packages/spdk
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

!34 Rebase to spdk v21.01.1 Maintenance Release

Browse Source 
From: @swf504
Reviewed-by: @liuzhiqiang26
Signed-off-by: @liuzhiqiang26
This commit is contained in:
openeuler-ci-bot 2021-11-24 08:41:22 +00:00 committed by Gitee
commit ef270196f5
31 changed files with 17 additions and 1016 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

0
0018-blobstore-fix-memleak-problem-in-blob_load_cpl.patch → 0001-blobstore-fix-memleak-problem-in-blob_load_cpl.patch
View File

65
0001-lib-env_dpdk-fix-the-enum-rte_kernel_driver-definiti.patch
View File

@ -1,65 +0,0 @@
From 5932cd76b357acbfcd9bfa56f4e7f5b3225815e9 Mon Sep 17 00:00:00 2001
From: sunshihao520 <sunshihao@huawei.com>
Date: Thu, 4 Feb 2021 14:52:43 +0800
Subject: [PATCH 01/15] lib:env_dpdk fix the enum rte_kernel_driver definition
deference between dpdk 19.11 and 20.11
In dpdk 19.11, rte_kernel_driver is the old version, add version check before use the members.
Signed-off-by: sunshihao <sunshihao@huawei.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6252 (master)
(cherry picked from commit b1687cd45680d1573be9f66e748014fd9c5e9e4e)
Change-Id: Ic1db37cc0760c7d03692fd2cdcbb6ff1e41f872d
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6321
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: sunshihao <sunshihao@huawei.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
---
lib/env_dpdk/memory.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/lib/env_dpdk/memory.c b/lib/env_dpdk/memory.c
index f49dcf4..c923483 100644
--- a/lib/env_dpdk/memory.c
+++ b/lib/env_dpdk/memory.c
@@ -1238,7 +1238,11 @@ vtophys_iommu_device_event(const char *device_name,
RTE_DEV_FOREACH(dev, "bus=pci", &dev_iter) {
if (strcmp(dev->name, device_name) == 0) {
struct rte_pci_device *pci_dev = RTE_DEV_TO_PCI(dev);
+#if RTE_VERSION < RTE_VERSION_NUM(20, 11, 0, 0)
+ if (pci_dev->kdrv == RTE_KDRV_VFIO) {
+#else
if (pci_dev->kdrv == RTE_PCI_KDRV_VFIO) {
+#endif
/* This is a new PCI device using vfio */
g_vfio.device_ref++;
}
@@ -1267,7 +1271,11 @@ vtophys_iommu_device_event(const char *device_name,
RTE_DEV_FOREACH(dev, "bus=pci", &dev_iter) {
if (strcmp(dev->name, device_name) == 0) {
struct rte_pci_device *pci_dev = RTE_DEV_TO_PCI(dev);
+#if RTE_VERSION < RTE_VERSION_NUM(20, 11, 0, 0)
+ if (pci_dev->kdrv == RTE_KDRV_VFIO) {
+#else
if (pci_dev->kdrv == RTE_PCI_KDRV_VFIO) {
+#endif
/* This is a PCI device using vfio */
g_vfio.device_ref--;
}
@@ -1362,7 +1370,11 @@ vtophys_iommu_init(void)
RTE_DEV_FOREACH(dev, "bus=pci", &dev_iter) {
struct rte_pci_device *pci_dev = RTE_DEV_TO_PCI(dev);
+#if RTE_VERSION < RTE_VERSION_NUM(20, 11, 0, 0)
+ if (pci_dev->kdrv == RTE_KDRV_VFIO) {
+#else
if (pci_dev->kdrv == RTE_PCI_KDRV_VFIO) {
+#endif
/* This is a PCI device using vfio */
g_vfio.device_ref++;
}
--
1.8.3.1

0
0019-blobstore-fix-potential-memleak-problem-in-blob_seri.patch → 0002-blobstore-fix-potential-memleak-problem-in-blob_seri.patch
View File

40
0002-env_dpdk-add-rte_ethdev-dependency.patch
View File

@ -1,40 +0,0 @@
From bf8f81bc254642110b009e802e479be1ea149423 Mon Sep 17 00:00:00 2001
From: Jim Harris <james.r.harris@intel.com>
Date: Mon, 1 Feb 2021 14:28:30 +0000
Subject: [PATCH 02/15] env_dpdk: add rte_ethdev dependency
As of DPDK main commit 682a6454 ("power: add ethdev
power management"), rte_power depends on rte_ethdev.
So add rte_ethdev library, but only on Linux since
that is the only place where we use rte_power.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6210 (master)
(cherry picked from commit 710838103bb8ba6e9938dd6aa66c72c6ecb3e928)
Change-Id: Ibc3473d6ec75c14a8c0c5517616950017cb2ea96
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6320
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
---
lib/env_dpdk/env.mk | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/env_dpdk/env.mk b/lib/env_dpdk/env.mk
index 6dae04e..f44f515 100644
--- a/lib/env_dpdk/env.mk
+++ b/lib/env_dpdk/env.mk
@@ -51,7 +51,7 @@ DPDK_INC := -I$(DPDK_INC_DIR)
DPDK_LIB_LIST = rte_eal rte_mempool rte_ring rte_mbuf rte_pci rte_bus_pci rte_mempool_ring
ifeq ($(OS),Linux)
-DPDK_LIB_LIST += rte_power
+DPDK_LIB_LIST += rte_power rte_ethdev
endif
# DPDK 20.05 eal dependency
--
1.8.3.1

0
0020-idxd-fix-memleak-problem-in-spdk_idxd_configure_chan.patch → 0003-idxd-fix-memleak-problem-in-spdk_idxd_configure_chan.patch
View File

38
0003-pkg-spdk.spec-Add-ncurses-devel-to-BuildRequires.patch
View File

@ -1,38 +0,0 @@
From fa2ceb127060000d26758db36e5e13c4c7d97f77 Mon Sep 17 00:00:00 2001
From: sunshihao520 <sunshihao@huawei.com>
Date: Thu, 4 Feb 2021 19:49:03 +0800
Subject: [PATCH 03/15] pkg:spdk.spec Add ncurses-devel to BuildRequires
ncurses.h is needed in spdk_top, add ncurses-devel to BuildRequires.
Signed-off-by: sunshihao <sunshihao@huawei.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6258 (master)
(cherry picked from commit ccbdd1750f3da9be85b2018896fab5c88cb7f686)
Change-Id: I22754838788fe237a29ab74a675a94ab7007bd33
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6388
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: sunshihao <sunshihao@huawei.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
---
pkg/spdk.spec | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/pkg/spdk.spec b/pkg/spdk.spec
index 2237635..23de4f8 100644
--- a/pkg/spdk.spec
+++ b/pkg/spdk.spec
@@ -22,7 +22,7 @@ License: BSD
ExclusiveArch: x86_64
BuildRequires: gcc gcc-c++ make
-BuildRequires: dpdk-devel, numactl-devel
+BuildRequires: dpdk-devel, numactl-devel, ncurses-devel
BuildRequires: libiscsi-devel, libaio-devel, openssl-devel, libuuid-devel
BuildRequires: libibverbs-devel, librdmacm-devel
%if %{with doc}
--
1.8.3.1

0
0021-idxd-fix-one-memleak-problem-in-spdk_idxd_get_channe.patch → 0004-idxd-fix-one-memleak-problem-in-spdk_idxd_get_channe.patch
View File

51
0004-lib-vhost-Add-version-check-when-use-RTE_VHOST_USER_.patch
View File

@ -1,51 +0,0 @@
From 0c08337dd060fd42390c6fab26f5a08d0d6269bf Mon Sep 17 00:00:00 2001
From: sunshihao520 <sunshihao@huawei.com>
Date: Tue, 9 Feb 2021 15:17:24 +0800
Subject: [PATCH 04/15] lib:vhost Add version check when use
RTE_VHOST_USER_ASYNC_COPY in 21.01
In dpdk 19.11 version, RTE_VHOST_USER_ASYNC_COPY is not define.
After dpdk 20.08, we can use RTE_VHOST_USER_ASYNC_COPY.
Use version check to avoid this problem.
Signed-off-by: sunshihao <sunshihao@huawei.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6274 (master)
(cherry picked from commit 9529803fe0f32d69752cf01eeed590056bd43637)
Change-Id: Iaf9914e8380f3d54cded1e2f16af6a7dc3504f95
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6255
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
---
lib/vhost/rte_vhost_compat.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/lib/vhost/rte_vhost_compat.c b/lib/vhost/rte_vhost_compat.c
index 85cce3d..3c9f691 100644
--- a/lib/vhost/rte_vhost_compat.c
+++ b/lib/vhost/rte_vhost_compat.c
@@ -46,6 +46,7 @@
#include "spdk/barrier.h"
#include "spdk/vhost.h"
#include "vhost_internal.h"
+#include <rte_version.h>
#include "spdk_internal/vhost_user.h"
@@ -331,7 +332,11 @@ vhost_register_unix_socket(const char *path, const char *ctrl_name,
}
}
+#if RTE_VERSION < RTE_VERSION_NUM(20, 8, 0, 0)
+ if (rte_vhost_driver_register(path, 0) != 0) {
+#else
if (rte_vhost_driver_register(path, RTE_VHOST_USER_ASYNC_COPY) != 0) {
+#endif
SPDK_ERRLOG("Could not register controller %s with vhost library\n", ctrl_name);
SPDK_ERRLOG("Check if domain socket %s already exists\n", path);
return -EIO;
--
1.8.3.1

0
0022-ioat-fix-potential-double-free-problem-in-ioat_chann.patch → 0005-ioat-fix-potential-double-free-problem-in-ioat_chann.patch
View File

89
0005-lib-nvme-Remove-qpair-from-all-lists-before-freeing-.patch
View File

@ -1,89 +0,0 @@
From d889786d0044daff1344ed919bc7bab7d88da6ab Mon Sep 17 00:00:00 2001
From: Vasuki Manikarnike <vasuki.manikarnike@hpe.com>
Date: Thu, 4 Feb 2021 15:54:37 +0000
Subject: [PATCH 05/15] lib/nvme: Remove qpair from all lists before freeing
it.
Fixes #1777.
When a qpair cannot be allocated because the transport connection fails,
the qpair was freed without unlinking it from the other structures.
This was leading to a segfault when attempting to create and free other
qpairs.
Also added a unit test to cover this case.
Signed-off-by: Vasuki Manikarnike <vasuki.manikarnike@hpe.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6272 (master)
(cherry picked from commit d92c2f118aa9ccbd242538a01b2ce0e2cf0c6b04)
Change-Id: I74b78d1847f90117248b07203b43a11ff5cfa5d6
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6393
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
---
lib/nvme/nvme_ctrlr.c | 4 ++++
test/unit/lib/nvme/nvme_ctrlr.c/nvme_ctrlr_ut.c | 22 ++++++++++++++++++++++
2 files changed, 26 insertions(+)
diff --git a/lib/nvme/nvme_ctrlr.c b/lib/nvme/nvme_ctrlr.c
index 1960b0a..ff27771 100644
--- a/lib/nvme/nvme_ctrlr.c
+++ b/lib/nvme/nvme_ctrlr.c
@@ -418,8 +418,12 @@ spdk_nvme_ctrlr_alloc_io_qpair(struct spdk_nvme_ctrlr *ctrlr,
rc = spdk_nvme_ctrlr_connect_io_qpair(ctrlr, qpair);
if (rc != 0) {
SPDK_ERRLOG("nvme_transport_ctrlr_connect_io_qpair() failed\n");
+ nvme_robust_mutex_lock(&ctrlr->ctrlr_lock);
+ nvme_ctrlr_proc_remove_io_qpair(qpair);
TAILQ_REMOVE(&ctrlr->active_io_qpairs, qpair, tailq);
+ spdk_bit_array_set(ctrlr->free_io_qids, qpair->id);
nvme_transport_ctrlr_delete_io_qpair(ctrlr, qpair);
+ nvme_robust_mutex_unlock(&ctrlr->ctrlr_lock);
return NULL;
}
diff --git a/test/unit/lib/nvme/nvme_ctrlr.c/nvme_ctrlr_ut.c b/test/unit/lib/nvme/nvme_ctrlr.c/nvme_ctrlr_ut.c
index 3e2321d..16a4c1b 100644
--- a/test/unit/lib/nvme/nvme_ctrlr.c/nvme_ctrlr_ut.c
+++ b/test/unit/lib/nvme/nvme_ctrlr.c/nvme_ctrlr_ut.c
@@ -2162,6 +2162,27 @@ test_nvme_ctrlr_init_set_keep_alive_timeout(void)
nvme_ctrlr_destruct(&ctrlr);
}
+static void
+test_alloc_io_qpair_fail(void)
+{
+ struct spdk_nvme_ctrlr ctrlr = {};
+ struct spdk_nvme_qpair *q0;
+
+ setup_qpairs(&ctrlr, 1);
+
+ /* Modify the connect_qpair return code to inject a failure */
+ g_connect_qpair_return_code = 1;
+
+ /* Attempt to allocate a qpair, this should fail */
+ q0 = spdk_nvme_ctrlr_alloc_io_qpair(&ctrlr, NULL, 0);
+ SPDK_CU_ASSERT_FATAL(q0 == NULL);
+
+ /* Verify that the qpair is removed from the lists */
+ SPDK_CU_ASSERT_FATAL(TAILQ_EMPTY(&ctrlr.active_io_qpairs));
+
+ cleanup_qpairs(&ctrlr);
+}
+
int main(int argc, char **argv)
{
CU_pSuite suite = NULL;
@@ -2200,6 +2221,7 @@ int main(int argc, char **argv)
CU_ADD_TEST(suite, test_nvme_ctrlr_init_set_nvmf_ioccsz);
CU_ADD_TEST(suite, test_nvme_ctrlr_init_set_num_queues);
CU_ADD_TEST(suite, test_nvme_ctrlr_init_set_keep_alive_timeout);
+ CU_ADD_TEST(suite, test_alloc_io_qpair_fail);
CU_basic_set_mode(CU_BRM_VERBOSE);
CU_basic_run_tests();
--
1.8.3.1

50
0006-lib-env_dpdk-add-rte_net-dependency.patch
View File

@ -1,50 +0,0 @@
From f8acbbb0ac154fa812e6297e86fad58381d4a163 Mon Sep 17 00:00:00 2001
From: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Date: Fri, 12 Feb 2021 03:45:16 -0500
Subject: [PATCH 06/15] lib/env_dpdk: add rte_net dependency
Vhost is enabled by default, so rte_net was always included.
When disabled, rte_power failed as it depends on rte_ethdev and rte_net.
rte_vhost was only possible to enable on Linux, so there
is no conflict with adding it next to rte_power under this condition.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6398 (master)
(cherry picked from commit 232f34d754e08b9673fd3585786ca8cd55653b85)
Change-Id: I2e183004d6457e404471740a0540dcb08aa738d8
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6430
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
---
lib/env_dpdk/env.mk | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/env_dpdk/env.mk b/lib/env_dpdk/env.mk
index f44f515..292dd91 100644
--- a/lib/env_dpdk/env.mk
+++ b/lib/env_dpdk/env.mk
@@ -51,7 +51,7 @@ DPDK_INC := -I$(DPDK_INC_DIR)
DPDK_LIB_LIST = rte_eal rte_mempool rte_ring rte_mbuf rte_pci rte_bus_pci rte_mempool_ring
ifeq ($(OS),Linux)
-DPDK_LIB_LIST += rte_power rte_ethdev
+DPDK_LIB_LIST += rte_power rte_ethdev rte_net
endif
# DPDK 20.05 eal dependency
@@ -101,7 +101,7 @@ endif
LINK_HASH=n
ifeq ($(CONFIG_VHOST),y)
-DPDK_LIB_LIST += rte_vhost rte_net
+DPDK_LIB_LIST += rte_vhost
LINK_HASH=y
ifneq ($(DPDK_FRAMEWORK),y)
DPDK_LIB_LIST += rte_cryptodev
--
1.8.3.1

0
0023-nvmf-check-return-value-of-strdup-in-spdk_nvmf_subsy.patch → 0006-nvmf-check-return-value-of-strdup-in-spdk_nvmf_subsy.patch
View File

0
0024-nvmf-check-return-value-of-strdup-in-spdk_nvmf_subsy.patch → 0007-nvmf-check-return-value-of-strdup-in-spdk_nvmf_subsy.patch
View File

38
0007-pkg-add-python3-requires-in-spdk.spec.patch
View File

@ -1,38 +0,0 @@
From 4eca99edc2e23b4d6250907b62e8feda66414fd9 Mon Sep 17 00:00:00 2001
From: sunshihao520 <sunshihao@huawei.com>
Date: Tue, 9 Feb 2021 15:45:30 +0800
Subject: [PATCH 07/15] pkg: add python3 requires in spdk.spec
When use rpm -i to install spdk-tools, it requires python3-configshell
and python3-pexpect. Add this requires to spec to avoid install failure.
Signed-off-by: sunshihao <sunshihao@huawei.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6336 (master)
(cherry picked from commit a4ee18886d728de55070fd8c08f947be4004c207)
Change-Id: I7fc7e9aab35c3fd730401282b72b1604f765a73f
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6431
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: sunshihao <sunshihao@huawei.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
---
pkg/spdk.spec | 1 +
1 file changed, 1 insertion(+)
diff --git a/pkg/spdk.spec b/pkg/spdk.spec
index 23de4f8..0fc46e7 100644
--- a/pkg/spdk.spec
+++ b/pkg/spdk.spec
@@ -32,6 +32,7 @@ BuildRequires: doxygen mscgen graphviz
# Install dependencies
Requires: dpdk >= 19.11, numactl-libs, openssl-libs
Requires: libiscsi, libaio, libuuid
+Requires: python3-configshell, python3-pexpect
# NVMe over Fabrics
Requires: librdmacm, librdmacm
Requires(post): /sbin/ldconfig
--
1.8.3.1

0
0025-nvmf-fix-fd-leakage-problem-in-nvmf_vfio_user_listen.patch → 0008-nvmf-fix-fd-leakage-problem-in-nvmf_vfio_user_listen.patch
View File

36
0008-sock-add-enable_quickack-and-enable_placement_id-whe.patch
View File

@ -1,36 +0,0 @@
From 830f0bab074f397bd13eec39fbef63b550a302b2 Mon Sep 17 00:00:00 2001
From: Richael Zhuang <richael.zhuang@arm.com>
Date: Thu, 28 Jan 2021 17:19:15 +0800
Subject: [PATCH 08/15] sock: add enable_quickack and enable_placement_id when
writing json config
Signed-off-by: Richael Zhuang <richael.zhuang@arm.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6144 (master)
(cherry picked from commit c0a3cf7b88851cf51dff8145546dd3c7e2741c9d)
Change-Id: If67dcb6bf7b9742d2c96481057b472e31e87e87d
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6392
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
---
lib/sock/sock.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/lib/sock/sock.c b/lib/sock/sock.c
index ef4678f..8aa2817 100644
--- a/lib/sock/sock.c
+++ b/lib/sock/sock.c
@@ -832,6 +832,8 @@ spdk_sock_write_config_json(struct spdk_json_write_ctx *w)
spdk_json_write_named_uint32(w, "send_buf_size", opts.send_buf_size);
spdk_json_write_named_bool(w, "enable_recv_pipe", opts.enable_recv_pipe);
spdk_json_write_named_bool(w, "enable_zerocopy_send", opts.enable_zerocopy_send);
+ spdk_json_write_named_bool(w, "enable_quickack", opts.enable_quickack);
+ spdk_json_write_named_bool(w, "enable_placement_id", opts.enable_placement_id);
spdk_json_write_object_end(w);
spdk_json_write_object_end(w);
} else {
--
1.8.3.1

41
0009-bdev-ocssd-Fix-the-bug-that-no-media-event-is-pushed.patch
View File

@ -1,41 +0,0 @@
From 66ce3d8005f9a37f55c36146bb73476c6f6f1d22 Mon Sep 17 00:00:00 2001
From: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Date: Wed, 17 Feb 2021 20:57:38 +0900
Subject: [PATCH 09/15] bdev/ocssd: Fix the bug that no media event is pushed
to the target bdev
Recent refactoring added a critical bug that no media event is pushed
to the target bdev.
Fix the bug by changing return to break.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6455 (master)
(cherry picked from commit bd8de45801cc0b45dedddef6534cd35493163b29)
Change-Id: Ia674ea2c9c21d08dd23b50a0f726da55011d4be4
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6616
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
---
module/bdev/nvme/bdev_ocssd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/module/bdev/nvme/bdev_ocssd.c b/module/bdev/nvme/bdev_ocssd.c
index 1cf1705..748b454 100644
--- a/module/bdev/nvme/bdev_ocssd.c
+++ b/module/bdev/nvme/bdev_ocssd.c
@@ -967,7 +967,7 @@ bdev_ocssd_push_media_events(struct nvme_bdev_ns *nvme_ns,
TAILQ_FOREACH(nvme_bdev, &nvme_ns->bdevs, tailq) {
ocssd_bdev = SPDK_CONTAINEROF(nvme_bdev, struct ocssd_bdev, nvme_bdev);
if (bdev_ocssd_lba_in_range(ocssd_bdev, ocssd_ns, chunk_entry->lba)) {
- return;
+ break;
}
}
--
1.8.3.1

0
0026-posix-set-fd-to-1-after-close-fd-in-posix_sock_creat.patch → 0009-posix-set-fd-to-1-after-close-fd-in-posix_sock_creat.patch
View File

128
0010-lib-iscsi-return-immediately-from-iscsi_parse_params.patch
View File

@ -1,128 +0,0 @@
From c2c8b7d986c86ff2ec031873d3b08e7e5eae22ea Mon Sep 17 00:00:00 2001
From: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Date: Mon, 8 Feb 2021 10:28:44 -0500
Subject: [PATCH 10/15] lib/iscsi: return immediately from iscsi_parse_params
if len is 0
The spec does not disallow TEXT PDUs with no data. In that
case, just return immediately from iscsi_parse_params.
This avoids a NULL pointer dereference with a TEXT PDU that has
no data, but CONTINUE flag is set.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6319 (master)
(cherry picked from commit f3fd56fc3c73ee7595bbe7c69cf8048fe2577e1a)
Change-Id: I2605293daf171633a45132d7b5532fdfc9128aff
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6601
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
---
lib/iscsi/param.c | 10 +++++++++
test/unit/lib/iscsi/iscsi.c/iscsi_ut.c | 38 ++++++++++++++++++++++++++++++++++
test/unit/lib/iscsi/param.c/param_ut.c | 8 +++++++
3 files changed, 56 insertions(+)
diff --git a/lib/iscsi/param.c b/lib/iscsi/param.c
index 1a353c4..ab2c9cd 100644
--- a/lib/iscsi/param.c
+++ b/lib/iscsi/param.c
@@ -315,6 +315,16 @@ iscsi_parse_params(struct iscsi_param **params, const uint8_t *data,
char *p;
int i;
+ /* Spec does not disallow TEXT PDUs with zero length, just return
+ * immediately in that case, since there is no param data to parse
+ * and any existing partial parameter would remain as-is.
+ */
+ if (len == 0) {
+ return 0;
+ }
+
+ assert(data != NULL);
+
/* strip the partial text parameters if previous PDU have C enabled */
if (partial_parameter && *partial_parameter) {
for (i = 0; i < len && data[i] != '\0'; i++) {
diff --git a/test/unit/lib/iscsi/iscsi.c/iscsi_ut.c b/test/unit/lib/iscsi/iscsi.c/iscsi_ut.c
index c901e9e..56082a0 100644
--- a/test/unit/lib/iscsi/iscsi.c/iscsi_ut.c
+++ b/test/unit/lib/iscsi/iscsi.c/iscsi_ut.c
@@ -1993,6 +1993,43 @@ pdu_hdr_op_data_test(void)
g_task_pool_is_empty = false;
}
+/* Test an ISCSI_OP_TEXT PDU with CONTINUE bit set but
+ * no data.
+ */
+static void
+empty_text_with_cbit_test(void)
+{
+ struct spdk_iscsi_sess sess = {};
+ struct spdk_iscsi_conn conn = {};
+ struct spdk_scsi_dev dev = {};
+ struct spdk_iscsi_pdu *req_pdu;
+ int rc;
+
+ req_pdu = iscsi_get_pdu(&conn);
+
+ sess.ExpCmdSN = 0;
+ sess.MaxCmdSN = 64;
+ sess.session_type = SESSION_TYPE_NORMAL;
+ sess.MaxBurstLength = 1024;
+
+ conn.full_feature = 1;
+ conn.sess = &sess;
+ conn.dev = &dev;
+ conn.state = ISCSI_CONN_STATE_RUNNING;
+
+ memset(&req_pdu->bhs, 0, sizeof(req_pdu->bhs));
+ req_pdu->bhs.opcode = ISCSI_OP_TEXT;
+ req_pdu->bhs.flags = ISCSI_TEXT_CONTINUE;
+
+ rc = iscsi_pdu_hdr_handle(&conn, req_pdu);
+ CU_ASSERT(rc == 0);
+ CU_ASSERT(!req_pdu->is_rejected);
+ rc = iscsi_pdu_payload_handle(&conn, req_pdu);
+ CU_ASSERT(rc == 0);
+
+ iscsi_put_pdu(req_pdu);
+}
+
int
main(int argc, char **argv)
{
@@ -2024,6 +2061,7 @@ main(int argc, char **argv)
CU_ADD_TEST(suite, pdu_hdr_op_task_mgmt_test);
CU_ADD_TEST(suite, pdu_hdr_op_nopout_test);
CU_ADD_TEST(suite, pdu_hdr_op_data_test);
+ CU_ADD_TEST(suite, empty_text_with_cbit_test);
CU_basic_set_mode(CU_BRM_VERBOSE);
CU_basic_run_tests();
diff --git a/test/unit/lib/iscsi/param.c/param_ut.c b/test/unit/lib/iscsi/param.c/param_ut.c
index ccf6264..9b3e820 100644
--- a/test/unit/lib/iscsi/param.c/param_ut.c
+++ b/test/unit/lib/iscsi/param.c/param_ut.c
@@ -264,6 +264,14 @@ parse_valid_test(void)
EXPECT_VAL("F", "IIII");
CU_ASSERT_PTR_NULL(partial_parameter);
+ /* partial parameter: NULL data */
+ /* It is technically allowed to have a TEXT PDU with no data, yet
+ * CONTINUE bit is enabled - make sure we handle that case correctly.
+ */
+ rc = iscsi_parse_params(&params, NULL, 0, true, &partial_parameter);
+ CU_ASSERT(rc == 0);
+ CU_ASSERT_PTR_NULL(partial_parameter);
+
/* Second partial parameter is the only parameter */
PARSE("OOOO", true, &partial_parameter);
CU_ASSERT_STRING_EQUAL(partial_parameter, "OOOO");
--
1.8.3.1

0
0027-spdk_top-check-return-value-of-strdup-in-store_last_.patch → 0010-spdk_top-check-return-value-of-strdup-in-store_last_.patch
View File

57
0011-nbd-set-io-timeout.patch
View File

@ -1,57 +0,0 @@
From b7a8bc02e3530adc92f5287285330d926f999b5a Mon Sep 17 00:00:00 2001
From: Liu Xiaodong <xiaodong.liu@intel.com>
Date: Mon, 1 Feb 2021 08:19:14 -0500
Subject: [PATCH 11/15] nbd: set io timeout
After setting io timeout, host can avoid nbd io
stuck or kernel hang occasionally caused by nbd
stop or underlying bdev removal.
Signed-off-by: Liu Xiaodong <xiaodong.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6191 (master)
(cherry picked from commit b7c33b5e610fd8894a0d4a67894f71c750f357d1)
Change-Id: I4ba2a0af7ff7bed369cdaf86121f082136dc1a0b
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6394
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: <dongx.yi@intel.com>
---
lib/nbd/nbd.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/lib/nbd/nbd.c b/lib/nbd/nbd.c
index 2da9066..3a9299d 100644
--- a/lib/nbd/nbd.c
+++ b/lib/nbd/nbd.c
@@ -50,6 +50,7 @@
#define GET_IO_LOOP_COUNT 16
#define NBD_BUSY_WAITING_MS 1000
#define NBD_BUSY_POLLING_INTERVAL_US 20000
+#define NBD_IO_TIMEOUT_S 60
enum nbd_io_state_t {
/* Receiving or ready to receive nbd request header */
@@ -922,6 +923,17 @@ nbd_start_complete(struct spdk_nbd_start_ctx *ctx)
goto err;
}
+#ifdef NBD_SET_TIMEOUT
+ rc = ioctl(ctx->nbd->dev_fd, NBD_SET_TIMEOUT, NBD_IO_TIMEOUT_S);
+ if (rc == -1) {
+ SPDK_ERRLOG("ioctl(NBD_SET_TIMEOUT) failed: %s\n", spdk_strerror(errno));
+ rc = -errno;
+ goto err;
+ }
+#else
+ SPDK_NOTICELOG("ioctl(NBD_SET_TIMEOUT) is not supported.\n");
+#endif
+
#ifdef NBD_FLAG_SEND_TRIM
rc = ioctl(ctx->nbd->dev_fd, NBD_SET_FLAGS, NBD_FLAG_SEND_TRIM);
if (rc == -1) {
--
1.8.3.1

0
0028-uring-set-fd-to-1-after-close-fd-in-uring_sock_creat.patch → 0011-uring-set-fd-to-1-after-close-fd-in-uring_sock_creat.patch
View File

39
0012-lib-util-Fix-valgrind-error-reported-on-ARM-platform.patch
View File

@ -1,39 +0,0 @@
From 0765bbc3fbfa06a511eaf931468e813dfd787697 Mon Sep 17 00:00:00 2001
From: Rui Chang <rui.chang@arm.com>
Date: Tue, 2 Feb 2021 10:18:47 +0800
Subject: [PATCH 12/15] lib/util: Fix valgrind error reported on ARM platform
Signed-off-by: Rui Chang <rui.chang@arm.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6214 (master)
(cherry picked from commit 1293377629ad9ae95373178c919dcbe7d0fc2a95)
Change-Id: I244a1b0350b4426666e390781b500de65d250c09
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6391
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Rui Chang <rui.chang@arm.com>
---
lib/util/fd_group.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/util/fd_group.c b/lib/util/fd_group.c
index d3f9104..217ca43 100644
--- a/lib/util/fd_group.c
+++ b/lib/util/fd_group.c
@@ -85,8 +85,8 @@ int
spdk_fd_group_add(struct spdk_fd_group *fgrp,
int efd, spdk_fd_fn fn, void *arg)
{
- struct event_handler *ehdlr;
- struct epoll_event epevent;
+ struct event_handler *ehdlr = NULL;
+ struct epoll_event epevent = {0};
int rc;
/* parameter checking */
--
1.8.3.1

104
0013-lib-vhost-force-cpumask-to-be-subset-of-application-.patch
View File

@ -1,104 +0,0 @@
From c5bf8a0423325b96d7ede271965c117cd3790cde Mon Sep 17 00:00:00 2001
From: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Date: Mon, 1 Feb 2021 07:32:46 -0500
Subject: [PATCH 13/15] lib/vhost: force cpumask to be subset of application's
CPU mask
Documentation for vhost target states that CPU mask must be a subset
of application CPU mask. This wasn't enforced right now and allowed
the cpumask on controller creation so long as at least single
CPU core overlapped with application's CPU mask.
This might have been misleading and covered up user configuration errors.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6212 (master)
(cherry picked from commit b74363bcdfb6c0925dd01a24574c4fc82b22e58f)
Change-Id: I03f959ec37efd0be9b98cff9c93c5f996b04af35
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6390
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
---
lib/vhost/vhost.c | 11 +++++++++++
test/unit/lib/vhost/vhost.c/vhost_ut.c | 4 ++++
test/vhost/other/negative.sh | 10 ++++++++++
3 files changed, 25 insertions(+)
diff --git a/lib/vhost/vhost.c b/lib/vhost/vhost.c
index b3aee72..18c17b3 100644
--- a/lib/vhost/vhost.c
+++ b/lib/vhost/vhost.c
@@ -888,6 +888,7 @@ static int
vhost_parse_core_mask(const char *mask, struct spdk_cpuset *cpumask)
{
int rc;
+ struct spdk_cpuset negative_vhost_mask;
if (cpumask == NULL) {
return -1;
@@ -904,6 +905,16 @@ vhost_parse_core_mask(const char *mask, struct spdk_cpuset *cpumask)
return -1;
}
+ spdk_cpuset_copy(&negative_vhost_mask, &g_vhost_core_mask);
+ spdk_cpuset_negate(&negative_vhost_mask);
+ spdk_cpuset_and(&negative_vhost_mask, cpumask);
+
+ if (spdk_cpuset_count(&negative_vhost_mask) != 0) {
+ SPDK_ERRLOG("one of selected cpu is outside of core mask(=%s)\n",
+ spdk_cpuset_fmt(&g_vhost_core_mask));
+ return -1;
+ }
+
spdk_cpuset_and(cpumask, &g_vhost_core_mask);
if (spdk_cpuset_count(cpumask) == 0) {
diff --git a/test/unit/lib/vhost/vhost.c/vhost_ut.c b/test/unit/lib/vhost/vhost.c/vhost_ut.c
index 79a6d20..45ce0a4 100644
--- a/test/unit/lib/vhost/vhost.c/vhost_ut.c
+++ b/test/unit/lib/vhost/vhost.c/vhost_ut.c
@@ -269,6 +269,10 @@ create_controller_test(void)
spdk_cpuset_set_cpu(&g_vhost_core_mask, 0, true);
+ /* Create device with incorrect cpumask partially outside of application cpumask */
+ ret = alloc_vdev(&vdev, "vdev_name_0", "0xff");
+ SPDK_CU_ASSERT_FATAL(ret != 0);
+
/* Create device with no name */
ret = alloc_vdev(&vdev, NULL, "0x1");
CU_ASSERT(ret != 0);
diff --git a/test/vhost/other/negative.sh b/test/vhost/other/negative.sh
index ab7fa9d..2c219e7 100755
--- a/test/vhost/other/negative.sh
+++ b/test/vhost/other/negative.sh
@@ -93,6 +93,11 @@ if $rpc_py vhost_create_scsi_controller vhost.invalid.cpumask --cpumask 0x2; the
error "Creating scsi controller with incorrect cpumask succeeded, but it shouldn't"
fi
+notice "Trying to create scsi controller with incorrect cpumask partially outside of application cpumask"
+if $rpc_py vhost_create_scsi_controller vhost.invalid.cpumask --cpumask 0xff; then
+ error "Creating scsi controller with incorrect cpumask succeeded, but it shouldn't"
+fi
+
notice "Trying to remove device from nonexistent scsi controller"
if $rpc_py vhost_scsi_controller_remove_target vhost.nonexistent.name 0; then
error "Removing device from nonexistent scsi controller succeeded, but it shouldn't"
@@ -172,6 +177,11 @@ if $rpc_py vhost_create_blk_controller vhost.invalid.cpumask Malloc0 --cpumask 0
error "Creating block controller with incorrect cpumask succeeded, but it shouldn't"
fi
+notice "Trying to create block controller with incorrect cpumask partially outside of application cpumask"
+if $rpc_py vhost_create_blk_controller vhost.invalid.cpumask Malloc0 --cpumask 0xff; then
+ error "Creating block controller with incorrect cpumask succeeded, but it shouldn't"
+fi
+
notice "Trying to remove nonexistent block controller"
if $rpc_py vhost_delete_controller vhost.nonexistent.name; then
error "Removing nonexistent block controller succeeded, but it shouldn't"
--
1.8.3.1

42
0014-autorun-allow-pass-configuration-file-path.patch
View File

@ -1,42 +0,0 @@
From 7cbc04ea01fd56bac975bf4c1a1634ff7e1a75f1 Mon Sep 17 00:00:00 2001
From: Pawel Piatek <pawelx.piatek@intel.com>
Date: Wed, 10 Feb 2021 05:47:39 +0100
Subject: [PATCH 14/15] autorun: allow pass configuration file path
Allow to pass configuration file as first parameter with
fallback to previous behaviour where config file is read
from home directory.
Signed-off-by: Pawel Piatek <pawelx.piatek@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6361 (master)
(cherry picked from commit 23c8f667a22e0a6a79099765bfdf8256497a5028)
Change-Id: I45972b04fe5fc978bf3b2bede45aae363be3069c
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6545
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Pawel Piatek <pawelx.piatek@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
---
autorun.sh | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/autorun.sh b/autorun.sh
index cdc27a3..c85fdb2 100755
--- a/autorun.sh
+++ b/autorun.sh
@@ -4,7 +4,8 @@ set -e
rootdir=$(readlink -f $(dirname $0))
-conf=~/autorun-spdk.conf
+default_conf=~/autorun-spdk.conf
+conf=${1:-${default_conf}}
# If the configuration of tests is not provided, no tests will be carried out.
if [[ ! -f $conf ]]; then
--
1.8.3.1

63
0015-spdk_top-fix-app-crashing-on-tab-selection-with-TAB-.patch
View File

@ -1,63 +0,0 @@
From d1bc0054cdebc60d469e570610590c13c2234591 Mon Sep 17 00:00:00 2001
From: Krzysztof Karas <krzysztof.karas@intel.com>
Date: Thu, 4 Feb 2021 10:30:00 +0100
Subject: [PATCH 15/15] spdk_top: fix app crashing on tab selection with TAB
key
Fixes spdk_top crashing problem upon selecting a row in one tab,
then switching with TAB key to a different tab with fewer rows
and calling pop-up details window with ENTER key.
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6367 (master)
(cherry picked from commit 6a343482193bf728d3e82bd5b1994bf767533aef)
Change-Id: I52c5729eca0ce0eaad88767ff3add471a780140a
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6729
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
---
app/spdk_top/spdk_top.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/app/spdk_top/spdk_top.c b/app/spdk_top/spdk_top.c
index c996235..402c2a5 100644
--- a/app/spdk_top/spdk_top.c
+++ b/app/spdk_top/spdk_top.c
@@ -2027,6 +2027,7 @@ show_thread(uint8_t current_page)
get_data();
+ assert(thread_number < g_threads_stats.threads.threads_count);
for (i = 0; i < g_threads_stats.threads.threads_count; i++) {
thread_info[i] = &g_threads_stats.threads.thread_info[i];
}
@@ -2067,6 +2068,7 @@ show_core(uint8_t current_page)
get_data();
+ assert(core_number < g_cores_stats.cores.cores_count);
for (i = 0; i < g_cores_stats.cores.cores_count; i++) {
core_info[i] = &g_cores_stats.cores.core[i];
}
@@ -2178,6 +2180,7 @@ show_poller(uint8_t current_page)
get_data();
prepare_poller_data(current_page, pollers, &count, current_page);
+ assert(poller_number < count);
poller_win = newwin(POLLER_WIN_HEIGHT, POLLER_WIN_WIDTH,
(g_max_row - poller_counter) / 2, (g_max_col - POLLER_WIN_HOR_POS) / 2);
@@ -2300,6 +2303,7 @@ show_stats(void)
} else {
active_tab = THREADS_TAB;
}
+ g_selected_row = 0;
current_page = 0;
switch_tab(active_tab);
break;
--
1.8.3.1

52
0016-blobfs-check-return-value-of-strdup-in-blobfs_fuse_s.patch
View File

@ -1,52 +0,0 @@
From 65f41dc6b49cb2d8b7bb9e3951f6f4fcf5c93eee Mon Sep 17 00:00:00 2001
From: Zhiqiang Liu <liuzhiqiang26@huawei.com>
Date: Sun, 13 Jun 2021 18:25:14 +0800
Subject: [PATCH 16/28] blobfs: check return value of strdup in
blobfs_fuse_start()
In blobfs_fuse_start(), bfuse->bdev_name and bfuse->mountpoint
are allocated by calling strdup(), which may return NULL.
Here, we will go to err if strdup() returns NULL.
Signed-off-by: Zhiqiang Liu <liuzhiqiang26@huawei.com>
Change-Id: I0599254b3436a310ddd26732312281f07a4972ec
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8303
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
---
module/blobfs/bdev/blobfs_fuse.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/module/blobfs/bdev/blobfs_fuse.c b/module/blobfs/bdev/blobfs_fuse.c
index 1666549..176f81e 100644
--- a/module/blobfs/bdev/blobfs_fuse.c
+++ b/module/blobfs/bdev/blobfs_fuse.c
@@ -301,15 +301,19 @@ blobfs_fuse_start(const char *bdev_name, const char *mountpoint, struct spdk_fil
return -ENOMEM;
}
- rc = fuse_parse_cmdline(&args, &opts);
- assert(rc == 0);
-
bfuse->bdev_name = strdup(bdev_name);
bfuse->mountpoint = strdup(mountpoint);
+ if (!bfuse->bdev_name || !bfuse->mountpoint) {
+ rc = -ENOMEM;
+ goto err;
+ }
bfuse->fs = fs;
bfuse->cb_fn = cb_fn;
bfuse->cb_arg = cb_arg;
+ rc = fuse_parse_cmdline(&args, &opts);
+ assert(rc == 0);
+
fuse_handle = fuse_new(&args, &spdk_fuse_oper, sizeof(spdk_fuse_oper), NULL);
fuse_opt_free_args(&args);
if (fuse_handle == NULL) {
--
1.8.3.1

51
0017-blobfs-check-return-value-of-strdup-in-spdk_fs_creat.patch
View File

@ -1,51 +0,0 @@
From 199309e555028889c4cf5bb02d3d5b1278bb3ce5 Mon Sep 17 00:00:00 2001
From: Zhiqiang Liu <liuzhiqiang26@huawei.com>
Date: Sun, 13 Jun 2021 16:17:32 +0800
Subject: [PATCH 17/28] blobfs: check return value of strdup in
spdk_fs_create_file_async()
In spdk_fs_create_file_async(), file->name is set to strdup(name).
We should check whether file->name is equal to NULL.
Signed-off-by: Zhiqiang Liu <liuzhiqiang26@huawei.com>
Change-Id: I2219cc353eb4711290aee2599505f57af9088bb2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8302
Community-CI: Mellanox Build Bot
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
---
lib/blobfs/blobfs.c | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/lib/blobfs/blobfs.c b/lib/blobfs/blobfs.c
index c9bcde8..65c92a9 100644
--- a/lib/blobfs/blobfs.c
+++ b/lib/blobfs/blobfs.c
@@ -1100,6 +1100,8 @@ spdk_fs_create_file_async(struct spdk_filesystem *fs, const char *name,
req = alloc_fs_request(fs->md_target.md_fs_channel);
if (req == NULL) {
SPDK_ERRLOG("Cannot allocate create async req for file=%s\n", name);
+ TAILQ_REMOVE(&fs->files, file, tailq);
+ file_free(file);
cb_fn(cb_arg, -ENOMEM);
return;
}
@@ -1110,6 +1112,14 @@ spdk_fs_create_file_async(struct spdk_filesystem *fs, const char *name,
args->arg = cb_arg;
file->name = strdup(name);
+ if (!file->name) {
+ SPDK_ERRLOG("Cannot allocate file->name for file=%s\n", name);
+ free_fs_request(req);
+ TAILQ_REMOVE(&fs->files, file, tailq);
+ file_free(file);
+ cb_fn(cb_arg, -ENOMEM);
+ return;
+ }
_file_build_trace_arg_name(file);
spdk_bs_create_blob(fs->bs, fs_create_blob_create_cb, args);
}
--
1.8.3.1

49
spdk.spec
View File

@ -2,40 +2,23 @@
%bcond_with doc
Name: spdk
Version: 21.01
Release: 4
Version: 21.01.1
Release: 1
Summary: Set of libraries and utilities for high performance user-mode storage
License: BSD and MIT
URL: http://spdk.io
Source0: https://github.com/spdk/spdk/archive/v%{version}.tar.gz
Patch1: 0001-lib-env_dpdk-fix-the-enum-rte_kernel_driver-definiti.patch
Patch2: 0002-env_dpdk-add-rte_ethdev-dependency.patch
Patch3: 0003-pkg-spdk.spec-Add-ncurses-devel-to-BuildRequires.patch
Patch4: 0004-lib-vhost-Add-version-check-when-use-RTE_VHOST_USER_.patch
Patch5: 0005-lib-nvme-Remove-qpair-from-all-lists-before-freeing-.patch
Patch6: 0006-lib-env_dpdk-add-rte_net-dependency.patch
Patch7: 0007-pkg-add-python3-requires-in-spdk.spec.patch
Patch8: 0008-sock-add-enable_quickack-and-enable_placement_id-whe.patch
Patch9: 0009-bdev-ocssd-Fix-the-bug-that-no-media-event-is-pushed.patch
Patch10: 0010-lib-iscsi-return-immediately-from-iscsi_parse_params.patch
Patch11: 0011-nbd-set-io-timeout.patch
Patch12: 0012-lib-util-Fix-valgrind-error-reported-on-ARM-platform.patch
Patch13: 0013-lib-vhost-force-cpumask-to-be-subset-of-application-.patch
Patch14: 0014-autorun-allow-pass-configuration-file-path.patch
Patch15: 0015-spdk_top-fix-app-crashing-on-tab-selection-with-TAB-.patch
Patch16: 0016-blobfs-check-return-value-of-strdup-in-blobfs_fuse_s.patch
Patch17: 0017-blobfs-check-return-value-of-strdup-in-spdk_fs_creat.patch
Patch18: 0018-blobstore-fix-memleak-problem-in-blob_load_cpl.patch
Patch19: 0019-blobstore-fix-potential-memleak-problem-in-blob_seri.patch
Patch20: 0020-idxd-fix-memleak-problem-in-spdk_idxd_configure_chan.patch
Patch21: 0021-idxd-fix-one-memleak-problem-in-spdk_idxd_get_channe.patch
Patch22: 0022-ioat-fix-potential-double-free-problem-in-ioat_chann.patch
Patch23: 0023-nvmf-check-return-value-of-strdup-in-spdk_nvmf_subsy.patch
Patch24: 0024-nvmf-check-return-value-of-strdup-in-spdk_nvmf_subsy.patch
Patch25: 0025-nvmf-fix-fd-leakage-problem-in-nvmf_vfio_user_listen.patch
Patch26: 0026-posix-set-fd-to-1-after-close-fd-in-posix_sock_creat.patch
Patch27: 0027-spdk_top-check-return-value-of-strdup-in-store_last_.patch
Patch28: 0028-uring-set-fd-to-1-after-close-fd-in-uring_sock_creat.patch
Source0: https://github.com/spdk/spdk/archive/refs/tags/v%{version}.tar.gz
Patch1: 0001-blobstore-fix-memleak-problem-in-blob_load_cpl.patch
Patch2: 0002-blobstore-fix-potential-memleak-problem-in-blob_seri.patch
Patch3: 0003-idxd-fix-memleak-problem-in-spdk_idxd_configure_chan.patch
Patch4: 0004-idxd-fix-one-memleak-problem-in-spdk_idxd_get_channe.patch
Patch5: 0005-ioat-fix-potential-double-free-problem-in-ioat_chann.patch
Patch6: 0006-nvmf-check-return-value-of-strdup-in-spdk_nvmf_subsy.patch
Patch7: 0007-nvmf-check-return-value-of-strdup-in-spdk_nvmf_subsy.patch
Patch8: 0008-nvmf-fix-fd-leakage-problem-in-nvmf_vfio_user_listen.patch
Patch9: 0009-posix-set-fd-to-1-after-close-fd-in-posix_sock_creat.patch
Patch10: 0010-spdk_top-check-return-value-of-strdup-in-store_last_.patch
Patch11: 0011-uring-set-fd-to-1-after-close-fd-in-uring_sock_creat.patch
%define package_version %{version}-%{release}
@ -117,7 +100,6 @@ BuildArch: noarch
# add -q
%autosetup -n spdk-%{version} -p1
%build
./configure --prefix=%{_usr} \
--disable-tests \
@ -196,6 +178,9 @@ mv doc/output/html/ %{install_docdir}
%changelog
* Tue Nov 23 2021 Weifeng Su <suweifeng1@huawei.com> - 21.01.1-1
- rebase to v21.01.1 Maintenance LTS Version
* Sat Jul 24 2021 Zhiqiang Liu <liuzhiqiang26@huawei.com> - 21.01-4
- backport 13 bugfix from upstream

BIN
v21.01.1.tar.gz Normal file
View File

Binary file not shown.

BIN
v21.01.tar.gz
View File

Binary file not shown.