Message ID | 20170313165849.1378558-1-arnd@arndb.de |
---|---|
State | Changes Requested, archived |
Delegated to: | David Miller |
Headers | show |
On Mon, 13 Mar 2017, Arnd Bergmann wrote: > With posix timers having become optional, we get a build error with > the cpts time sync option of the CPSW driver: > > drivers/net/ethernet/ti/cpts.c: In function 'cpts_find_ts': > drivers/net/ethernet/ti/cpts.c:291:23: error: implicit declaration of function 'ptp_classify_raw';did you mean 'ptp_classifier_init'? [-Werror=implicit-function-declaration] > > It really makes no sense to build this driver if we can't use PTP, > so it's better to go back to 'select PTP_1588_CLOCK' but instead > add a dependency on POSIX_TIMERS. Adding 'depends on PTP_1588_CLOCK' > might also work, but has the risk of circular dependencies when > mixed with other drivers using 'imply'. Could you elaborate on that risk please? > Fixes: baa73d9e478f ("posix-timers: Make them configurable") > Cc: Nicolas Pitre <nicolas.pitre@linaro.org> > Cc: stable@vger.kernel.org > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- > Originally submitted on Dec 16, but not applied after I missed a > reply from Nico. I confirmed that this is still needed for v4.10 > and v4.11-rc2 and am resending it without changes. > --- > drivers/net/ethernet/ti/Kconfig | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/net/ethernet/ti/Kconfig b/drivers/net/ethernet/ti/Kconfig > index d42257fbe9d9..c114efcd1575 100644 > --- a/drivers/net/ethernet/ti/Kconfig > +++ b/drivers/net/ethernet/ti/Kconfig > @@ -76,7 +76,8 @@ config TI_CPSW > config TI_CPTS_ENABLE > bool "TI Common Platform Time Sync (CPTS) Support" > depends on TI_CPSW || TI_KEYSTONE_NETCP > - imply PTP_1588_CLOCK > + depends on POSIX_TIMERS > + select PTP_1588_CLOCK > ---help--- > This driver supports the Common Platform Time Sync unit of > the CPSW Ethernet Switch and Keystone 2 1g/10g Switch Subsystem. > -- > 2.9.0 > >
On Mon, Mar 13, 2017 at 9:09 PM, Nicolas Pitre <nicolas.pitre@linaro.org> wrote: > On Mon, 13 Mar 2017, Arnd Bergmann wrote: > >> With posix timers having become optional, we get a build error with >> the cpts time sync option of the CPSW driver: >> >> drivers/net/ethernet/ti/cpts.c: In function 'cpts_find_ts': >> drivers/net/ethernet/ti/cpts.c:291:23: error: implicit declaration of function 'ptp_classify_raw';did you mean 'ptp_classifier_init'? [-Werror=implicit-function-declaration] >> >> It really makes no sense to build this driver if we can't use PTP, >> so it's better to go back to 'select PTP_1588_CLOCK' but instead >> add a dependency on POSIX_TIMERS. Adding 'depends on PTP_1588_CLOCK' >> might also work, but has the risk of circular dependencies when >> mixed with other drivers using 'imply'. > > Could you elaborate on that risk please? I have seen many circular dependencies in the past that tend to be of type config FOO depends on A select B config BAR select A depends on B The best way to avoid this problem is to only ever use either 'select' or 'depends on' for any given dependency, but not both. In this case, almost all references to PTP_1588_CLOCK use 'select' or 'implies', so I don't want to introduce any more 'depends on'. Arnd
On Mon, 13 Mar 2017, Arnd Bergmann wrote: > On Mon, Mar 13, 2017 at 9:09 PM, Nicolas Pitre <nicolas.pitre@linaro.org> wrote: > > On Mon, 13 Mar 2017, Arnd Bergmann wrote: > > > >> With posix timers having become optional, we get a build error with > >> the cpts time sync option of the CPSW driver: > >> > >> drivers/net/ethernet/ti/cpts.c: In function 'cpts_find_ts': > >> drivers/net/ethernet/ti/cpts.c:291:23: error: implicit declaration of function 'ptp_classify_raw';did you mean 'ptp_classifier_init'? [-Werror=implicit-function-declaration] > >> > >> It really makes no sense to build this driver if we can't use PTP, > >> so it's better to go back to 'select PTP_1588_CLOCK' but instead > >> add a dependency on POSIX_TIMERS. Adding 'depends on PTP_1588_CLOCK' > >> might also work, but has the risk of circular dependencies when > >> mixed with other drivers using 'imply'. > > > > Could you elaborate on that risk please? > > I have seen many circular dependencies in the past that tend to be of type > > config FOO > depends on A > select B > > config BAR > select A > depends on B This is really a problem? I mean in this example there is nothing that prevents A or B to be enabled independently. Of course if you had: config A depends on B config B depends on A then the circular dependency is obvious. > The best way to avoid this problem is to only ever use either 'select' or > 'depends on' for any given dependency, but not both. In this case, almost > all references to PTP_1588_CLOCK use 'select' or 'implies', so I don't > want to introduce any more 'depends on'. I can't find any "select PTP_1588_CLOCK" in the tree. The "imply" keyword in itself doesn't create nor inforce any dependencies -- that's why it was created in the first place. So unless I'm mistaken I don't see any problem using "depends on PTP_1588_CLOCK" here. Nicolas
On Mon, 13 Mar 2017, Nicolas Pitre wrote: > So unless I'm mistaken I don't see any problem using "depends on > PTP_1588_CLOCK" here. Furthermore that wouldn't be a first. See for example PTP_1588_CLOCK_GIANFAR, PTP_1588_CLOCK_IXP46X, DP83640_PHY, etc. Nicolas
On Mon, Mar 13, 2017 at 9:51 PM, Nicolas Pitre <nicolas.pitre@linaro.org> wrote: > On Mon, 13 Mar 2017, Nicolas Pitre wrote: > >> So unless I'm mistaken I don't see any problem using "depends on >> PTP_1588_CLOCK" here. > > Furthermore that wouldn't be a first. See for example > PTP_1588_CLOCK_GIANFAR, PTP_1588_CLOCK_IXP46X, DP83640_PHY, etc. Ok, I'll give this a few rounds on the randconfig builder and submit it that way unless I run into problems. Arnd
diff --git a/drivers/net/ethernet/ti/Kconfig b/drivers/net/ethernet/ti/Kconfig index d42257fbe9d9..c114efcd1575 100644 --- a/drivers/net/ethernet/ti/Kconfig +++ b/drivers/net/ethernet/ti/Kconfig @@ -76,7 +76,8 @@ config TI_CPSW config TI_CPTS_ENABLE bool "TI Common Platform Time Sync (CPTS) Support" depends on TI_CPSW || TI_KEYSTONE_NETCP - imply PTP_1588_CLOCK + depends on POSIX_TIMERS + select PTP_1588_CLOCK ---help--- This driver supports the Common Platform Time Sync unit of the CPSW Ethernet Switch and Keystone 2 1g/10g Switch Subsystem.
With posix timers having become optional, we get a build error with the cpts time sync option of the CPSW driver: drivers/net/ethernet/ti/cpts.c: In function 'cpts_find_ts': drivers/net/ethernet/ti/cpts.c:291:23: error: implicit declaration of function 'ptp_classify_raw';did you mean 'ptp_classifier_init'? [-Werror=implicit-function-declaration] It really makes no sense to build this driver if we can't use PTP, so it's better to go back to 'select PTP_1588_CLOCK' but instead add a dependency on POSIX_TIMERS. Adding 'depends on PTP_1588_CLOCK' might also work, but has the risk of circular dependencies when mixed with other drivers using 'imply'. Fixes: baa73d9e478f ("posix-timers: Make them configurable") Cc: Nicolas Pitre <nicolas.pitre@linaro.org> Cc: stable@vger.kernel.org Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- Originally submitted on Dec 16, but not applied after I missed a reply from Nico. I confirmed that this is still needed for v4.10 and v4.11-rc2 and am resending it without changes. --- drivers/net/ethernet/ti/Kconfig | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)