b->dev, b->legacy_mem); > > if (error) > > -- > > 2.32.0 > >
On Thu, Aug 12, 2021 at 12:14:50PM -0500, Bjorn Helgaas wrote: > On Thu, Aug 12, 2021 at 11:17:12AM -0500, Bjorn Helgaas wrote: > > [+to Greg, please update sysfs_defferred_iomem_get_mapping-5.15] > > Actually, Greg, totally up to you, but if nobody else is depending on > the sysfs_defferred_iomem_get_mapping-5.15 branch, another possibility > would be for you to drop that branch and for me to merge the two > patches on it + Krzysztof's fix below + (hopefully) Krzysztof's PCI > static attribute work. I can not "drop" the branch as it is already merged into my driver-core-next branch that I can not rebase. I could revert it, but is that really needed? > That would make my v5.15 pull request easier because I simple-mindedly > base all my branches on -rc1 while your branch is based on -rc3. But > again, up to you. I put those two patches on a local branch: > > 94b34bc04c25 ("sysfs: Rename struct bin_attribute member to f_mapping") > 97e9dada53f1 ("sysfs: Invoke iomem_get_mapping() from the sysfs open callback") > > in case that seems better to you. I do not see what I can do with that here as it's already in my tree :( confused, greg k-h
On Fri, Aug 13, 2021 at 10:19:25AM +0200, Greg Kroah-Hartman wrote: > On Thu, Aug 12, 2021 at 12:14:50PM -0500, Bjorn Helgaas wrote: > > On Thu, Aug 12, 2021 at 11:17:12AM -0500, Bjorn Helgaas wrote: > > > [+to Greg, please update sysfs_defferred_iomem_get_mapping-5.15] > > > > Actually, Greg, totally up to you, but if nobody else is depending on > > the sysfs_defferred_iomem_get_mapping-5.15 branch, another possibility > > would be for you to drop that branch and for me to merge the two > > patches on it + Krzysztof's fix below + (hopefully) Krzysztof's PCI > > static attribute work. > > I can not "drop" the branch as it is already merged into my > driver-core-next branch that I can not rebase. I could revert it, but > is that really needed? Nope, I don't think a revert is warranted. I'll take care of getting Krzysztof's fix into v5.15. Bjorn
On Fri, Aug 13, 2021 at 09:29:01AM -0500, Bjorn Helgaas wrote: > On Fri, Aug 13, 2021 at 10:19:25AM +0200, Greg Kroah-Hartman wrote: > > On Thu, Aug 12, 2021 at 12:14:50PM -0500, Bjorn Helgaas wrote: > > > On Thu, Aug 12, 2021 at 11:17:12AM -0500, Bjorn Helgaas wrote: > > > > [+to Greg, please update sysfs_defferred_iomem_get_mapping-5.15] > > > > > > Actually, Greg, totally up to you, but if nobody else is depending on > > > the sysfs_defferred_iomem_get_mapping-5.15 branch, another possibility > > > would be for you to drop that branch and for me to merge the two > > > patches on it + Krzysztof's fix below + (hopefully) Krzysztof's PCI > > > static attribute work. > > > > I can not "drop" the branch as it is already merged into my > > driver-core-next branch that I can not rebase. I could revert it, but > > is that really needed? > > Nope, I don't think a revert is warranted. I'll take care of getting > Krzysztof's fix into v5.15. Thanks, I'll handle the merge issues after that happens in my tree. greg k-h
On Thu, Aug 12, 2021 at 01:21:44PM +0000, Krzysztof Wilczyński wrote: > Two legacy PCI sysfs objects "legacy_io" and "legacy_mem" were updated > to use an unified address space in the commit 636b21b50152 ("PCI: Revoke > mappings like devmem"). This allows for revocations to be managed from > a single place when drivers want to take over and mmap() a /dev/mem > range. > > Following the update, both of the sysfs objects should leverage the > iomem_get_mapping() function to get an appropriate address range, but > only the "legacy_io" has been correctly updated - the second attribute > seems to be using a wrong variable to pass the iomem_get_mapping() > function to. > > Thus, correct the variable name used so that the "legacy_mem" sysfs > object would also correctly call the iomem_get_mapping() function. > > Fixes: 636b21b50152 ("PCI: Revoke mappings like devmem") > Signed-off-by: Krzysztof Wilczyński <kw@linux.com> Applied to for-linus for v5.14, thanks! > --- > drivers/pci/pci-sysfs.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c > index 5d63df7c1820..7bbf2673c7f2 100644 > --- a/drivers/pci/pci-sysfs.c > +++ b/drivers/pci/pci-sysfs.c > @@ -978,7 +978,7 @@ void pci_create_legacy_files(struct pci_bus *b) > b->legacy_mem->size = 1024*1024; > b->legacy_mem->attr.mode = 0600; > b->legacy_mem->mmap = pci_mmap_legacy_mem; > - b->legacy_io->mapping = iomem_get_mapping(); > + b->legacy_mem->mapping = iomem_get_mapping(); > pci_adjust_legacy_attr(b, pci_mmap_mem); > error = device_create_bin_file(&b->dev, b->legacy_mem); > if (error) > -- > 2.32.0 >
diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c index 5d63df7c1820..7bbf2673c7f2 100644 --- a/drivers/pci/pci-sysfs.c +++ b/drivers/pci/pci-sysfs.c @@ -978,7 +978,7 @@ void pci_create_legacy_files(struct pci_bus *b) b->legacy_mem->size = 1024*1024; b->legacy_mem->attr.mode = 0600; b->legacy_mem->mmap = pci_mmap_legacy_mem; - b->legacy_io->mapping = iomem_get_mapping(); + b->legacy_mem->mapping = iomem_get_mapping(); pci_adjust_legacy_attr(b, pci_mmap_mem); error = device_create_bin_file(&b->dev, b->legacy_mem); if (error)
Two legacy PCI sysfs objects "legacy_io" and "legacy_mem" were updated to use an unified address space in the commit 636b21b50152 ("PCI: Revoke mappings like devmem"). This allows for revocations to be managed from a single place when drivers want to take over and mmap() a /dev/mem range. Following the update, both of the sysfs objects should leverage the iomem_get_mapping() function to get an appropriate address range, but only the "legacy_io" has been correctly updated - the second attribute seems to be using a wrong variable to pass the iomem_get_mapping() function to. Thus, correct the variable name used so that the "legacy_mem" sysfs object would also correctly call the iomem_get_mapping() function. Fixes: 636b21b50152 ("PCI: Revoke mappings like devmem") Signed-off-by: Krzysztof Wilczyński <kw@linux.com> --- drivers/pci/pci-sysfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)