| Message ID | 20180326092741.22716-1-yanaijie@huawei.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
On Mon, Mar 26, 2018 at 05:27:41PM +0800, Jason Yan wrote: > Now ata devices attached with sas controller do not have transport > class, so that we can not see any information of these ata devices in > /sys/class/ata_port(or ata_link or ata_device). > > Add transport class for the ata devices attached with sas controller. > The /sys/class directory will show the infomation of the ata devices > as follows: > > localhost:/sys/class # ls ata* > ata_device: > dev1.0 dev2.0 > > ata_link: > link1 link2 > > ata_port: > ata1 ata2 > > No functional change of the device scanning and io path. The ata > transport class was deleted when destroying the sas devices. > > Signed-off-by: Jason Yan <yanaijie@huawei.com> > CC: Dan Williams <dan.j.williams@intel.com> > CC: Tejun Heo <tj@kernel.org> Looks good to me on the ata side. Acked-by: Tejun Heo <tj@kernel.org> Thanks.
On Mon, Mar 26, 2018 at 2:27 AM, Jason Yan <yanaijie@huawei.com> wrote: > Now ata devices attached with sas controller do not have transport > class, so that we can not see any information of these ata devices in > /sys/class/ata_port(or ata_link or ata_device). > > Add transport class for the ata devices attached with sas controller. > The /sys/class directory will show the infomation of the ata devices > as follows: > > localhost:/sys/class # ls ata* > ata_device: > dev1.0 dev2.0 > > ata_link: > link1 link2 > > ata_port: > ata1 ata2 > > No functional change of the device scanning and io path. The ata > transport class was deleted when destroying the sas devices. Have you tested this with suspend / resume power management? I believe that is what stopped me from going down this path, when I first evaluated it.
On 2018/3/26 22:38, Dan Williams wrote: > On Mon, Mar 26, 2018 at 2:27 AM, Jason Yan <yanaijie@huawei.com> wrote: >> Now ata devices attached with sas controller do not have transport >> class, so that we can not see any information of these ata devices in >> /sys/class/ata_port(or ata_link or ata_device). >> >> Add transport class for the ata devices attached with sas controller. >> The /sys/class directory will show the infomation of the ata devices >> as follows: >> >> localhost:/sys/class # ls ata* >> ata_device: >> dev1.0 dev2.0 >> >> ata_link: >> link1 link2 >> >> ata_port: >> ata1 ata2 >> >> No functional change of the device scanning and io path. The ata >> transport class was deleted when destroying the sas devices. > > Have you tested this with suspend / resume power management? I believe > that is what stopped me from going down this path, when I first > evaluated it. > Thanks a lot, I will test it with suspend/resume power management later. >
Hi, Dan On 2018/3/26 22:38, Dan Williams wrote: > On Mon, Mar 26, 2018 at 2:27 AM, Jason Yan <yanaijie@huawei.com> wrote: >> Now ata devices attached with sas controller do not have transport >> class, so that we can not see any information of these ata devices in >> /sys/class/ata_port(or ata_link or ata_device). >> >> Add transport class for the ata devices attached with sas controller. >> The /sys/class directory will show the infomation of the ata devices >> as follows: >> >> localhost:/sys/class # ls ata* >> ata_device: >> dev1.0 dev2.0 >> >> ata_link: >> link1 link2 >> >> ata_port: >> ata1 ata2 >> >> No functional change of the device scanning and io path. The ata >> transport class was deleted when destroying the sas devices. > > Have you tested this with suspend / resume power management? I believe > that is what stopped me from going down this path, when I first > evaluated it. > I tested this patch with suspend/resume power management and did not see any regressions. I don't know if this test is enough. The log is as below: [ 1098.226768] hns-nic HISI00C2:01 eth1: link up [ 1183.162762] PM: suspend entry (s2idle) [ 1183.162766] PM: Syncing filesystems ... done. [ 1183.163224] Freezing user space processes ... (elapsed 0.001 seconds) done. [ 1183.164710] OOM killer disabled. [ 1183.164711] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done. [ 1183.166415] Suspending console(s) (use no_console_suspend to debug) [ 1183.166733] pci_bus 0004:88: 2-byte config write to 0004:88:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1183.166741] pci_bus 0005:00: 2-byte config write to 0005:00:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1183.166748] pci_bus 0006:c0: 2-byte config write to 0006:c0:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1183.166758] pci_bus 000a:10: 2-byte config write to 000a:10:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1183.166765] pci_bus 000c:20: 2-byte config write to 000c:20:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1183.166774] pci_bus 000d:30: 2-byte config write to 000d:30:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1183.242586] sd 0:0:11:0: [sdl] Synchronizing SCSI cache [ 1183.258568] sd 0:0:10:0: [sdk] Synchronizing SCSI cache [ 1183.258699] sd 0:0:10:0: [sdk] Stopping disk [ 1183.922568] sd 0:0:9:0: [sdj] Synchronizing SCSI cache [ 1183.922690] sd 0:0:9:0: [sdj] Stopping disk [ 1184.610567] sd 0:0:8:0: [sdi] Synchronizing SCSI cache [ 1184.626567] sd 0:0:7:0: [sdh] Synchronizing SCSI cache [ 1184.642564] sd 0:0:6:0: [sdg] Synchronizing SCSI cache [ 1184.658567] sd 0:0:5:0: [sdf] Synchronizing SCSI cache [ 1184.674565] sd 0:0:4:0: [sde] Synchronizing SCSI cache [ 1184.690567] sd 0:0:3:0: [sdd] Synchronizing SCSI cache [ 1184.706564] sd 0:0:2:0: [sdc] Synchronizing SCSI cache [ 1184.722566] sd 0:0:1:0: [sdb] Synchronizing SCSI cache [ 1184.738563] sd 0:0:0:0: [sda] Synchronizing SCSI cache [ 1184.738844] sas: Enter sas_scsi_recover_host busy: 0 failed: 0 [ 1184.738857] sas: ata1: end_device-0:0:9: dev error handler [ 1184.738860] sas: ata2: end_device-0:0:10: dev error handler [ 1184.738877] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1 [ 1184.739169] pci_bus 000d:30: 2-byte config write to 000d:30:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1184.739179] pci_bus 000c:20: 2-byte config write to 000c:20:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1184.739188] pci_bus 000a:10: 2-byte config write to 000a:10:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1184.739214] pci_bus 0006:c0: 2-byte config write to 0006:c0:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1184.739334] pcieport 0002:80:00.0: can't derive routing for PCI INT A [ 1184.739336] ixgbe 0002:81:00.0: PCI INT A: no GSI [ 1184.758563] PM: suspend debug: Waiting for 5 second(s). [ 1189.809010] sas: Enter sas_scsi_recover_host busy: 0 failed: 0 [ 1189.811180] sas: ata1: end_device-0:0:9: dev error handler [ 1189.811185] sas: ata1: end_device-0:0:9: Unable to reset ata device? [ 1189.811223] sas: ata2: end_device-0:0:10: dev error handler [ 1189.811226] sas: ata2: end_device-0:0:10: Unable to reset ata device? [ 1189.811474] sd 0:0:9:0: [sdj] Starting disk [ 1189.888654] ixgbe 0002:81:00.0: Multiqueue Enabled: Rx Queue count = 63, Tx Queue count = 63 XDP Queue count = 0 [ 1189.974214] ata2.00: configured for UDMA/133 [ 1189.974285] ata1.00: configured for UDMA/133 [ 1189.974295] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1 [ 1197.991137] sd 0:0:10:0: [sdk] Starting disk [ 1210.019736] pci_generic_config_write32: 2 callbacks suppressed [ 1210.019740] pci_bus 000d:30: 2-byte config write to 000d:30:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1210.019748] pci_bus 000c:20: 2-byte config write to 000c:20:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1210.019755] pci_bus 000a:10: 2-byte config write to 000a:10:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1210.019764] pci_bus 0006:c0: 2-byte config write to 0006:c0:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1210.019770] pci_bus 0005:00: 2-byte config write to 0005:00:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1210.019775] pci_bus 0004:88: 2-byte config write to 0004:88:00.0 offset 0x44 may corrupt adjacent RW1C bits [ 1210.019888] OOM killer enabled. [ 1210.019889] Restarting tasks ... done. [ 1210.021064] PM: suspend exit localhost:/sys/power # ls /sys/class/ata* /sys/class/ata_device: dev1.0 dev2.0 /sys/class/ata_link: link1 link2 /sys/class/ata_port: ata1 ata2 localhost:/sys/power # localhost:/sys/power # lsscsi [0:0:0:0] disk SEAGATE ST3600057SS 0008 /dev/sda [0:0:1:0] disk SEAGATE ST3600057SS 0008 /dev/sdb [0:0:2:0] disk SEAGATE ST3600057SS 0008 /dev/sdc [0:0:3:0] disk SEAGATE ST3600057SS 0008 /dev/sdd [0:0:4:0] disk SEAGATE ST3600057SS 0008 /dev/sde [0:0:5:0] disk SEAGATE ST3600057SS 0008 /dev/sdf [0:0:6:0] disk SEAGATE ST3600057SS 0008 /dev/sdg [0:0:7:0] disk SEAGATE ST3600057SS 0008 /dev/sdh [0:0:8:0] disk SEAGATE ST3600057SS 0008 /dev/sdi [0:0:9:0] disk ATA HGST HUS724040AL A8B0 /dev/sdj [0:0:10:0] disk ATA HGST HUS724040AL A8B0 /dev/sdk [0:0:11:0] disk SEAGATE ST3600057SS 0008 /dev/sdl [0:0:12:0] enclosu 12G SAS Expander RevB - Jason >
Jason, > Now ata devices attached with sas controller do not have transport > class, so that we can not see any information of these ata devices in > /sys/class/ata_port(or ata_link or ata_device). > > Add transport class for the ata devices attached with sas controller. > The /sys/class directory will show the infomation of the ata devices > as follows: Applied to 4.18/scsi-queue. Thank you!
diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c index 89a9d4a2efc8..1c9f80fbc51c 100644 --- a/drivers/ata/libata-scsi.c +++ b/drivers/ata/libata-scsi.c @@ -5051,6 +5051,18 @@ int ata_sas_port_init(struct ata_port *ap) } EXPORT_SYMBOL_GPL(ata_sas_port_init); +int ata_sas_tport_add(struct device *parent, struct ata_port *ap) +{ + return ata_tport_add(parent, ap); +} +EXPORT_SYMBOL_GPL(ata_sas_tport_add); + +void ata_sas_tport_delete(struct ata_port *ap) +{ + ata_tport_delete(ap); +} +EXPORT_SYMBOL_GPL(ata_sas_tport_delete); + /** * ata_sas_port_destroy - Destroy a SATA port allocated by ata_sas_port_alloc * @ap: SATA port to destroy diff --git a/drivers/scsi/libsas/sas_ata.c b/drivers/scsi/libsas/sas_ata.c index 2b3637b40dde..2457f12c7953 100644 --- a/drivers/scsi/libsas/sas_ata.c +++ b/drivers/scsi/libsas/sas_ata.c @@ -577,6 +577,11 @@ int sas_ata_init(struct domain_device *found_dev) ata_sas_port_destroy(ap); return rc; } + rc = ata_sas_tport_add(found_dev->sata_dev.ata_host.dev, ap); + if (rc) { + ata_sas_port_destroy(ap); + return rc; + } found_dev->sata_dev.ap = ap; return 0; diff --git a/drivers/scsi/libsas/sas_discover.c b/drivers/scsi/libsas/sas_discover.c index e4fd078e4175..1ed47ccfc280 100644 --- a/drivers/scsi/libsas/sas_discover.c +++ b/drivers/scsi/libsas/sas_discover.c @@ -314,6 +314,7 @@ void sas_free_device(struct kref *kref) kfree(dev->ex_dev.ex_phy); if (dev_is_sata(dev) && dev->sata_dev.ap) { + ata_sas_tport_delete(dev->sata_dev.ap); ata_sas_port_destroy(dev->sata_dev.ap); dev->sata_dev.ap = NULL; } diff --git a/include/linux/libata.h b/include/linux/libata.h index ed9826b21c5e..86858d0a372c 100644 --- a/include/linux/libata.h +++ b/include/linux/libata.h @@ -1129,6 +1129,8 @@ extern void ata_sas_async_probe(struct ata_port *ap); extern int ata_sas_sync_probe(struct ata_port *ap); extern int ata_sas_port_init(struct ata_port *); extern int ata_sas_port_start(struct ata_port *ap); +extern int ata_sas_tport_add(struct device *parent, struct ata_port *ap); +extern void ata_sas_tport_delete(struct ata_port *ap); extern void ata_sas_port_stop(struct ata_port *ap); extern int ata_sas_slave_configure(struct scsi_device *, struct ata_port *); extern int ata_sas_queuecmd(struct scsi_cmnd *cmd, struct ata_port *ap);
Now ata devices attached with sas controller do not have transport class, so that we can not see any information of these ata devices in /sys/class/ata_port(or ata_link or ata_device). Add transport class for the ata devices attached with sas controller. The /sys/class directory will show the infomation of the ata devices as follows: localhost:/sys/class # ls ata* ata_device: dev1.0 dev2.0 ata_link: link1 link2 ata_port: ata1 ata2 No functional change of the device scanning and io path. The ata transport class was deleted when destroying the sas devices. Signed-off-by: Jason Yan <yanaijie@huawei.com> CC: Dan Williams <dan.j.williams@intel.com> CC: Tejun Heo <tj@kernel.org> --- drivers/ata/libata-scsi.c | 12 ++++++++++++ drivers/scsi/libsas/sas_ata.c | 5 +++++ drivers/scsi/libsas/sas_discover.c | 1 + include/linux/libata.h | 2 ++ 4 files changed, 20 insertions(+)