phy->attached_sas_addr will be set to a zero-address when
phy->linkrate < SAS_LINK_RATE_1_5_GBPS but device-type != NO_DEVICE,
and it may trigger BUG() as follows when do revalidate with zero-address:
[562240.062536] sas: ex 500e004aaaaaaa1f phy0 new device attached
[562240.062616] sas: ex 500e004aaaaaaa1f phy00:U:5 attached: 0000000000000000 (stp)
[562240.062680] port-7:7:0: trying to add phy phy-7:7:19 fails: it's already part of another port
[562240.085064] ------------[ cut here ]------------
[562240.096612] kernel BUG at drivers/scsi/scsi_transport_sas.c:1083!
[562240.109611] Internal error: Oops - BUG: 0 [#1] SMP
[562240.343518] Process kworker/u256:3 (pid: 435909, stack limit = 0x0000000003bcbebf)
[562240.421714] Workqueue: 0000:b4:02.0_disco_q sas_revalidate_domain [libsas]
[562240.437173] pstate: 40c00009 (nZcv daif +PAN +UAO)
[562240.450478] pc : sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
[562240.465283] lr : sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
[562240.479751] sp : ffff0000300cfa70
[562240.674822] Call trace:
[562240.682709] sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
[562240.694013] sas_ex_get_linkrate.isra.5+0xcc/0x128 [libsas]
[562240.704957] sas_ex_discover_end_dev+0xfc/0x538 [libsas]
[562240.715508] sas_ex_discover_dev+0x3cc/0x4b8 [libsas]
[562240.725634] sas_ex_discover_devices+0x9c/0x1a8 [libsas]
[562240.735855] sas_ex_revalidate_domain+0x2f0/0x450 [libsas]
[562240.746123] sas_revalidate_domain+0x158/0x160 [libsas]
[562240.756014] process_one_work+0x1b4/0x448
[562240.764548] worker_thread+0x54/0x468
[562240.772562] kthread+0x134/0x138
[562240.779989] ret_from_fork+0x10/0x18
So set a zero-address for phy->attached_sas_addr only when
phy->attached_dev_type == NO_DEVICE.
Fixes: 7d1d86518118 ("[SCSI] libsas: fix false positive 'device attached' conditions")
Signed-off-by: Xingui Yang <[email protected]>
---
drivers/scsi/libsas/sas_expander.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/scsi/libsas/sas_expander.c b/drivers/scsi/libsas/sas_expander.c
index a2204674b680..5a81754d3768 100644
--- a/drivers/scsi/libsas/sas_expander.c
+++ b/drivers/scsi/libsas/sas_expander.c
@@ -239,8 +239,7 @@ static void sas_set_ex_phy(struct domain_device *dev, int phy_id,
/* help some expanders that fail to zero sas_address in the 'no
* device' case
*/
- if (phy->attached_dev_type == SAS_PHY_UNUSED ||
- phy->linkrate < SAS_LINK_RATE_1_5_GBPS)
+ if (phy->attached_dev_type == SAS_PHY_UNUSED)
memset(phy->attached_sas_addr, 0, SAS_ADDR_SIZE);
else
memcpy(phy->attached_sas_addr, dr->attached_sas_addr, SAS_ADDR_SIZE);
--
2.17.1
On 2023/10/20 10:42, Xingui Yang wrote:
> phy->attached_sas_addr will be set to a zero-address when
> phy->linkrate < SAS_LINK_RATE_1_5_GBPS but device-type != NO_DEVICE,
> and it may trigger BUG() as follows when do revalidate with zero-address:
Hi Xingui,
Why is this zero-addressed PHY added to another port? A zero-addressed
PHY should not belong to any port.
Thanks,
Jason
>
> [562240.062536] sas: ex 500e004aaaaaaa1f phy0 new device attached
> [562240.062616] sas: ex 500e004aaaaaaa1f phy00:U:5 attached: 0000000000000000 (stp)
> [562240.062680] port-7:7:0: trying to add phy phy-7:7:19 fails: it's already part of another port
> [562240.085064] ------------[ cut here ]------------
> [562240.096612] kernel BUG at drivers/scsi/scsi_transport_sas.c:1083!
> [562240.109611] Internal error: Oops - BUG: 0 [#1] SMP
> [562240.343518] Process kworker/u256:3 (pid: 435909, stack limit = 0x0000000003bcbebf)
> [562240.421714] Workqueue: 0000:b4:02.0_disco_q sas_revalidate_domain [libsas]
> [562240.437173] pstate: 40c00009 (nZcv daif +PAN +UAO)
> [562240.450478] pc : sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
> [562240.465283] lr : sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
> [562240.479751] sp : ffff0000300cfa70
> [562240.674822] Call trace:
> [562240.682709] sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
> [562240.694013] sas_ex_get_linkrate.isra.5+0xcc/0x128 [libsas]
> [562240.704957] sas_ex_discover_end_dev+0xfc/0x538 [libsas]
> [562240.715508] sas_ex_discover_dev+0x3cc/0x4b8 [libsas]
> [562240.725634] sas_ex_discover_devices+0x9c/0x1a8 [libsas]
> [562240.735855] sas_ex_revalidate_domain+0x2f0/0x450 [libsas]
> [562240.746123] sas_revalidate_domain+0x158/0x160 [libsas]
> [562240.756014] process_one_work+0x1b4/0x448
> [562240.764548] worker_thread+0x54/0x468
> [562240.772562] kthread+0x134/0x138
> [562240.779989] ret_from_fork+0x10/0x18
>
> So set a zero-address for phy->attached_sas_addr only when
> phy->attached_dev_type == NO_DEVICE.
>
> Fixes: 7d1d86518118 ("[SCSI] libsas: fix false positive 'device attached' conditions")
>
> Signed-off-by: Xingui Yang <[email protected]>
> ---
> drivers/scsi/libsas/sas_expander.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/drivers/scsi/libsas/sas_expander.c b/drivers/scsi/libsas/sas_expander.c
> index a2204674b680..5a81754d3768 100644
> --- a/drivers/scsi/libsas/sas_expander.c
> +++ b/drivers/scsi/libsas/sas_expander.c
> @@ -239,8 +239,7 @@ static void sas_set_ex_phy(struct domain_device *dev, int phy_id,
> /* help some expanders that fail to zero sas_address in the 'no
> * device' case
> */
> - if (phy->attached_dev_type == SAS_PHY_UNUSED ||
> - phy->linkrate < SAS_LINK_RATE_1_5_GBPS)
> + if (phy->attached_dev_type == SAS_PHY_UNUSED)
> memset(phy->attached_sas_addr, 0, SAS_ADDR_SIZE);
> else
> memcpy(phy->attached_sas_addr, dr->attached_sas_addr, SAS_ADDR_SIZE);
>
On 2023/10/20 11:24, Jason Yan wrote:
> On 2023/10/20 10:42, Xingui Yang wrote:
>> phy->attached_sas_addr will be set to a zero-address when
>> phy->linkrate < SAS_LINK_RATE_1_5_GBPS but device-type != NO_DEVICE,
>> and it may trigger BUG() as follows when do revalidate with zero-address:
>
> Hi Xingui,
>
> Why is this zero-addressed PHY added to another port? A zero-addressed
> PHY should not belong to any port.
Hi Jason,
[562240.062536] sas: ex 500e004aaaaaaa1f phy0 new device attached
[562240.062616] sas: ex 500e004aaaaaaa1f phy00:U:5 attached:
0000000000000000 (stp)
port-7:7:0: trying to add phy phy-7:7:19 fails: it's already part of
another port
we get phy0's data info:
crash> struct ex_phy 0xffff8020cead3000
struct ex_phy {
phy_id = 0,
phy_state = PHY_EMPTY,
attached_dev_type = SAS_END_DEVICE,
linkrate = SAS_PHY_RESET_IN_PROGRESS,
attached_sata_host = 0 '\000',
attached_sata_dev = 1 '\001',
attached_sata_ps = 0 '\000',
attached_tproto = SAS_PROTOCOL_NONE,
attached_iproto = SAS_PROTOCOL_NONE,
attached_sas_addr = "\000\000\000\000\000\000\000",
attached_phy_id = 0 '\000',
phy_change_count = 152,
routing_attr = TABLE_ROUTING,
virtual = 0 '\000',
last_da_index = -1,
phy = 0xffff8020ceac5000,
port = 0xffffa0300c6f8800
}
When a new device is attached, if the address is 0 and the device type
is not null, but stp or ssp, the device still creates a port in
sas_ex_discover_end_dev() and add all other expander phys with the same
sas_address 0 to this port in sas_ex_get_linkrate().
All of the phys is zero-address belongs to port-7:7:0 as follows:
crash> list sas_phy.port_siblings -s sas_phy.dev.kobj.name -s
sas_phy.identify.sas_address -h ffff8020ceac5000
ffff8020ceac5000
dev.kobj.name = 0xffff803017878b00 "phy-7:7:0",
identify.sas_address = 0,
ffff8020ceacb000
dev.kobj.name = 0xffff803017871980 "phy-7:7:1",
identify.sas_address = 0,
ffff8020ceade000
dev.kobj.name = 0xffff80301787f580 "phy-7:7:2",
identify.sas_address = 0,
ffff8020cead4000
dev.kobj.name = 0xffff80301787d980 "phy-7:7:3",
identify.sas_address = 0,
ffff8020ceacc800
dev.kobj.name = 0xffff80301787a480 "phy-7:7:4",
identify.sas_address = 0,
ffff8020ceac6000
dev.kobj.name = 0xffff803017872500 "phy-7:7:5",
identify.sas_address = 0,
ffff8020ceacd000
dev.kobj.name = 0xffff803041e62100 "phy-7:7:12",
identify.sas_address = 0,
ffff8020ceadb000
dev.kobj.name = 0xffff803041e61980 "phy-7:7:13",
identify.sas_address = 0,
ffff8020ceac8800
dev.kobj.name = 0xffff803041e61a00 "phy-7:7:14",
identify.sas_address = 0,
ffff8020cead1800
dev.kobj.name = 0xffff803041e69980 "phy-7:7:15",
identify.sas_address = 0,
Thanks,
Xingui
>
> Thanks,
> Jason
>
>>
>> [562240.062536] sas: ex 500e004aaaaaaa1f phy0 new device attached
>> [562240.062616] sas: ex 500e004aaaaaaa1f phy00:U:5 attached:
>> 0000000000000000 (stp)
>> [562240.062680] port-7:7:0: trying to add phy phy-7:7:19 fails: it's
>> already part of another port
>> [562240.085064] ------------[ cut here ]------------
>> [562240.096612] kernel BUG at drivers/scsi/scsi_transport_sas.c:1083!
>> [562240.109611] Internal error: Oops - BUG: 0 [#1] SMP
>> [562240.343518] Process kworker/u256:3 (pid: 435909, stack limit =
>> 0x0000000003bcbebf)
>> [562240.421714] Workqueue: 0000:b4:02.0_disco_q sas_revalidate_domain
>> [libsas]
>> [562240.437173] pstate: 40c00009 (nZcv daif +PAN +UAO)
>> [562240.450478] pc : sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
>> [562240.465283] lr : sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
>> [562240.479751] sp : ffff0000300cfa70
>> [562240.674822] Call trace:
>> [562240.682709] sas_port_add_phy+0x13c/0x168 [scsi_transport_sas]
>> [562240.694013] sas_ex_get_linkrate.isra.5+0xcc/0x128 [libsas]
>> [562240.704957] sas_ex_discover_end_dev+0xfc/0x538 [libsas]
>> [562240.715508] sas_ex_discover_dev+0x3cc/0x4b8 [libsas]
>> [562240.725634] sas_ex_discover_devices+0x9c/0x1a8 [libsas]
>> [562240.735855] sas_ex_revalidate_domain+0x2f0/0x450 [libsas]
>> [562240.746123] sas_revalidate_domain+0x158/0x160 [libsas]
>> [562240.756014] process_one_work+0x1b4/0x448
>> [562240.764548] worker_thread+0x54/0x468
>> [562240.772562] kthread+0x134/0x138
>> [562240.779989] ret_from_fork+0x10/0x18
>>
>> So set a zero-address for phy->attached_sas_addr only when
>> phy->attached_dev_type == NO_DEVICE.
>>
>> Fixes: 7d1d86518118 ("[SCSI] libsas: fix false positive 'device
>> attached' conditions")
>>
>> Signed-off-by: Xingui Yang <[email protected]>
>> ---
>> drivers/scsi/libsas/sas_expander.c | 3 +--
>> 1 file changed, 1 insertion(+), 2 deletions(-)
>>
>> diff --git a/drivers/scsi/libsas/sas_expander.c
>> b/drivers/scsi/libsas/sas_expander.c
>> index a2204674b680..5a81754d3768 100644
>> --- a/drivers/scsi/libsas/sas_expander.c
>> +++ b/drivers/scsi/libsas/sas_expander.c
>> @@ -239,8 +239,7 @@ static void sas_set_ex_phy(struct domain_device
>> *dev, int phy_id,
>> /* help some expanders that fail to zero sas_address in the 'no
>> * device' case
>> */
>> - if (phy->attached_dev_type == SAS_PHY_UNUSED ||
>> - phy->linkrate < SAS_LINK_RATE_1_5_GBPS)
>> + if (phy->attached_dev_type == SAS_PHY_UNUSED)
>> memset(phy->attached_sas_addr, 0, SAS_ADDR_SIZE);
>> else
>> memcpy(phy->attached_sas_addr, dr->attached_sas_addr,
>> SAS_ADDR_SIZE);
>>
> .