Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1449834

[PATCH 3.14 48/53] scsi: fix race between simultaneous decrements of ->host_failed

From Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Newsgroups linux.kernel
Subject [PATCH 3.14 48/53] scsi: fix race between simultaneous decrements of ->host_failed
Date 2016-07-25 23:00 +0200
Message-ID <rYV8M-Pn-81@gated-at.bofh.it> (permalink)
References <rYV8K-Pn-3@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw

3.14-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Wei Fang <fangwei1@huawei.com>

commit 72d8c36ec364c82bf1bf0c64dfa1041cfaf139f7 upstream.

sas_ata_strategy_handler() adds the works of the ata error handler to
system_unbound_wq. This workqueue asynchronously runs work items, so the
ata error handler will be performed concurrently on different CPUs. In
this case, ->host_failed will be decreased simultaneously in
scsi_eh_finish_cmd() on different CPUs, and become abnormal.

It will lead to permanently inequality between ->host_failed and
->host_busy, and scsi error handler thread won't start running. IO
errors after that won't be handled.

Since all scmds must have been handled in the strategy handler, just
remove the decrement in scsi_eh_finish_cmd() and zero ->host_busy after
the strategy handler to fix this race.

Fixes: 50824d6c5657 ("[SCSI] libsas: async ata-eh")
Signed-off-by: Wei Fang <fangwei1@huawei.com>
Reviewed-by: James Bottomley <jejb@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 Documentation/scsi/scsi_eh.txt |    8 ++++++--
 drivers/ata/libata-eh.c        |    2 +-
 drivers/scsi/scsi_error.c      |    4 +++-
 3 files changed, 10 insertions(+), 4 deletions(-)

--- a/Documentation/scsi/scsi_eh.txt
+++ b/Documentation/scsi/scsi_eh.txt
@@ -263,19 +263,23 @@ scmd->allowed.
 
  3. scmd recovered
     ACTION: scsi_eh_finish_cmd() is invoked to EH-finish scmd
-	- shost->host_failed--
 	- clear scmd->eh_eflags
 	- scsi_setup_cmd_retry()
 	- move from local eh_work_q to local eh_done_q
     LOCKING: none
+    CONCURRENCY: at most one thread per separate eh_work_q to
+		 keep queue manipulation lockless
 
  4. EH completes
     ACTION: scsi_eh_flush_done_q() retries scmds or notifies upper
-	    layer of failure.
+	    layer of failure. May be called concurrently but must have
+	    a no more than one thread per separate eh_work_q to
+	    manipulate the queue locklessly
 	- scmd is removed from eh_done_q and scmd->eh_entry is cleared
 	- if retry is necessary, scmd is requeued using
           scsi_queue_insert()
 	- otherwise, scsi_finish_command() is invoked for scmd
+	- zero shost->host_failed
     LOCKING: queue or finish function performs appropriate locking
 
 
--- a/drivers/ata/libata-eh.c
+++ b/drivers/ata/libata-eh.c
@@ -604,7 +604,7 @@ void ata_scsi_error(struct Scsi_Host *ho
 	ata_scsi_port_error_handler(host, ap);
 
 	/* finish or retry handled scmd's and clean up */
-	WARN_ON(host->host_failed || !list_empty(&eh_work_q));
+	WARN_ON(!list_empty(&eh_work_q));
 
 	DPRINTK("EXIT\n");
 }
--- a/drivers/scsi/scsi_error.c
+++ b/drivers/scsi/scsi_error.c
@@ -1111,7 +1111,6 @@ static int scsi_eh_action(struct scsi_cm
  */
 void scsi_eh_finish_cmd(struct scsi_cmnd *scmd, struct list_head *done_q)
 {
-	scmd->device->host->host_failed--;
 	scmd->eh_eflags = 0;
 	list_move_tail(&scmd->eh_entry, done_q);
 }
@@ -2193,6 +2192,9 @@ int scsi_error_handler(void *data)
 		else
 			scsi_unjam_host(shost);
 
+		/* All scmds have been handled */
+		shost->host_failed = 0;
+
 		/*
 		 * Note - if the above fails completely, the action is to take
 		 * individual devices offline and flush the queue of any

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH 3.14 00/53] 3.14.74-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 25/53] signal: remove warning about using SI_TKILL in rt_[tg]sigqueueinfo Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 16/53] kprobes/x86: Clear TF bit in fault on single-stepping Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 40/53] iio: accel: kxsd9: fix the usage of spi_w8r8() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 39/53] staging: iio: accel: fix error check Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 24/53] nfsd: check permissions when setting ACLs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 03/53] ipmr/ip6mr: Initialize the last assert time of mfc entries. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 05/53] usb: quirks: Add no-lpm quirk for Acer C120 LED Projector Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 17/53] x86/amd_nb: Fix boot crash on non-AMD systems Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 48/53] scsi: fix race between simultaneous decrements of ->host_failed Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:00 +0200
  [PATCH 3.14 49/53] Fix reconnect to not defer smb3 session reconnect long after socket reconnect Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:10 +0200
  [PATCH 3.14 47/53] ALSA: ctl: Stop notification after disconnection Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:10 +0200
  [PATCH 3.14 36/53] perf/x86: Honor the architectural performance monitoring version Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 52/53] tmpfs: fix regression hang in fallocate undo Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 43/53] iio:ad7266: Fix probe deferral for vref Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 15/53] x86, build: copy ldlinux.c32 to image.iso Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 50/53] xen/acpi: allow xen-acpi-processor driver to load on Xen 4.7 Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 31/53] HID: hiddev: validate num_values for HIDIOCGUSAGES, HIDIOCSUSAGES commands Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 13/53] powerpc/tm: Always reclaim in start_thread() for exec() class syscalls Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 41/53] iio:ad7266: Fix broken regulator error handling Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 46/53] ALSA: au88x0: Fix calculation in vortex_wtdma_bufshift() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 45/53] ALSA: dummy: Fix a use-after-free at closing Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 11/53] powerpc/iommu: Remove the dependency on EEH struct in DDW mechanism Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 53/53] s390/seccomp: fix error return for filtered system calls Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 44/53] tty/vt/keyboard: fix OOB access in do_compute_shiftstate() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 12/53] powerpc/pseries: Fix PCI config address for DDW Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 32/53] tracing: Handle NULL formats in hold_module_trace_bprintk_format() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 14/53] USB: EHCI: declare hostpc register as zero-length array Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 37/53] perf/x86: Fix undefined shift on 32-bit kernels Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 29/53] kvm: Fix irq route entries exceeding KVM_MAX_IRQ_ROUTES Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 34/53] drm/radeon: fix asic initialization for virtualized environments Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 51/53] tmpfs: dont undo fallocate past its last page Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 30/53] HID: elo: kill not flush the work Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 23/53] posix_acl: Add set_posix_acl Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 42/53] iio:ad7266: Fix support for optional regulators Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-25 23:20 +0200
  [PATCH 3.14 04/53] net: alx: Work around the DMA RX overflow issue Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 21/53] mm: Export migrate_page_move_mapping and migrate_page_copy Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 22/53] UBIFS: Implement ->migratepage() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 20/53] ARM: 8578/1: mm: ensure pmd_present only checks the valid bit Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 19/53] NFS: Fix another OPEN_DOWNGRADE bug Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 01/53] crypto: ux500 - memmove the right size Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 07/53] usb: musb: Ensure rx reinit occurs for shared_fifo endpoints Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 18/53] make nfs_atomic_open() call d_drop() on all ->open_context() errors. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 02/53] sit: correct IP protocol used in ipip6_err Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 10/53] IB/mlx4: Properly initialize GRH TClass and FlowLabel in AHs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 28/53] KEYS: potential uninitialized variable Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 09/53] mac80211_hwsim: Add missing check for HWSIM_ATTR_SIGNAL Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  [PATCH 3.14 06/53] usb: musb: Stop bulk endpoint while queue is rotated Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-07-26 00:50 +0200
  Re: [PATCH 3.14 00/53] 3.14.74-stable review Shuah Khan <shuah.kh@samsung.com> - 2016-07-26 04:10 +0200
  Re: [PATCH 3.14 00/53] 3.14.74-stable review Guenter Roeck <linux@roeck-us.net> - 2016-07-26 16:00 +0200

csiph-web