| Message ID | 20240815125905.1667148-1-vadfed@meta.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net,v3,1/2] ptp: ocp: adjust sysfs entries to expose tty information | expand |
On Thu, Aug 15, 2024 at 05:59:04AM -0700, Vadim Fedorenko wrote: > Starting v6.8 the serial port subsystem changed the hierarchy of devices > and symlinks are not working anymore. Previous discussion made it clear > that the idea of symlinks for tty devices was wrong by design. Implement > additional attributes to expose the information. Fixes tag points to the > commit which introduced the change. > > Fixes: b286f4e87e32 ("serial: core: Move tty and serdev to be children of serial core port device") > Signed-off-by: Vadim Fedorenko <vadfed@meta.com> Complicated fixes, nice! Thanks for doing this. One question: > +static ssize_t > +ptp_ocp_tty_show(struct device *dev, struct device_attribute *attr, char *buf) > +{ > + struct dev_ext_attribute *ea = to_ext_attr(attr); > + struct ptp_ocp *bp = dev_get_drvdata(dev); > + struct ptp_ocp_serial_port *port; > + > + return sysfs_emit(buf, "ttyS%d", bp->port[(uintptr_t)ea->var].line); "uintptr_t"? That's not a normal kernel type. var is of type "void *" so can't this just be "int" here? Or am I reading this wrong? thanks, greg k-h
On Thu, Aug 15, 2024 at 05:59:04AM -0700, Vadim Fedorenko wrote: > Starting v6.8 the serial port subsystem changed the hierarchy of devices > and symlinks are not working anymore. Previous discussion made it clear > that the idea of symlinks for tty devices was wrong by design. Implement > additional attributes to expose the information. Fixes tag points to the > commit which introduced the change. Hi Vadim, Would it be possible to provide a link to the discussion(s)? > > Fixes: b286f4e87e32 ("serial: core: Move tty and serdev to be children of serial core port device") > Signed-off-by: Vadim Fedorenko <vadfed@meta.com> ... > diff --git a/drivers/ptp/ptp_ocp.c b/drivers/ptp/ptp_ocp.c ... > @@ -3346,6 +3352,54 @@ static EXT_ATTR_RO(freq, frequency, 1); > static EXT_ATTR_RO(freq, frequency, 2); > static EXT_ATTR_RO(freq, frequency, 3); > > +static ssize_t > +ptp_ocp_tty_show(struct device *dev, struct device_attribute *attr, char *buf) > +{ > + struct dev_ext_attribute *ea = to_ext_attr(attr); > + struct ptp_ocp *bp = dev_get_drvdata(dev); > + struct ptp_ocp_serial_port *port; nit: Port is unused in this function, it should be removed. > + > + return sysfs_emit(buf, "ttyS%d", bp->port[(uintptr_t)ea->var].line); > +} ...
On 15/08/2024 14:54, Simon Horman wrote: > On Thu, Aug 15, 2024 at 05:59:04AM -0700, Vadim Fedorenko wrote: >> Starting v6.8 the serial port subsystem changed the hierarchy of devices >> and symlinks are not working anymore. Previous discussion made it clear >> that the idea of symlinks for tty devices was wrong by design. Implement >> additional attributes to expose the information. Fixes tag points to the >> commit which introduced the change. > > Hi Vadim, > > Would it be possible to provide a link to the discussion(s)? > Hi Simon, Yeah, sure: the initial RFC of serial port subsystem changes is https://lore.kernel.org/linux-serial/20231024113624.54364-1-tony@atomide.com/ the merged version is (serial port subsystem): https://lore.kernel.org/linux-serial/20231113080758.30346-1-tony@atomide.com/ The first update to ptp_ocp driver is https://lore.kernel.org/linux-serial/20240510110405.15115-1-vadim.fedorenko@linux.dev/ >> >> Fixes: b286f4e87e32 ("serial: core: Move tty and serdev to be children of serial core port device") >> Signed-off-by: Vadim Fedorenko <vadfed@meta.com> > > ... > >> diff --git a/drivers/ptp/ptp_ocp.c b/drivers/ptp/ptp_ocp.c > > ... > >> @@ -3346,6 +3352,54 @@ static EXT_ATTR_RO(freq, frequency, 1); >> static EXT_ATTR_RO(freq, frequency, 2); >> static EXT_ATTR_RO(freq, frequency, 3); >> >> +static ssize_t >> +ptp_ocp_tty_show(struct device *dev, struct device_attribute *attr, char *buf) >> +{ >> + struct dev_ext_attribute *ea = to_ext_attr(attr); >> + struct ptp_ocp *bp = dev_get_drvdata(dev); >> + struct ptp_ocp_serial_port *port; > > nit: Port is unused in this function, it should be removed. Ah, yeah, it was last second change, didn't clean it fully, thanks for pointing. > >> + >> + return sysfs_emit(buf, "ttyS%d", bp->port[(uintptr_t)ea->var].line); >> +} > > ...
On 15/08/2024 14:41, Greg Kroah-Hartman wrote: > On Thu, Aug 15, 2024 at 05:59:04AM -0700, Vadim Fedorenko wrote: >> Starting v6.8 the serial port subsystem changed the hierarchy of devices >> and symlinks are not working anymore. Previous discussion made it clear >> that the idea of symlinks for tty devices was wrong by design. Implement >> additional attributes to expose the information. Fixes tag points to the >> commit which introduced the change. >> >> Fixes: b286f4e87e32 ("serial: core: Move tty and serdev to be children of serial core port device") >> Signed-off-by: Vadim Fedorenko <vadfed@meta.com> > > Complicated fixes, nice! Thanks for doing this. One question: > >> +static ssize_t >> +ptp_ocp_tty_show(struct device *dev, struct device_attribute *attr, char *buf) >> +{ >> + struct dev_ext_attribute *ea = to_ext_attr(attr); >> + struct ptp_ocp *bp = dev_get_drvdata(dev); >> + struct ptp_ocp_serial_port *port; >> + >> + return sysfs_emit(buf, "ttyS%d", bp->port[(uintptr_t)ea->var].line); > > "uintptr_t"? That's not a normal kernel type. var is of type "void *" > so can't this just be "int" here? Or am I reading this wrong? Well, yes, looks like it can be used as pure int provided enum is int and we simply use "void *" as a transport. I'll send v4 with this fix and the one Simon pointed in 24hr (as per netdev policy). Thanks! > thanks, > > greg k-h
On 15. 08. 24, 14:59, Vadim Fedorenko wrote: > Starting v6.8 the serial port subsystem changed the hierarchy of devices > and symlinks are not working anymore. Previous discussion made it clear > that the idea of symlinks for tty devices was wrong by design. Care to link it here? > Implement > additional attributes to expose the information. Fixes tag points to the > commit which introduced the change. ... > --- a/drivers/ptp/ptp_ocp.c > +++ b/drivers/ptp/ptp_ocp.c > @@ -316,6 +316,15 @@ struct ptp_ocp_serial_port { > #define OCP_SERIAL_LEN 6 > #define OCP_SMA_NUM 4 > > +enum { > + PORT_GNSS, > + PORT_GNSS2, > + PORT_MAC, /* miniature atomic clock */ > + PORT_NMEA, > + > + PORT_NUM_MAX > +}; > + The conversion to the array needs to go to a separate patch, apparently. > +static ssize_t > +ptp_ocp_tty_show(struct device *dev, struct device_attribute *attr, char *buf) > +{ > + struct dev_ext_attribute *ea = to_ext_attr(attr); > + struct ptp_ocp *bp = dev_get_drvdata(dev); > + struct ptp_ocp_serial_port *port; > + > + return sysfs_emit(buf, "ttyS%d", bp->port[(uintptr_t)ea->var].line); uintptr_t is unusual as Greg points out. It is a correct type per C99 to cast from/to pointers, but we usually do "unsigned long". (int wouldn't work as it has a different size (on 64bit).) But looking at the code, uintptr_t is used all over. So perhaps use that to be consistent? > @@ -3960,16 +4017,16 @@ ptp_ocp_summary_show(struct seq_file *s, void *data) > bp = dev_get_drvdata(dev); > > seq_printf(s, "%7s: /dev/ptp%d\n", "PTP", ptp_clock_index(bp->ptp)); > - if (bp->gnss_port.line != -1) > + if (bp->port[PORT_GNSS].line != -1) > seq_printf(s, "%7s: /dev/ttyS%d\n", "GNSS1", > - bp->gnss_port.line); > - if (bp->gnss2_port.line != -1) > + bp->port[PORT_GNSS].line); > + if (bp->port[PORT_GNSS2].line != -1) > seq_printf(s, "%7s: /dev/ttyS%d\n", "GNSS2", > - bp->gnss2_port.line); > - if (bp->mac_port.line != -1) > - seq_printf(s, "%7s: /dev/ttyS%d\n", "MAC", bp->mac_port.line); > - if (bp->nmea_port.line != -1) > - seq_printf(s, "%7s: /dev/ttyS%d\n", "NMEA", bp->nmea_port.line); > + bp->port[PORT_GNSS2].line); > + if (bp->port[PORT_MAC].line != -1) > + seq_printf(s, "%7s: /dev/ttyS%d\n", "MAC", bp->port[PORT_MAC].line); > + if (bp->port[PORT_NMEA].line != -1) > + seq_printf(s, "%7s: /dev/ttyS%d\n", "NMEA", bp->port[PORT_NMEA].line); Perhaps you can introduce some to_name() function (mapping enum -> const char *)? Can this code be then a three-line for loop? > @@ -4419,20 +4460,21 @@ ptp_ocp_info(struct ptp_ocp *bp) > > ptp_ocp_phc_info(bp); > > - ptp_ocp_serial_info(dev, "GNSS", bp->gnss_port.line, > - bp->gnss_port.baud); > - ptp_ocp_serial_info(dev, "GNSS2", bp->gnss2_port.line, > - bp->gnss2_port.baud); > - ptp_ocp_serial_info(dev, "MAC", bp->mac_port.line, bp->mac_port.baud); > - if (bp->nmea_out && bp->nmea_port.line != -1) { > - bp->nmea_port.baud = -1; > + ptp_ocp_serial_info(dev, "GNSS", bp->port[PORT_GNSS].line, > + bp->port[PORT_GNSS].baud); > + ptp_ocp_serial_info(dev, "GNSS2", bp->port[PORT_GNSS2].line, > + bp->port[PORT_GNSS2].baud); > + ptp_ocp_serial_info(dev, "MAC", bp->port[PORT_MAC].line, > + bp->port[PORT_MAC].baud); > + if (bp->nmea_out && bp->port[PORT_NMEA].line != -1) { > + bp->port[PORT_NMEA].baud = -1; > > reg = ioread32(&bp->nmea_out->uart_baud); > if (reg < ARRAY_SIZE(nmea_baud)) > - bp->nmea_port.baud = nmea_baud[reg]; > + bp->port[PORT_NMEA].baud = nmea_baud[reg]; > > - ptp_ocp_serial_info(dev, "NMEA", bp->nmea_port.line, > - bp->nmea_port.baud); > + ptp_ocp_serial_info(dev, "NMEA", bp->port[PORT_NMEA].line, > + bp->port[PORT_NMEA].baud); Maybe even here with if (iterator == PORT_NMEA)?
Hi Vadim, kernel test robot noticed the following build warnings: [auto build test WARNING on net/main] url: https://github.com/intel-lab-lkp/linux/commits/Vadim-Fedorenko/docs-ABI-update-OCP-TimeCard-sysfs-entries/20240815-210131 base: net/main patch link: https://lore.kernel.org/r/20240815125905.1667148-1-vadfed%40meta.com patch subject: [PATCH net v3 1/2] ptp: ocp: adjust sysfs entries to expose tty information config: loongarch-allmodconfig (https://download.01.org/0day-ci/archive/20240816/202408161934.aQsUDy88-lkp@intel.com/config) compiler: loongarch64-linux-gcc (GCC) 14.1.0 reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240816/202408161934.aQsUDy88-lkp@intel.com/reproduce) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <lkp@intel.com> | Closes: https://lore.kernel.org/oe-kbuild-all/202408161934.aQsUDy88-lkp@intel.com/ All warnings (new ones prefixed by >>): drivers/ptp/ptp_ocp.c: In function 'ptp_ocp_tty_show': >> drivers/ptp/ptp_ocp.c:3360:37: warning: unused variable 'port' [-Wunused-variable] 3360 | struct ptp_ocp_serial_port *port; | ^~~~ vim +/port +3360 drivers/ptp/ptp_ocp.c 3354 3355 static ssize_t 3356 ptp_ocp_tty_show(struct device *dev, struct device_attribute *attr, char *buf) 3357 { 3358 struct dev_ext_attribute *ea = to_ext_attr(attr); 3359 struct ptp_ocp *bp = dev_get_drvdata(dev); > 3360 struct ptp_ocp_serial_port *port; 3361 3362 return sysfs_emit(buf, "ttyS%d", bp->port[(uintptr_t)ea->var].line); 3363 } 3364
On 16/08/2024 06:26, Jiri Slaby wrote: > On 15. 08. 24, 14:59, Vadim Fedorenko wrote: >> Starting v6.8 the serial port subsystem changed the hierarchy of devices >> and symlinks are not working anymore. Previous discussion made it clear >> that the idea of symlinks for tty devices was wrong by design. > > Care to link it here? Sure, with the next version. > >> Implement >> additional attributes to expose the information. Fixes tag points to the >> commit which introduced the change. > ... >> --- a/drivers/ptp/ptp_ocp.c >> +++ b/drivers/ptp/ptp_ocp.c >> @@ -316,6 +316,15 @@ struct ptp_ocp_serial_port { >> #define OCP_SERIAL_LEN 6 >> #define OCP_SMA_NUM 4 >> +enum { >> + PORT_GNSS, >> + PORT_GNSS2, >> + PORT_MAC, /* miniature atomic clock */ >> + PORT_NMEA, >> + >> + PORT_NUM_MAX >> +}; >> + > > The conversion to the array needs to go to a separate patch, apparently. I'm not sure here. I'm trying to fix the regression introduced back in 6.8. The conversion to the array itself doesn't solve the issue, it's pure net-next material. But the simple fix was NACKed previously, that's why I had to introduce such a big change. I would like to keep these changes all together in one patch. > >> +static ssize_t >> +ptp_ocp_tty_show(struct device *dev, struct device_attribute *attr, >> char *buf) >> +{ >> + struct dev_ext_attribute *ea = to_ext_attr(attr); >> + struct ptp_ocp *bp = dev_get_drvdata(dev); >> + struct ptp_ocp_serial_port *port; >> + >> + return sysfs_emit(buf, "ttyS%d", bp->port[(uintptr_t)ea->var].line); > > uintptr_t is unusual as Greg points out. It is a correct type per C99 to > cast from/to pointers, but we usually do "unsigned long". (int wouldn't > work as it has a different size (on 64bit).) > > But looking at the code, uintptr_t is used all over. So perhaps use that > to be consistent? so, now I'm lost. I'm ok with both options, just let me know which one will have no objections to be merged... > >> @@ -3960,16 +4017,16 @@ ptp_ocp_summary_show(struct seq_file *s, void >> *data) >> bp = dev_get_drvdata(dev); >> seq_printf(s, "%7s: /dev/ptp%d\n", "PTP", ptp_clock_index(bp- >> >ptp)); >> - if (bp->gnss_port.line != -1) >> + if (bp->port[PORT_GNSS].line != -1) >> seq_printf(s, "%7s: /dev/ttyS%d\n", "GNSS1", >> - bp->gnss_port.line); >> - if (bp->gnss2_port.line != -1) >> + bp->port[PORT_GNSS].line); >> + if (bp->port[PORT_GNSS2].line != -1) >> seq_printf(s, "%7s: /dev/ttyS%d\n", "GNSS2", >> - bp->gnss2_port.line); >> - if (bp->mac_port.line != -1) >> - seq_printf(s, "%7s: /dev/ttyS%d\n", "MAC", bp->mac_port.line); >> - if (bp->nmea_port.line != -1) >> - seq_printf(s, "%7s: /dev/ttyS%d\n", "NMEA", bp->nmea_port.line); >> + bp->port[PORT_GNSS2].line); >> + if (bp->port[PORT_MAC].line != -1) >> + seq_printf(s, "%7s: /dev/ttyS%d\n", "MAC", bp- >> >port[PORT_MAC].line); >> + if (bp->port[PORT_NMEA].line != -1) >> + seq_printf(s, "%7s: /dev/ttyS%d\n", "NMEA", bp- >> >port[PORT_NMEA].line); > > Perhaps you can introduce some to_name() function (mapping enum -> const > char *)? Can this code be then a three-line for loop? yeah, I can do it, for sure. > >> @@ -4419,20 +4460,21 @@ ptp_ocp_info(struct ptp_ocp *bp) >> ptp_ocp_phc_info(bp); >> - ptp_ocp_serial_info(dev, "GNSS", bp->gnss_port.line, >> - bp->gnss_port.baud); >> - ptp_ocp_serial_info(dev, "GNSS2", bp->gnss2_port.line, >> - bp->gnss2_port.baud); >> - ptp_ocp_serial_info(dev, "MAC", bp->mac_port.line, bp- >> >mac_port.baud); >> - if (bp->nmea_out && bp->nmea_port.line != -1) { >> - bp->nmea_port.baud = -1; >> + ptp_ocp_serial_info(dev, "GNSS", bp->port[PORT_GNSS].line, >> + bp->port[PORT_GNSS].baud); >> + ptp_ocp_serial_info(dev, "GNSS2", bp->port[PORT_GNSS2].line, >> + bp->port[PORT_GNSS2].baud); >> + ptp_ocp_serial_info(dev, "MAC", bp->port[PORT_MAC].line, >> + bp->port[PORT_MAC].baud); >> + if (bp->nmea_out && bp->port[PORT_NMEA].line != -1) { >> + bp->port[PORT_NMEA].baud = -1; >> reg = ioread32(&bp->nmea_out->uart_baud); >> if (reg < ARRAY_SIZE(nmea_baud)) >> - bp->nmea_port.baud = nmea_baud[reg]; >> + bp->port[PORT_NMEA].baud = nmea_baud[reg]; >> - ptp_ocp_serial_info(dev, "NMEA", bp->nmea_port.line, >> - bp->nmea_port.baud); >> + ptp_ocp_serial_info(dev, "NMEA", bp->port[PORT_NMEA].line, >> + bp->port[PORT_NMEA].baud); > > Maybe even here with if (iterator == PORT_NMEA)? > Should be ok, will make it in the next revision. Thanks, Vadim
diff --git a/drivers/ptp/ptp_ocp.c b/drivers/ptp/ptp_ocp.c index ee2ced88ab34..7d50b49bf684 100644 --- a/drivers/ptp/ptp_ocp.c +++ b/drivers/ptp/ptp_ocp.c @@ -316,6 +316,15 @@ struct ptp_ocp_serial_port { #define OCP_SERIAL_LEN 6 #define OCP_SMA_NUM 4 +enum { + PORT_GNSS, + PORT_GNSS2, + PORT_MAC, /* miniature atomic clock */ + PORT_NMEA, + + PORT_NUM_MAX +}; + struct ptp_ocp { struct pci_dev *pdev; struct device dev; @@ -357,10 +366,7 @@ struct ptp_ocp { struct delayed_work sync_work; int id; int n_irqs; - struct ptp_ocp_serial_port gnss_port; - struct ptp_ocp_serial_port gnss2_port; - struct ptp_ocp_serial_port mac_port; /* miniature atomic clock */ - struct ptp_ocp_serial_port nmea_port; + struct ptp_ocp_serial_port port[PORT_NUM_MAX]; bool fw_loader; u8 fw_tag; u16 fw_version; @@ -655,28 +661,28 @@ static struct ocp_resource ocp_fb_resource[] = { }, }, { - OCP_SERIAL_RESOURCE(gnss_port), + OCP_SERIAL_RESOURCE(port[PORT_GNSS]), .offset = 0x00160000 + 0x1000, .irq_vec = 3, .extra = &(struct ptp_ocp_serial_port) { .baud = 115200, }, }, { - OCP_SERIAL_RESOURCE(gnss2_port), + OCP_SERIAL_RESOURCE(port[PORT_GNSS2]), .offset = 0x00170000 + 0x1000, .irq_vec = 4, .extra = &(struct ptp_ocp_serial_port) { .baud = 115200, }, }, { - OCP_SERIAL_RESOURCE(mac_port), + OCP_SERIAL_RESOURCE(port[PORT_MAC]), .offset = 0x00180000 + 0x1000, .irq_vec = 5, .extra = &(struct ptp_ocp_serial_port) { .baud = 57600, }, }, { - OCP_SERIAL_RESOURCE(nmea_port), + OCP_SERIAL_RESOURCE(port[PORT_NMEA]), .offset = 0x00190000 + 0x1000, .irq_vec = 10, }, { @@ -740,7 +746,7 @@ static struct ocp_resource ocp_art_resource[] = { .offset = 0x01000000, .size = 0x10000, }, { - OCP_SERIAL_RESOURCE(gnss_port), + OCP_SERIAL_RESOURCE(port[PORT_GNSS]), .offset = 0x00160000 + 0x1000, .irq_vec = 3, .extra = &(struct ptp_ocp_serial_port) { .baud = 115200, @@ -839,7 +845,7 @@ static struct ocp_resource ocp_art_resource[] = { }, }, { - OCP_SERIAL_RESOURCE(mac_port), + OCP_SERIAL_RESOURCE(port[PORT_MAC]), .offset = 0x00190000, .irq_vec = 7, .extra = &(struct ptp_ocp_serial_port) { .baud = 9600, @@ -950,14 +956,14 @@ static struct ocp_resource ocp_adva_resource[] = { .offset = 0x00220000, .size = 0x1000, }, { - OCP_SERIAL_RESOURCE(gnss_port), + OCP_SERIAL_RESOURCE(port[PORT_GNSS]), .offset = 0x00160000 + 0x1000, .irq_vec = 3, .extra = &(struct ptp_ocp_serial_port) { .baud = 9600, }, }, { - OCP_SERIAL_RESOURCE(mac_port), + OCP_SERIAL_RESOURCE(port[PORT_MAC]), .offset = 0x00180000 + 0x1000, .irq_vec = 5, .extra = &(struct ptp_ocp_serial_port) { .baud = 115200, @@ -3346,6 +3352,54 @@ static EXT_ATTR_RO(freq, frequency, 1); static EXT_ATTR_RO(freq, frequency, 2); static EXT_ATTR_RO(freq, frequency, 3); +static ssize_t +ptp_ocp_tty_show(struct device *dev, struct device_attribute *attr, char *buf) +{ + struct dev_ext_attribute *ea = to_ext_attr(attr); + struct ptp_ocp *bp = dev_get_drvdata(dev); + struct ptp_ocp_serial_port *port; + + return sysfs_emit(buf, "ttyS%d", bp->port[(uintptr_t)ea->var].line); +} + +static umode_t +ptp_ocp_timecard_tty_is_visible(struct kobject *kobj, struct attribute *attr, int n) +{ + struct ptp_ocp *bp = dev_get_drvdata(kobj_to_dev(kobj)); + struct ptp_ocp_serial_port *port; + struct device_attribute *dattr; + struct dev_ext_attribute *ea; + + if (strncmp(attr->name, "tty", 3)) + return attr->mode; + + dattr = container_of(attr, struct device_attribute, attr); + ea = container_of(dattr, struct dev_ext_attribute, attr); + port = &bp->port[(uintptr_t)ea->var]; + return port->line == -1 ? 0 : 0444; +} +#define EXT_TTY_ATTR_RO(_name, _val) \ + struct dev_ext_attribute dev_attr_tty##_name = \ + { __ATTR(tty##_name, 0444, ptp_ocp_tty_show, NULL), (void *)_val } + +static EXT_TTY_ATTR_RO(GNSS, PORT_GNSS); +static EXT_TTY_ATTR_RO(GNSS2, PORT_GNSS2); +static EXT_TTY_ATTR_RO(MAC, PORT_MAC); +static EXT_TTY_ATTR_RO(NMEA, PORT_NMEA); +static struct attribute *ptp_ocp_timecard_tty_attrs[] = { + &dev_attr_ttyGNSS.attr.attr, + &dev_attr_ttyGNSS2.attr.attr, + &dev_attr_ttyMAC.attr.attr, + &dev_attr_ttyNMEA.attr.attr, + NULL, +}; + +static const struct attribute_group ptp_ocp_timecard_tty_group = { + .name = "tty", + .attrs = ptp_ocp_timecard_tty_attrs, + .is_visible = ptp_ocp_timecard_tty_is_visible, +}; + static ssize_t serialnum_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -3775,6 +3829,7 @@ static const struct attribute_group fb_timecard_group = { static const struct ocp_attr_group fb_timecard_groups[] = { { .cap = OCP_CAP_BASIC, .group = &fb_timecard_group }, + { .cap = OCP_CAP_BASIC, .group = &ptp_ocp_timecard_tty_group }, { .cap = OCP_CAP_SIGNAL, .group = &fb_timecard_signal0_group }, { .cap = OCP_CAP_SIGNAL, .group = &fb_timecard_signal1_group }, { .cap = OCP_CAP_SIGNAL, .group = &fb_timecard_signal2_group }, @@ -3814,6 +3869,7 @@ static const struct attribute_group art_timecard_group = { static const struct ocp_attr_group art_timecard_groups[] = { { .cap = OCP_CAP_BASIC, .group = &art_timecard_group }, + { .cap = OCP_CAP_BASIC, .group = &ptp_ocp_timecard_tty_group }, { }, }; @@ -3841,6 +3897,7 @@ static const struct attribute_group adva_timecard_group = { static const struct ocp_attr_group adva_timecard_groups[] = { { .cap = OCP_CAP_BASIC, .group = &adva_timecard_group }, + { .cap = OCP_CAP_BASIC, .group = &ptp_ocp_timecard_tty_group }, { .cap = OCP_CAP_SIGNAL, .group = &fb_timecard_signal0_group }, { .cap = OCP_CAP_SIGNAL, .group = &fb_timecard_signal1_group }, { .cap = OCP_CAP_FREQ, .group = &fb_timecard_freq0_group }, @@ -3960,16 +4017,16 @@ ptp_ocp_summary_show(struct seq_file *s, void *data) bp = dev_get_drvdata(dev); seq_printf(s, "%7s: /dev/ptp%d\n", "PTP", ptp_clock_index(bp->ptp)); - if (bp->gnss_port.line != -1) + if (bp->port[PORT_GNSS].line != -1) seq_printf(s, "%7s: /dev/ttyS%d\n", "GNSS1", - bp->gnss_port.line); - if (bp->gnss2_port.line != -1) + bp->port[PORT_GNSS].line); + if (bp->port[PORT_GNSS2].line != -1) seq_printf(s, "%7s: /dev/ttyS%d\n", "GNSS2", - bp->gnss2_port.line); - if (bp->mac_port.line != -1) - seq_printf(s, "%7s: /dev/ttyS%d\n", "MAC", bp->mac_port.line); - if (bp->nmea_port.line != -1) - seq_printf(s, "%7s: /dev/ttyS%d\n", "NMEA", bp->nmea_port.line); + bp->port[PORT_GNSS2].line); + if (bp->port[PORT_MAC].line != -1) + seq_printf(s, "%7s: /dev/ttyS%d\n", "MAC", bp->port[PORT_MAC].line); + if (bp->port[PORT_NMEA].line != -1) + seq_printf(s, "%7s: /dev/ttyS%d\n", "NMEA", bp->port[PORT_NMEA].line); memset(sma_val, 0xff, sizeof(sma_val)); if (bp->sma_map1) { @@ -4279,7 +4336,7 @@ ptp_ocp_dev_release(struct device *dev) static int ptp_ocp_device_init(struct ptp_ocp *bp, struct pci_dev *pdev) { - int err; + int i, err; mutex_lock(&ptp_ocp_lock); err = idr_alloc(&ptp_ocp_idr, bp, 0, 0, GFP_KERNEL); @@ -4292,10 +4349,10 @@ ptp_ocp_device_init(struct ptp_ocp *bp, struct pci_dev *pdev) bp->ptp_info = ptp_ocp_clock_info; spin_lock_init(&bp->lock); - bp->gnss_port.line = -1; - bp->gnss2_port.line = -1; - bp->mac_port.line = -1; - bp->nmea_port.line = -1; + + for (i = 0; i < PORT_NUM_MAX; i++) + bp->port[i].line = -1; + bp->pdev = pdev; device_initialize(&bp->dev); @@ -4352,22 +4409,6 @@ ptp_ocp_complete(struct ptp_ocp *bp) struct pps_device *pps; char buf[32]; - if (bp->gnss_port.line != -1) { - sprintf(buf, "ttyS%d", bp->gnss_port.line); - ptp_ocp_link_child(bp, buf, "ttyGNSS"); - } - if (bp->gnss2_port.line != -1) { - sprintf(buf, "ttyS%d", bp->gnss2_port.line); - ptp_ocp_link_child(bp, buf, "ttyGNSS2"); - } - if (bp->mac_port.line != -1) { - sprintf(buf, "ttyS%d", bp->mac_port.line); - ptp_ocp_link_child(bp, buf, "ttyMAC"); - } - if (bp->nmea_port.line != -1) { - sprintf(buf, "ttyS%d", bp->nmea_port.line); - ptp_ocp_link_child(bp, buf, "ttyNMEA"); - } sprintf(buf, "ptp%d", ptp_clock_index(bp->ptp)); ptp_ocp_link_child(bp, buf, "ptp"); @@ -4419,20 +4460,21 @@ ptp_ocp_info(struct ptp_ocp *bp) ptp_ocp_phc_info(bp); - ptp_ocp_serial_info(dev, "GNSS", bp->gnss_port.line, - bp->gnss_port.baud); - ptp_ocp_serial_info(dev, "GNSS2", bp->gnss2_port.line, - bp->gnss2_port.baud); - ptp_ocp_serial_info(dev, "MAC", bp->mac_port.line, bp->mac_port.baud); - if (bp->nmea_out && bp->nmea_port.line != -1) { - bp->nmea_port.baud = -1; + ptp_ocp_serial_info(dev, "GNSS", bp->port[PORT_GNSS].line, + bp->port[PORT_GNSS].baud); + ptp_ocp_serial_info(dev, "GNSS2", bp->port[PORT_GNSS2].line, + bp->port[PORT_GNSS2].baud); + ptp_ocp_serial_info(dev, "MAC", bp->port[PORT_MAC].line, + bp->port[PORT_MAC].baud); + if (bp->nmea_out && bp->port[PORT_NMEA].line != -1) { + bp->port[PORT_NMEA].baud = -1; reg = ioread32(&bp->nmea_out->uart_baud); if (reg < ARRAY_SIZE(nmea_baud)) - bp->nmea_port.baud = nmea_baud[reg]; + bp->port[PORT_NMEA].baud = nmea_baud[reg]; - ptp_ocp_serial_info(dev, "NMEA", bp->nmea_port.line, - bp->nmea_port.baud); + ptp_ocp_serial_info(dev, "NMEA", bp->port[PORT_NMEA].line, + bp->port[PORT_NMEA].baud); } } @@ -4441,9 +4483,6 @@ ptp_ocp_detach_sysfs(struct ptp_ocp *bp) { struct device *dev = &bp->dev; - sysfs_remove_link(&dev->kobj, "ttyGNSS"); - sysfs_remove_link(&dev->kobj, "ttyGNSS2"); - sysfs_remove_link(&dev->kobj, "ttyMAC"); sysfs_remove_link(&dev->kobj, "ptp"); sysfs_remove_link(&dev->kobj, "pps"); } @@ -4473,14 +4512,9 @@ ptp_ocp_detach(struct ptp_ocp *bp) for (i = 0; i < 4; i++) if (bp->signal_out[i]) ptp_ocp_unregister_ext(bp->signal_out[i]); - if (bp->gnss_port.line != -1) - serial8250_unregister_port(bp->gnss_port.line); - if (bp->gnss2_port.line != -1) - serial8250_unregister_port(bp->gnss2_port.line); - if (bp->mac_port.line != -1) - serial8250_unregister_port(bp->mac_port.line); - if (bp->nmea_port.line != -1) - serial8250_unregister_port(bp->nmea_port.line); + for (i = 0; i < PORT_NUM_MAX; i++) + if (bp->port[i].line != -1) + serial8250_unregister_port(bp->port[i].line); platform_device_unregister(bp->spi_flash); platform_device_unregister(bp->i2c_ctrl); if (bp->i2c_clk)
Starting v6.8 the serial port subsystem changed the hierarchy of devices and symlinks are not working anymore. Previous discussion made it clear that the idea of symlinks for tty devices was wrong by design. Implement additional attributes to expose the information. Fixes tag points to the commit which introduced the change. Fixes: b286f4e87e32 ("serial: core: Move tty and serdev to be children of serial core port device") Signed-off-by: Vadim Fedorenko <vadfed@meta.com> --- v2 -> v3: - replace serial ports definitions with array and enum for index - replace pointer math with direct array access - nit in documentation spelling v1 -> v2: - add Documentation/ABI changes --- drivers/ptp/ptp_ocp.c | 158 +++++++++++++++++++++++++----------------- 1 file changed, 96 insertions(+), 62 deletions(-)