| Message ID | 20180528154448.2494709-1-arnd@arndb.de |
|---|---|
| State | Accepted |
| Commit | 73acc0315cd72174729141856c6607d1c91419cb |
| Headers | show |
| Series | ARM: mcpm, perf/arm-cci: export mcpm_is_available | expand |
Hi Arnd, Russell, [+Nico and Robin] On Mon, May 28, 2018 at 05:44:36PM +0200, Arnd Bergmann wrote: > Now that the ARM CCI PMU driver can be built as a loadable module, > we get a link failure when MCPM is enabled: > > ERROR: "mcpm_is_available" [drivers/perf/arm-cci.ko] undefined! > > The simplest fix is to export that helper function. > > Fixes: 8b0c93c20ef7 ("perf/arm-cci: Allow building as a module") > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- > The patch that caused this is currently part of the arm-perf/for-next/perf > branch, it would be good to have the fix there as well. > --- > arch/arm/common/mcpm_entry.c | 2 ++ > 1 file changed, 2 insertions(+) I'm happy to take this via the arm perf tree if others are ok with that. Alternatively, I can revert the offending commit if there are objections to exporting the symbol. Russell: do you any preference? Thanks, Will > diff --git a/arch/arm/common/mcpm_entry.c b/arch/arm/common/mcpm_entry.c > index ed9e87ddbb06..037a4479b8c3 100644 > --- a/arch/arm/common/mcpm_entry.c > +++ b/arch/arm/common/mcpm_entry.c > @@ -9,6 +9,7 @@ > * published by the Free Software Foundation. > */ > > +#include <linux/export.h> > #include <linux/kernel.h> > #include <linux/init.h> > #include <linux/irqflags.h> > @@ -174,6 +175,7 @@ bool mcpm_is_available(void) > { > return (platform_ops) ? true : false; > } > +EXPORT_SYMBOL_GPL(mcpm_is_available); > > /* > * We can't use regular spinlocks. In the switcher case, it is possible > -- > 2.9.0 >
On Tue, May 29, 2018 at 04:30:14PM +0100, Will Deacon wrote: > Hi Arnd, Russell, [+Nico and Robin] > > On Mon, May 28, 2018 at 05:44:36PM +0200, Arnd Bergmann wrote: > > Now that the ARM CCI PMU driver can be built as a loadable module, > > we get a link failure when MCPM is enabled: > > > > ERROR: "mcpm_is_available" [drivers/perf/arm-cci.ko] undefined! > > > > The simplest fix is to export that helper function. > > > > Fixes: 8b0c93c20ef7 ("perf/arm-cci: Allow building as a module") > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > > --- > > The patch that caused this is currently part of the arm-perf/for-next/perf > > branch, it would be good to have the fix there as well. > > --- > > arch/arm/common/mcpm_entry.c | 2 ++ > > 1 file changed, 2 insertions(+) > > I'm happy to take this via the arm perf tree if others are ok with that. > Alternatively, I can revert the offending commit if there are objections > to exporting the symbol. > > Russell: do you any preference? As it claims to fix 8b0c93c20ef7, which I don't have, I can't take this patch. Do we know which tree has this? -- RMK's Patch system: http://www.armlinux.org.uk/developer/patches/ FTTC broadband for 0.8mile line in suburbia: sync at 8.8Mbps down 630kbps up According to speedtest.net: 8.21Mbps down 510kbps up
Hi Russell, On Tue, May 29, 2018 at 04:33:24PM +0100, Russell King - ARM Linux wrote: > On Tue, May 29, 2018 at 04:30:14PM +0100, Will Deacon wrote: > > Hi Arnd, Russell, [+Nico and Robin] > > > > On Mon, May 28, 2018 at 05:44:36PM +0200, Arnd Bergmann wrote: > > > Now that the ARM CCI PMU driver can be built as a loadable module, > > > we get a link failure when MCPM is enabled: > > > > > > ERROR: "mcpm_is_available" [drivers/perf/arm-cci.ko] undefined! > > > > > > The simplest fix is to export that helper function. > > > > > > Fixes: 8b0c93c20ef7 ("perf/arm-cci: Allow building as a module") > > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > > > --- > > > The patch that caused this is currently part of the arm-perf/for-next/perf > > > branch, it would be good to have the fix there as well. > > > --- > > > arch/arm/common/mcpm_entry.c | 2 ++ > > > 1 file changed, 2 insertions(+) > > > > I'm happy to take this via the arm perf tree if others are ok with that. > > Alternatively, I can revert the offending commit if there are objections > > to exporting the symbol. > > > > Russell: do you any preference? > > As it claims to fix 8b0c93c20ef7, which I don't have, I can't take this > patch. Do we know which tree has this? Yes, sorry, it's in my for-next/perf branch: https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git/commit/?h=for-next/perf&id=8b0c93c20ef78f15d8b760964ff79bda7f68c610 which is now in -next. I'm happy to take the mcpm patch on top with your ack, but if you have conflicting changes (or would prefer not to export the symbol to modules) then I can just revert the patch in my tree for now. Will
On Tue, May 29, 2018 at 04:41:20PM +0100, Will Deacon wrote: > Hi Russell, > > On Tue, May 29, 2018 at 04:33:24PM +0100, Russell King - ARM Linux wrote: > > On Tue, May 29, 2018 at 04:30:14PM +0100, Will Deacon wrote: > > > Hi Arnd, Russell, [+Nico and Robin] > > > > > > On Mon, May 28, 2018 at 05:44:36PM +0200, Arnd Bergmann wrote: > > > > Now that the ARM CCI PMU driver can be built as a loadable module, > > > > we get a link failure when MCPM is enabled: > > > > > > > > ERROR: "mcpm_is_available" [drivers/perf/arm-cci.ko] undefined! > > > > > > > > The simplest fix is to export that helper function. > > > > > > > > Fixes: 8b0c93c20ef7 ("perf/arm-cci: Allow building as a module") > > > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > > > > --- > > > > The patch that caused this is currently part of the arm-perf/for-next/perf > > > > branch, it would be good to have the fix there as well. > > > > --- > > > > arch/arm/common/mcpm_entry.c | 2 ++ > > > > 1 file changed, 2 insertions(+) > > > > > > I'm happy to take this via the arm perf tree if others are ok with that. > > > Alternatively, I can revert the offending commit if there are objections > > > to exporting the symbol. > > > > > > Russell: do you any preference? > > > > As it claims to fix 8b0c93c20ef7, which I don't have, I can't take this > > patch. Do we know which tree has this? > > Yes, sorry, it's in my for-next/perf branch: > > https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git/commit/?h=for-next/perf&id=8b0c93c20ef78f15d8b760964ff79bda7f68c610 > > which is now in -next. > > I'm happy to take the mcpm patch on top with your ack, but if you have > conflicting changes (or would prefer not to export the symbol to modules) > then I can just revert the patch in my tree for now. I have no changes to that file, so: Acked-by: Russell King <rmk+kernel@armlinux.org.uk> Thanks. -- RMK's Patch system: http://www.armlinux.org.uk/developer/patches/ FTTC broadband for 0.8mile line in suburbia: sync at 8.8Mbps down 630kbps up According to speedtest.net: 8.21Mbps down 510kbps up
On Tue, May 29, 2018 at 04:42:51PM +0100, Russell King - ARM Linux wrote: > On Tue, May 29, 2018 at 04:41:20PM +0100, Will Deacon wrote: > > Hi Russell, > > > > On Tue, May 29, 2018 at 04:33:24PM +0100, Russell King - ARM Linux wrote: > > > On Tue, May 29, 2018 at 04:30:14PM +0100, Will Deacon wrote: > > > > Hi Arnd, Russell, [+Nico and Robin] > > > > > > > > On Mon, May 28, 2018 at 05:44:36PM +0200, Arnd Bergmann wrote: > > > > > Now that the ARM CCI PMU driver can be built as a loadable module, > > > > > we get a link failure when MCPM is enabled: > > > > > > > > > > ERROR: "mcpm_is_available" [drivers/perf/arm-cci.ko] undefined! > > > > > > > > > > The simplest fix is to export that helper function. > > > > > > > > > > Fixes: 8b0c93c20ef7 ("perf/arm-cci: Allow building as a module") > > > > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > > > > > --- > > > > > The patch that caused this is currently part of the arm-perf/for-next/perf > > > > > branch, it would be good to have the fix there as well. > > > > > --- > > > > > arch/arm/common/mcpm_entry.c | 2 ++ > > > > > 1 file changed, 2 insertions(+) > > > > > > > > I'm happy to take this via the arm perf tree if others are ok with that. > > > > Alternatively, I can revert the offending commit if there are objections > > > > to exporting the symbol. > > > > > > > > Russell: do you any preference? > > > > > > As it claims to fix 8b0c93c20ef7, which I don't have, I can't take this > > > patch. Do we know which tree has this? > > > > Yes, sorry, it's in my for-next/perf branch: > > > > https://git.kernel.org/pub/scm/linux/kernel/git/will/linux.git/commit/?h=for-next/perf&id=8b0c93c20ef78f15d8b760964ff79bda7f68c610 > > > > which is now in -next. > > > > I'm happy to take the mcpm patch on top with your ack, but if you have > > conflicting changes (or would prefer not to export the symbol to modules) > > then I can just revert the patch in my tree for now. > > I have no changes to that file, so: > > Acked-by: Russell King <rmk+kernel@armlinux.org.uk> Thanks, I've picked up the fix from Arnd with your Ack and will push out this evening. Will
On Tue, 29 May 2018, Will Deacon wrote: > Hi Arnd, Russell, [+Nico and Robin] > > On Mon, May 28, 2018 at 05:44:36PM +0200, Arnd Bergmann wrote: > > Now that the ARM CCI PMU driver can be built as a loadable module, > > we get a link failure when MCPM is enabled: > > > > ERROR: "mcpm_is_available" [drivers/perf/arm-cci.ko] undefined! > > > > The simplest fix is to export that helper function. > > > > Fixes: 8b0c93c20ef7 ("perf/arm-cci: Allow building as a module") > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > > --- > > The patch that caused this is currently part of the arm-perf/for-next/perf > > branch, it would be good to have the fix there as well. > > --- > > arch/arm/common/mcpm_entry.c | 2 ++ > > 1 file changed, 2 insertions(+) > > I'm happy to take this via the arm perf tree if others are ok with that. > Alternatively, I can revert the offending commit if there are objections > to exporting the symbol. Looks fine to me. Acked-by: Nicolas Pitre <nico@linaro.org> Nicolas
diff --git a/arch/arm/common/mcpm_entry.c b/arch/arm/common/mcpm_entry.c index ed9e87ddbb06..037a4479b8c3 100644 --- a/arch/arm/common/mcpm_entry.c +++ b/arch/arm/common/mcpm_entry.c @@ -9,6 +9,7 @@ * published by the Free Software Foundation. */ +#include <linux/export.h> #include <linux/kernel.h> #include <linux/init.h> #include <linux/irqflags.h> @@ -174,6 +175,7 @@ bool mcpm_is_available(void) { return (platform_ops) ? true : false; } +EXPORT_SYMBOL_GPL(mcpm_is_available); /* * We can't use regular spinlocks. In the switcher case, it is possible
Now that the ARM CCI PMU driver can be built as a loadable module, we get a link failure when MCPM is enabled: ERROR: "mcpm_is_available" [drivers/perf/arm-cci.ko] undefined! The simplest fix is to export that helper function. Fixes: 8b0c93c20ef7 ("perf/arm-cci: Allow building as a module") Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- The patch that caused this is currently part of the arm-perf/for-next/perf branch, it would be good to have the fix there as well. --- arch/arm/common/mcpm_entry.c | 2 ++ 1 file changed, 2 insertions(+) -- 2.9.0