* [PATCH AUTOSEL 5.3 08/99] nvme: fix possible deadlock when nvme_update_formats fails
[not found] <20191026131600.2507-1-sashal@kernel.org>
@ 2019-10-26 13:14 ` Sasha Levin
2019-10-26 13:14 ` [PATCH AUTOSEL 5.3 09/99] nvme: retain split access workaround for capability reads Sasha Levin
` (4 subsequent siblings)
5 siblings, 0 replies; 6+ messages in thread
From: Sasha Levin @ 2019-10-26 13:14 UTC (permalink / raw)
To: linux-kernel, stable; +Cc: Sasha Levin, Judy Brock, Sagi Grimberg, linux-nvme
From: Sagi Grimberg <sagi@grimberg.me>
[ Upstream commit 6abff1b9f7b8884a46b7bd80b49e7af0b5625aeb ]
nvme_update_formats may fail to revalidate the namespace and
attempt to remove the namespace. This may lead to a deadlock
as nvme_ns_remove will attempt to acquire the subsystem lock
which is already acquired by the passthru command with effects.
Move the invalid namepsace removal to after the passthru command
releases the subsystem lock.
Reported-by: Judy Brock <judy.brock@samsung.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/nvme/host/core.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index d3d6b7bd69033..7284bee560a0f 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1304,8 +1304,6 @@ static void nvme_update_formats(struct nvme_ctrl *ctrl)
if (ns->disk && nvme_revalidate_disk(ns->disk))
nvme_set_queue_dying(ns);
up_read(&ctrl->namespaces_rwsem);
-
- nvme_remove_invalid_namespaces(ctrl, NVME_NSID_ALL);
}
static void nvme_passthru_end(struct nvme_ctrl *ctrl, u32 effects)
@@ -1321,6 +1319,7 @@ static void nvme_passthru_end(struct nvme_ctrl *ctrl, u32 effects)
nvme_unfreeze(ctrl);
nvme_mpath_unfreeze(ctrl->subsys);
mutex_unlock(&ctrl->subsys->lock);
+ nvme_remove_invalid_namespaces(ctrl, NVME_NSID_ALL);
mutex_unlock(&ctrl->scan_lock);
}
if (effects & NVME_CMD_EFFECTS_CCC)
--
2.20.1
_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH AUTOSEL 5.3 09/99] nvme: retain split access workaround for capability reads
[not found] <20191026131600.2507-1-sashal@kernel.org>
2019-10-26 13:14 ` [PATCH AUTOSEL 5.3 08/99] nvme: fix possible deadlock when nvme_update_formats fails Sasha Levin
@ 2019-10-26 13:14 ` Sasha Levin
2019-10-26 13:15 ` [PATCH AUTOSEL 5.3 65/99] nvme-pci: Free tagset if no IO queues Sasha Levin
` (3 subsequent siblings)
5 siblings, 0 replies; 6+ messages in thread
From: Sasha Levin @ 2019-10-26 13:14 UTC (permalink / raw)
To: linux-kernel, stable
Cc: Sasha Levin, Sagi Grimberg, linux-nvme, Ard Biesheuvel
From: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[ Upstream commit 3a8ecc935efabdad106b5e06d07b150c394b4465 ]
Commit 7fd8930f26be4
"nvme: add a common helper to read Identify Controller data"
has re-introduced an issue that we have attempted to work around in the
past, in commit a310acd7a7ea ("NVMe: use split lo_hi_{read,write}q").
The problem is that some PCIe NVMe controllers do not implement 64-bit
outbound accesses correctly, which is why the commit above switched
to using lo_hi_[read|write]q for all 64-bit BAR accesses occuring in
the code.
In the mean time, the NVMe subsystem has been refactored, and now calls
into the PCIe support layer for NVMe via a .reg_read64() method, which
fails to use lo_hi_readq(), and thus reintroduces the problem that the
workaround above aimed to address.
Given that, at the moment, .reg_read64() is only used to read the
capability register [which is known to tolerate split reads], let's
switch .reg_read64() to lo_hi_readq() as well.
This fixes a boot issue on some ARM boxes with NVMe behind a Synopsys
DesignWare PCIe host controller.
Fixes: 7fd8930f26be4 ("nvme: add a common helper to read Identify Controller data")
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/nvme/host/pci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
index 732d5b63ec054..48a71909763be 100644
--- a/drivers/nvme/host/pci.c
+++ b/drivers/nvme/host/pci.c
@@ -2620,7 +2620,7 @@ static int nvme_pci_reg_write32(struct nvme_ctrl *ctrl, u32 off, u32 val)
static int nvme_pci_reg_read64(struct nvme_ctrl *ctrl, u32 off, u64 *val)
{
- *val = readq(to_nvme_dev(ctrl)->bar + off);
+ *val = lo_hi_readq(to_nvme_dev(ctrl)->bar + off);
return 0;
}
--
2.20.1
_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH AUTOSEL 5.3 65/99] nvme-pci: Free tagset if no IO queues
[not found] <20191026131600.2507-1-sashal@kernel.org>
2019-10-26 13:14 ` [PATCH AUTOSEL 5.3 08/99] nvme: fix possible deadlock when nvme_update_formats fails Sasha Levin
2019-10-26 13:14 ` [PATCH AUTOSEL 5.3 09/99] nvme: retain split access workaround for capability reads Sasha Levin
@ 2019-10-26 13:15 ` Sasha Levin
2019-10-26 13:15 ` [PATCH AUTOSEL 5.3 66/99] nvme: Prevent resets during paused controller state Sasha Levin
` (2 subsequent siblings)
5 siblings, 0 replies; 6+ messages in thread
From: Sasha Levin @ 2019-10-26 13:15 UTC (permalink / raw)
To: linux-kernel, stable
Cc: Sasha Levin, Sagi Grimberg, James Smart, linux-nvme,
Edmund Nadolski, Keith Busch, Christoph Hellwig
From: Keith Busch <kbusch@kernel.org>
[ Upstream commit 770597ecb2075390c01c425b8b1f551347f1bd70 ]
If a controller becomes degraded after a reset, we will not be able to
perform any IO. We currently teardown previously created request
queues and namespaces, but we had kept the unusable tagset. Free
it after all queues using it have been released.
Tested-by: Edmund Nadolski <edmund.nadolski@intel.com>
Reviewed-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/nvme/host/pci.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
index 48a71909763be..f94cc01e85a66 100644
--- a/drivers/nvme/host/pci.c
+++ b/drivers/nvme/host/pci.c
@@ -2438,14 +2438,20 @@ static void nvme_release_prp_pools(struct nvme_dev *dev)
dma_pool_destroy(dev->prp_small_pool);
}
+static void nvme_free_tagset(struct nvme_dev *dev)
+{
+ if (dev->tagset.tags)
+ blk_mq_free_tag_set(&dev->tagset);
+ dev->ctrl.tagset = NULL;
+}
+
static void nvme_pci_free_ctrl(struct nvme_ctrl *ctrl)
{
struct nvme_dev *dev = to_nvme_dev(ctrl);
nvme_dbbuf_dma_free(dev);
put_device(dev->dev);
- if (dev->tagset.tags)
- blk_mq_free_tag_set(&dev->tagset);
+ nvme_free_tagset(dev);
if (dev->ctrl.admin_q)
blk_put_queue(dev->ctrl.admin_q);
kfree(dev->queues);
@@ -2564,6 +2570,7 @@ static void nvme_reset_work(struct work_struct *work)
nvme_kill_queues(&dev->ctrl);
nvme_remove_namespaces(&dev->ctrl);
new_state = NVME_CTRL_ADMIN_ONLY;
+ nvme_free_tagset(dev);
} else {
nvme_start_queues(&dev->ctrl);
nvme_wait_freeze(&dev->ctrl);
--
2.20.1
_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH AUTOSEL 5.3 66/99] nvme: Prevent resets during paused controller state
[not found] <20191026131600.2507-1-sashal@kernel.org>
` (2 preceding siblings ...)
2019-10-26 13:15 ` [PATCH AUTOSEL 5.3 65/99] nvme-pci: Free tagset if no IO queues Sasha Levin
@ 2019-10-26 13:15 ` Sasha Levin
2019-10-26 13:15 ` [PATCH AUTOSEL 5.3 74/99] nvmet-loop: fix possible leakage during error flow Sasha Levin
2019-10-26 13:15 ` [PATCH AUTOSEL 5.3 75/99] nvme-tcp: " Sasha Levin
5 siblings, 0 replies; 6+ messages in thread
From: Sasha Levin @ 2019-10-26 13:15 UTC (permalink / raw)
To: linux-kernel, stable
Cc: Keith Busch, Sasha Levin, Edmund Nadolski, Christoph Hellwig, linux-nvme
From: Keith Busch <kbusch@kernel.org>
[ Upstream commit 4c75f877853cfa81b12374a07208e07b077f39b8 ]
A paused controller is doing critical internal activation work in the
background. Prevent subsequent controller resets from occurring during
this period by setting the controller state to RESETTING first. A helper
function, nvme_try_sched_reset_work(), is introduced for these paths so
they may continue with scheduling the reset_work after they've completed
their uninterruptible critical section.
Tested-by: Edmund Nadolski <edmund.nadolski@intel.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/nvme/host/core.c | 29 +++++++++++++++++++++++++----
1 file changed, 25 insertions(+), 4 deletions(-)
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 7284bee560a0f..b214df2f05850 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -118,6 +118,21 @@ static void nvme_queue_scan(struct nvme_ctrl *ctrl)
queue_work(nvme_wq, &ctrl->scan_work);
}
+/*
+ * Use this function to proceed with scheduling reset_work for a controller
+ * that had previously been set to the resetting state. This is intended for
+ * code paths that can't be interrupted by other reset attempts. A hot removal
+ * may prevent this from succeeding.
+ */
+static int nvme_try_sched_reset(struct nvme_ctrl *ctrl)
+{
+ if (ctrl->state != NVME_CTRL_RESETTING)
+ return -EBUSY;
+ if (!queue_work(nvme_reset_wq, &ctrl->reset_work))
+ return -EBUSY;
+ return 0;
+}
+
int nvme_reset_ctrl(struct nvme_ctrl *ctrl)
{
if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_RESETTING))
@@ -3687,13 +3702,13 @@ static void nvme_fw_act_work(struct work_struct *work)
if (time_after(jiffies, fw_act_timeout)) {
dev_warn(ctrl->device,
"Fw activation timeout, reset controller\n");
- nvme_reset_ctrl(ctrl);
- break;
+ nvme_try_sched_reset(ctrl);
+ return;
}
msleep(100);
}
- if (ctrl->state != NVME_CTRL_LIVE)
+ if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_LIVE))
return;
nvme_start_queues(ctrl);
@@ -3713,7 +3728,13 @@ static void nvme_handle_aen_notice(struct nvme_ctrl *ctrl, u32 result)
nvme_queue_scan(ctrl);
break;
case NVME_AER_NOTICE_FW_ACT_STARTING:
- queue_work(nvme_wq, &ctrl->fw_act_work);
+ /*
+ * We are (ab)using the RESETTING state to prevent subsequent
+ * recovery actions from interfering with the controller's
+ * firmware activation.
+ */
+ if (nvme_change_ctrl_state(ctrl, NVME_CTRL_RESETTING))
+ queue_work(nvme_wq, &ctrl->fw_act_work);
break;
#ifdef CONFIG_NVME_MULTIPATH
case NVME_AER_NOTICE_ANA:
--
2.20.1
_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH AUTOSEL 5.3 74/99] nvmet-loop: fix possible leakage during error flow
[not found] <20191026131600.2507-1-sashal@kernel.org>
` (3 preceding siblings ...)
2019-10-26 13:15 ` [PATCH AUTOSEL 5.3 66/99] nvme: Prevent resets during paused controller state Sasha Levin
@ 2019-10-26 13:15 ` Sasha Levin
2019-10-26 13:15 ` [PATCH AUTOSEL 5.3 75/99] nvme-tcp: " Sasha Levin
5 siblings, 0 replies; 6+ messages in thread
From: Sasha Levin @ 2019-10-26 13:15 UTC (permalink / raw)
To: linux-kernel, stable
Cc: Keith Busch, Max Gurtovoy, Christoph Hellwig, linux-nvme, Sasha Levin
From: Max Gurtovoy <maxg@mellanox.com>
[ Upstream commit 5812d04c4c7455627d8722e04ab99a737cfe9713 ]
During nvme_loop_queue_rq error flow, one must call nvme_cleanup_cmd since
it's symmetric to nvme_setup_cmd.
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/nvme/target/loop.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/nvme/target/loop.c b/drivers/nvme/target/loop.c
index 0940c5024a345..7b857c3f67879 100644
--- a/drivers/nvme/target/loop.c
+++ b/drivers/nvme/target/loop.c
@@ -157,8 +157,10 @@ static blk_status_t nvme_loop_queue_rq(struct blk_mq_hw_ctx *hctx,
iod->sg_table.sgl = iod->first_sgl;
if (sg_alloc_table_chained(&iod->sg_table,
blk_rq_nr_phys_segments(req),
- iod->sg_table.sgl, SG_CHUNK_SIZE))
+ iod->sg_table.sgl, SG_CHUNK_SIZE)) {
+ nvme_cleanup_cmd(req);
return BLK_STS_RESOURCE;
+ }
iod->req.sg = iod->sg_table.sgl;
iod->req.sg_cnt = blk_rq_map_sg(req->q, req, iod->sg_table.sgl);
--
2.20.1
_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH AUTOSEL 5.3 75/99] nvme-tcp: fix possible leakage during error flow
[not found] <20191026131600.2507-1-sashal@kernel.org>
` (4 preceding siblings ...)
2019-10-26 13:15 ` [PATCH AUTOSEL 5.3 74/99] nvmet-loop: fix possible leakage during error flow Sasha Levin
@ 2019-10-26 13:15 ` Sasha Levin
5 siblings, 0 replies; 6+ messages in thread
From: Sasha Levin @ 2019-10-26 13:15 UTC (permalink / raw)
To: linux-kernel, stable
Cc: Keith Busch, Max Gurtovoy, Christoph Hellwig, linux-nvme, Sasha Levin
From: Max Gurtovoy <maxg@mellanox.com>
[ Upstream commit 28a4cac48c7e897a0b4e7d79a53a8e4fe40337ae ]
During nvme_tcp_setup_cmd_pdu error flow, one must call nvme_cleanup_cmd
since it's symmetric to nvme_setup_cmd.
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/nvme/host/tcp.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c
index 606b13d35d16f..988a426169036 100644
--- a/drivers/nvme/host/tcp.c
+++ b/drivers/nvme/host/tcp.c
@@ -2093,6 +2093,7 @@ static blk_status_t nvme_tcp_setup_cmd_pdu(struct nvme_ns *ns,
ret = nvme_tcp_map_data(queue, rq);
if (unlikely(ret)) {
+ nvme_cleanup_cmd(rq);
dev_err(queue->ctrl->ctrl.device,
"Failed to map data (%d)\n", ret);
return ret;
--
2.20.1
_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme
^ permalink raw reply related [flat|nested] 6+ messages in thread