From 93079162bf0ed2934c7b0c3ee93ba894df8fb3cd Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Wed, 11 Dec 2013 17:06:14 +0100 Subject: scsi_transport_srp: Fix a race condition The rport timers must be stopped before the SRP initiator destroys the resources associated with the SCSI host. This is necessary because otherwise the callback functions invoked from the SRP transport layer could trigger a use-after-free. Stopping the rport timers before invoking scsi_remove_host() can trigger long delays in the SCSI error handler if a transport layer failure occurs while scsi_remove_host() is in progress. Hence move the code for stopping the rport timers from srp_rport_release() into a new function and invoke that function after scsi_remove_host() has finished. This patch fixes the following sporadic kernel crash: kernel BUG at include/asm-generic/dma-mapping-common.h:64! invalid opcode: 0000 [#1] SMP RIP: 0010:[] [] srp_unmap_data+0x121/0x130 [ib_srp] Call Trace: [] srp_free_req+0x3c/0x80 [ib_srp] [] srp_finish_req+0x48/0x70 [ib_srp] [] srp_terminate_io+0x4b/0x60 [ib_srp] [] __rport_fail_io_fast+0x75/0x80 [scsi_transport_srp] [] rport_fast_io_fail_timedout+0x88/0xc0 [scsi_transport_srp] [] worker_thread+0x170/0x2a0 [] kthread+0x96/0xa0 [] child_rip+0xa/0x20 Signed-off-by: Bart Van Assche Signed-off-by: Roland Dreier --- include/scsi/scsi_transport_srp.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include/scsi') diff --git a/include/scsi/scsi_transport_srp.h b/include/scsi/scsi_transport_srp.h index 4ebf6913b7b2..f24e763fa430 100644 --- a/include/scsi/scsi_transport_srp.h +++ b/include/scsi/scsi_transport_srp.h @@ -19,7 +19,7 @@ struct srp_rport_identifiers { * @SRP_RPORT_BLOCKED: Transport layer not operational; fast I/O fail timer * is running and I/O has been blocked. * @SRP_RPORT_FAIL_FAST: Fast I/O fail timer has expired; fail I/O fast. - * @SRP_RPORT_LOST: Device loss timer has expired; port is being removed. + * @SRP_RPORT_LOST: Port is being removed. */ enum srp_rport_state { SRP_RPORT_RUNNING, @@ -48,7 +48,6 @@ struct srp_rport { struct mutex mutex; enum srp_rport_state state; - bool deleted; int reconnect_delay; int failed_reconnects; struct delayed_work reconnect_work; @@ -101,6 +100,7 @@ extern int srp_tmo_valid(int reconnect_delay, int fast_io_fail_tmo, extern int srp_reconnect_rport(struct srp_rport *rport); extern void srp_start_tl_fail_timers(struct srp_rport *rport); extern void srp_remove_host(struct Scsi_Host *); +extern void srp_stop_rport_timers(struct srp_rport *rport); /** * srp_chkready() - evaluate the transport layer state before I/O -- cgit v1.2.1 From 0c7f82189d2249153ee302fe4d02de15f165d220 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Mon, 13 Jan 2014 12:39:35 +0100 Subject: scsi_transport_srp: Fix kernel-doc warnings The following command has been used to verify that the kernel-doc tool no longer complains about undocumented fields: scripts/kernel-doc -html drivers/scsi/scsi_transport_srp.c \ include/scsi/scsi_transport_srp.h >srp-transport-doc.html Signed-off-by: Bart Van Assche Acked-by: Sebastian Riemer Acked-by: Randy Dunlap Signed-off-by: Roland Dreier --- include/scsi/scsi_transport_srp.h | 32 ++++++++++++++++++++++++++++---- 1 file changed, 28 insertions(+), 4 deletions(-) (limited to 'include/scsi') diff --git a/include/scsi/scsi_transport_srp.h b/include/scsi/scsi_transport_srp.h index f24e763fa430..b11da5c1331e 100644 --- a/include/scsi/scsi_transport_srp.h +++ b/include/scsi/scsi_transport_srp.h @@ -29,10 +29,26 @@ enum srp_rport_state { }; /** - * struct srp_rport - * @lld_data: LLD private data. - * @mutex: Protects against concurrent rport reconnect / fast_io_fail / - * dev_loss_tmo activity. + * struct srp_rport - SRP initiator or target port + * + * Fields that are relevant for SRP initiator and SRP target drivers: + * @dev: Device associated with this rport. + * @port_id: 16-byte port identifier. + * @roles: Role of this port - initiator or target. + * + * Fields that are only relevant for SRP initiator drivers: + * @lld_data: LLD private data. + * @mutex: Protects against concurrent rport reconnect / + * fast_io_fail / dev_loss_tmo activity. + * @state: rport state. + * @deleted: Whether or not srp_rport_del() has already been invoked. + * @reconnect_delay: Reconnect delay in seconds. + * @failed_reconnects: Number of failed reconnect attempts. + * @reconnect_work: Work structure used for scheduling reconnect attempts. + * @fast_io_fail_tmo: Fast I/O fail timeout in seconds. + * @dev_loss_tmo: Device loss timeout in seconds. + * @fast_io_fail_work: Work structure used for scheduling fast I/O fail work. + * @dev_loss_work: Work structure used for scheduling device loss work. */ struct srp_rport { /* for initiator and target drivers */ @@ -59,6 +75,8 @@ struct srp_rport { /** * struct srp_function_template + * + * Fields that are only relevant for SRP initiator drivers: * @has_rport_state: Whether or not to create the state, fast_io_fail_tmo and * dev_loss_tmo sysfs attribute for an rport. * @reset_timer_if_blocked: Whether or srp_timed_out() should reset the command @@ -70,6 +88,11 @@ struct srp_rport { * srp_reconnect_rport(). * @terminate_rport_io: Callback function for terminating all outstanding I/O * requests for an rport. + * @rport_delete: Callback function that deletes an rport. + * + * Fields that are only relevant for SRP target drivers: + * @tsk_mgmt_response: Callback function for sending a task management response. + * @it_nexus_response: Callback function for processing an IT nexus response. */ struct srp_function_template { /* for initiator drivers */ @@ -104,6 +127,7 @@ extern void srp_stop_rport_timers(struct srp_rport *rport); /** * srp_chkready() - evaluate the transport layer state before I/O + * @rport: SRP target port pointer. * * Returns a SCSI result code that can be returned by the LLD queuecommand() * implementation. The role of this function is similar to that of -- cgit v1.2.1