Skip to contentSkip to conversation

INFO: trying to register non-static key in rxe_cleanup_task (2)

syzbot Superseded syzbot 4 participants 7 messages Feb 28, 2023 – Apr 4, 2023

Subsystems (2)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

syzbot

author syzbot
#0 lore raw

Hello,

syzbot found the following issue on:

HEAD commit:    982818426a0f Merge tag 'arm-fixes-6.3-1' of git://git.kern..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=1194800f480000
kernel config:  https://syzkaller.appspot.com/x/.config?x=90a4de3f96747e3f
dashboard link: https://syzkaller.appspot.com/bug?extid=cfcc1a3c85be15a40cba
compiler:       gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.2

Unfortunately, I don't have any reproducer for this issue yet.

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/225d8c8e9264/disk-98281842.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/87a9e2a89842/vmlinux-98281842.xz
kernel image: https://storage.googleapis.com/syzbot-assets/39bdeb741f2e/bzImage-98281842.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:

Reported-by [email protected]

Node 1 hugepages_total=2 hugepages_free=2 hugepages_surp=0 hugepages_size=2048kB 14586 total pagecache pages
0 pages in swap cache
Free swap = 0kB
Total swap = 0kB
2097051 pages RAM
0 pages HighMem/MovableOnly
392145 pages reserved
0 pages cma reserved
INFO: trying to register non-static key.
The code is fine but needs lockdep annotation, or maybe
you didn't initialize this object before use?
turning off the locking correctness validator.

CPU: 1 PID: 5486 Comm: syz-executor.2 Not tainted 6.2.0-syzkaller-12765-g982818426a0f #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 02/16/2023
Call Trace:
 <TASK>
 __dump_stack lib/dump_stack.c:88 [inline]
 dump_stack_lvl+0xd9/0x150 lib/dump_stack.c:106
 assign_lock_key kernel/locking/lockdep.c:982 [inline]
 register_lock_class+0xdb6/0x1120 kernel/locking/lockdep.c:1295
 __lock_acquire+0x108/0x5d40 kernel/locking/lockdep.c:4935
 lock_acquire kernel/locking/lockdep.c:5669 [inline]
 lock_acquire+0x1e3/0x670 kernel/locking/lockdep.c:5634
 __raw_spin_lock_bh include/linux/spinlock_api_smp.h:126 [inline]
 _raw_spin_lock_bh+0x33/0x40 kernel/locking/spinlock.c:178
 spin_lock_bh include/linux/spinlock.h:355 [inline]
 rxe_cleanup_task+0x73/0xc0 drivers/infiniband/sw/rxe/rxe_task.c:119
 rxe_qp_do_cleanup+0x8c/0x7c0 drivers/infiniband/sw/rxe/rxe_qp.c:776
 execute_in_process_context+0x3b/0x150 kernel/workqueue.c:3458
 __rxe_cleanup+0x21e/0x370 drivers/infiniband/sw/rxe/rxe_pool.c:233
 rxe_create_qp+0x2c2/0x340 drivers/infiniband/sw/rxe/rxe_verbs.c:430
 create_qp+0x5ac/0x970 drivers/infiniband/core/verbs.c:1233
 ib_create_qp_kernel+0xa1/0x310 drivers/infiniband/core/verbs.c:1344
 ib_create_qp include/rdma/ib_verbs.h:3743 [inline]
 create_mad_qp+0x177/0x380 drivers/infiniband/core/mad.c:2905
 ib_mad_port_open drivers/infiniband/core/mad.c:2986 [inline]
 ib_mad_init_device+0xf40/0x1a90 drivers/infiniband/core/mad.c:3077
 add_client_context+0x405/0x5e0 drivers/infiniband/core/device.c:721
 enable_device_and_get+0x1cd/0x3b0 drivers/infiniband/core/device.c:1332
 ib_register_device drivers/infiniband/core/device.c:1420 [inline]
 ib_register_device+0x8b1/0xbc0 drivers/infiniband/core/device.c:1366
 rxe_register_device+0x317/0x3f0 drivers/infiniband/sw/rxe/rxe_verbs.c:1096
 rxe_net_add+0x90/0xf0 drivers/infiniband/sw/rxe/rxe_net.c:524
 rxe_newlink+0xd5/0x140 drivers/infiniband/sw/rxe/rxe.c:195
 nldev_newlink+0x332/0x5e0 drivers/infiniband/core/nldev.c:1731
 rdma_nl_rcv_msg+0x371/0x6a0 drivers/infiniband/core/netlink.c:195
 rdma_nl_rcv_skb.constprop.0.isra.0+0x2fc/0x440 drivers/infiniband/core/netlink.c:239
 netlink_unicast_kernel net/netlink/af_netlink.c:1339 [inline]
 netlink_unicast+0x547/0x7f0 net/netlink/af_netlink.c:1365
 netlink_sendmsg+0x925/0xe30 net/netlink/af_netlink.c:1942
 sock_sendmsg_nosec net/socket.c:722 [inline]
 sock_sendmsg+0xde/0x190 net/socket.c:745
 ____sys_sendmsg+0x71c/0x900 net/socket.c:2504
 ___sys_sendmsg+0x110/0x1b0 net/socket.c:2558
 __sys_sendmsg+0xf7/0x1c0 net/socket.c:2587
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x39/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7fd95868c0f9
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 f1 19 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fd9594b3168 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 00007fd9587abf80 RCX: 00007fd95868c0f9
RDX: 0000000000000000 RSI: 0000000020000200 RDI: 0000000000000003
RBP: 00007fd9586e7ae9 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 00007fff9db7c9ef R14: 00007fd9594b3300 R15: 0000000000022000
 </TASK>

infiniband syz1: Couldn't create ib_mad QP1
infiniband syz1: Couldn't open port 1
RDS/IB: syz1: added
smc: adding ib device syz1 with port count 1
smc: ib device syz1 port 1 has pnetid
syz-executor.2 (5486) used greatest stack depth: 22840 bytes left

---
This report is generated by a bot. It may contain errors. See https://goo.gl/tpsmEJ for more information about syzbot. syzbot engineers can be reached at [email protected].

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

#1 lore raw
Quoted 45 lines · syzbot

Hello,

syzbot found the following issue on:

HEAD commit:    982818426a0f Merge tag 'arm-fixes-6.3-1' of git://git.kern..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=1194800f480000
kernel config:  https://syzkaller.appspot.com/x/.config?x=90a4de3f96747e3f
dashboard link: https://syzkaller.appspot.com/bug?extid=cfcc1a3c85be15a40cba
compiler:       gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.2

Unfortunately, I don't have any reproducer for this issue yet.

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/225d8c8e9264/disk-98281842.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/87a9e2a89842/vmlinux-98281842.xz
kernel image: https://storage.googleapis.com/syzbot-assets/39bdeb741f2e/bzImage-98281842.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:

Reported-by [email protected]

Node 1 hugepages_total=2 hugepages_free=2 hugepages_surp=0 hugepages_size=2048kB 14586 total pagecache pages
0 pages in swap cache
Free swap = 0kB
Total swap = 0kB
2097051 pages RAM
0 pages HighMem/MovableOnly
392145 pages reserved
0 pages cma reserved
INFO: trying to register non-static key.
The code is fine but needs lockdep annotation, or maybe
you didn't initialize this object before use?
turning off the locking correctness validator.

CPU: 1 PID: 5486 Comm: syz-executor.2 Not tainted 6.2.0-syzkaller-12765-g982818426a0f #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 02/16/2023
Call Trace:
  <TASK>
  __dump_stack lib/dump_stack.c:88 [inline]
  dump_stack_lvl+0xd9/0x150 lib/dump_stack.c:106
  assign_lock_key kernel/locking/lockdep.c:982 [inline]
  register_lock_class+0xdb6/0x1120 kernel/locking/lockdep.c:1295
  __lock_acquire+0x108/0x5d40 kernel/locking/lockdep.c:4935
  lock_acquire kernel/locking/lockdep.c:5669 [inline]
  lock_acquire+0x1e3/0x670 kernel/locking/lockdep.c:5634
Quoted 8 lines
  __raw_spin_lock_bh include/linux/spinlock_api_smp.h:126 [inline]
  _raw_spin_lock_bh+0x33/0x40 kernel/locking/spinlock.c:178
  spin_lock_bh include/linux/spinlock.h:355 [inline]
  rxe_cleanup_task+0x73/0xc0 drivers/infiniband/sw/rxe/rxe_task.c:119
  rxe_qp_do_cleanup+0x8c/0x7c0 drivers/infiniband/sw/rxe/rxe_qp.c:776
  execute_in_process_context+0x3b/0x150 kernel/workqueue.c:3458
  __rxe_cleanup+0x21e/0x370 drivers/infiniband/sw/rxe/rxe_pool.c:233
  rxe_create_qp+0x2c2/0x340 drivers/infiniband/sw/rxe/rxe_verbs.c:430

It took me a lot of time to reproduce this problem. And finally I can reproduce this problem.

The root cause is:

In this function rxe_create_qp, if some problem occur in rxe_qp_init_req, rxe_init_task functions are not executed. But error handler of rxe_create_qp still calls rxe_cleanup_task. This problem will occur.

static int rxe_create_qp
{
...

  err = rxe_qp_from_init {
        ...
        err = rxe_qp_init_req {
        ... some errors occur <---This will go to the error handler of 
rxe_create_qp
        rxe_init_task(&qp->req.task, qp, rxe_requester); <---It is not 
executed.
        rxe_init_task(&qp->comp.task, qp, rxe_completer);<---It is not 
executed.
        ...
        }
        if (err)
          goto err1;
        ...
        err = rxe_qp_init_resp{
        ...
        rxe_init_task(&qp->resp.task, qp, rxe_responder);<---It is not 
executed.
        ...
        }
        if (err)
           goto err2;
        }
        if (err)
                goto qp_init;
...
qp_init:
        rxe_cleanup(qp);  ---> rxe_cleanup_task
        return err;
}

Zhu Yanjun

Quoted 51 lines
  create_qp+0x5ac/0x970 drivers/infiniband/core/verbs.c:1233
  ib_create_qp_kernel+0xa1/0x310 drivers/infiniband/core/verbs.c:1344
  ib_create_qp include/rdma/ib_verbs.h:3743 [inline]
  create_mad_qp+0x177/0x380 drivers/infiniband/core/mad.c:2905
  ib_mad_port_open drivers/infiniband/core/mad.c:2986 [inline]
  ib_mad_init_device+0xf40/0x1a90 drivers/infiniband/core/mad.c:3077
  add_client_context+0x405/0x5e0 drivers/infiniband/core/device.c:721
  enable_device_and_get+0x1cd/0x3b0 drivers/infiniband/core/device.c:1332
  ib_register_device drivers/infiniband/core/device.c:1420 [inline]
  ib_register_device+0x8b1/0xbc0 drivers/infiniband/core/device.c:1366
  rxe_register_device+0x317/0x3f0 drivers/infiniband/sw/rxe/rxe_verbs.c:1096
  rxe_net_add+0x90/0xf0 drivers/infiniband/sw/rxe/rxe_net.c:524
  rxe_newlink+0xd5/0x140 drivers/infiniband/sw/rxe/rxe.c:195
  nldev_newlink+0x332/0x5e0 drivers/infiniband/core/nldev.c:1731
  rdma_nl_rcv_msg+0x371/0x6a0 drivers/infiniband/core/netlink.c:195
  rdma_nl_rcv_skb.constprop.0.isra.0+0x2fc/0x440 drivers/infiniband/core/netlink.c:239
  netlink_unicast_kernel net/netlink/af_netlink.c:1339 [inline]
  netlink_unicast+0x547/0x7f0 net/netlink/af_netlink.c:1365
  netlink_sendmsg+0x925/0xe30 net/netlink/af_netlink.c:1942
  sock_sendmsg_nosec net/socket.c:722 [inline]
  sock_sendmsg+0xde/0x190 net/socket.c:745
  ____sys_sendmsg+0x71c/0x900 net/socket.c:2504
  ___sys_sendmsg+0x110/0x1b0 net/socket.c:2558
  __sys_sendmsg+0xf7/0x1c0 net/socket.c:2587
  do_syscall_x64 arch/x86/entry/common.c:50 [inline]
  do_syscall_64+0x39/0xb0 arch/x86/entry/common.c:80
  entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7fd95868c0f9
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 f1 19 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fd9594b3168 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 00007fd9587abf80 RCX: 00007fd95868c0f9
RDX: 0000000000000000 RSI: 0000000020000200 RDI: 0000000000000003
RBP: 00007fd9586e7ae9 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 00007fff9db7c9ef R14: 00007fd9594b3300 R15: 0000000000022000
  </TASK>

infiniband syz1: Couldn't create ib_mad QP1
infiniband syz1: Couldn't open port 1
RDS/IB: syz1: added
smc: adding ib device syz1 with port count 1
smc: ib device syz1 port 1 has pnetid
syz-executor.2 (5486) used greatest stack depth: 22840 bytes left

---
This report is generated by a bot. It may contain errors. See https://goo.gl/tpsmEJ for more information about syzbot. syzbot engineers can be reached at [email protected].

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

#2 lore raw

PATCH 1/1 RDMA/rxe: Fix the error "trying to register non-static key in rxe_cleanup_task"

From: Zhu Yanjun <[email protected]>

In the function rxe_create_qp(), rxe_qp_from_init() is called to initialize qp, internally things like rxe_init_task are not setup until rxe_qp_init_req().

If an error occures before this point then the unwind will call rxe_cleanup() and eventually to rxe_qp_do_cleanup()/rxe_cleanup_task() which will oops when trying to access the uninitialized spinlock.

If rxe_init_task is not executed, rxe_cleanup_task will not be called.

Fixes 8700e3e7c485 ("Soft RoCE driver")
Fixes 2d4b21e0a291 ("IB/rxe: Prevent from completer to operate on non valid QP")
---
 drivers/infiniband/sw/rxe/rxe_qp.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/infiniband/sw/rxe/rxe_qp.c +12 -3
index ab72db68b58f..7856c02c1b46 100644
@@ -176,6 +176,10 @@ static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp,
 	spin_lock_init(&qp->rq.producer_lock);
 	spin_lock_init(&qp->rq.consumer_lock);
 
+	memset(&qp->req.task, 0, sizeof(struct rxe_task));
+	memset(&qp->comp.task, 0, sizeof(struct rxe_task));
+	memset(&qp->resp.task, 0, sizeof(struct rxe_task));
+
 	atomic_set(&qp->ssn, 0);
 	atomic_set(&qp->skb_out, 0);
 }
@@ -773,15 +777,20 @@ static void rxe_qp_do_cleanup(struct work_struct *work)
 
 	qp->valid = 0;
 	qp->qp_timeout_jiffies = 0;
-	rxe_cleanup_task(&qp->resp.task);
+
+	if (qp->resp.task.func)
+		rxe_cleanup_task(&qp->resp.task);
 
 	if (qp_type(qp) == IB_QPT_RC) {
 		del_timer_sync(&qp->retrans_timer);
 		del_timer_sync(&qp->rnr_nak_timer);
 	}
 
-	rxe_cleanup_task(&qp->req.task);
-	rxe_cleanup_task(&qp->comp.task);
+	if (qp->req.task.func)
+		rxe_cleanup_task(&qp->req.task);
+
+	if (qp->comp.task.func)
+		rxe_cleanup_task(&qp->comp.task);
 
 	/* flush out any receive wr's or pending requests */
 	if (qp->req.task.func)
Signature · 1 line
2.27.0
#3 lore raw

Re: [PATCH 1/1] RDMA/rxe: Fix the error "trying to register non-static key in rxe_cleanup_task"

Quoted 27 lines · Zhu Yanjun

From: Zhu Yanjun <[email protected]>

In the function rxe_create_qp(), rxe_qp_from_init() is called to initialize qp, internally things like rxe_init_task are not setup until rxe_qp_init_req().

If an error occures before this point then the unwind will call rxe_cleanup() and eventually to rxe_qp_do_cleanup()/rxe_cleanup_task() which will oops when trying to access the uninitialized spinlock.

If rxe_init_task is not executed, rxe_cleanup_task will not be called.

Fixes 8700e3e7c485 ("Soft RoCE driver")
Fixes 2d4b21e0a291 ("IB/rxe: Prevent from completer to operate on non valid QP")
---
 drivers/infiniband/sw/rxe/rxe_qp.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/infiniband/sw/rxe/rxe_qp.c +0 -0
index ab72db68b58f..7856c02c1b46 100644

drivers/infiniband/sw/rxe/rxe_qp.c static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp, · line 181

Quoted 6 lines
drivers/infiniband/sw/rxe/rxe_qp.c +3 -0
@@ -176,6 +176,10 @@ static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp,
 	spin_lock_init(&qp->rq.producer_lock);
 	spin_lock_init(&qp->rq.consumer_lock);
 
+	memset(&qp->req.task, 0, sizeof(struct rxe_task));
+	memset(&qp->comp.task, 0, sizeof(struct rxe_task));
+	memset(&qp->resp.task, 0, sizeof(struct rxe_task));

IMHO QP is already zeroed here.

Please don't send patches as reply-to.

Thanks

Quoted 31 lines
drivers/infiniband/sw/rxe/rxe_qp.c +9 -3
+
 	atomic_set(&qp->ssn, 0);
 	atomic_set(&qp->skb_out, 0);
 }
@@ -773,15 +777,20 @@ static void rxe_qp_do_cleanup(struct work_struct *work)
 
 	qp->valid = 0;
 	qp->qp_timeout_jiffies = 0;
-	rxe_cleanup_task(&qp->resp.task);
+
+	if (qp->resp.task.func)
+		rxe_cleanup_task(&qp->resp.task);
 
 	if (qp_type(qp) == IB_QPT_RC) {
 		del_timer_sync(&qp->retrans_timer);
 		del_timer_sync(&qp->rnr_nak_timer);
 	}
 
-	rxe_cleanup_task(&qp->req.task);
-	rxe_cleanup_task(&qp->comp.task);
+	if (qp->req.task.func)
+		rxe_cleanup_task(&qp->req.task);
+
+	if (qp->comp.task.func)
+		rxe_cleanup_task(&qp->comp.task);
 
 	/* flush out any receive wr's or pending requests */
 	if (qp->req.task.func)
Signature · 1 line
2.27.0
#4 lore raw

Re: [PATCH 1/1] RDMA/rxe: Fix the error "trying to register non-static key in rxe_cleanup_task"

Quoted 31 lines · Leon Romanovsky
Quoted 30 lines · Zhu Yanjun

From: Zhu Yanjun <[email protected]>

In the function rxe_create_qp(), rxe_qp_from_init() is called to initialize qp, internally things like rxe_init_task are not setup until rxe_qp_init_req().

If an error occures before this point then the unwind will call rxe_cleanup() and eventually to rxe_qp_do_cleanup()/rxe_cleanup_task() which will oops when trying to access the uninitialized spinlock.

If rxe_init_task is not executed, rxe_cleanup_task will not be called.

Fixes 8700e3e7c485 ("Soft RoCE driver")
Fixes 2d4b21e0a291 ("IB/rxe: Prevent from completer to operate on non valid QP")
---
  drivers/infiniband/sw/rxe/rxe_qp.c | 15 ++++++++++++---
  1 file changed, 12 insertions(+), 3 deletions(-)
drivers/infiniband/sw/rxe/rxe_qp.c +0 -0
index ab72db68b58f..7856c02c1b46 100644
@@ -176,6 +176,10 @@ static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp,
  	spin_lock_init(&qp->rq.producer_lock);
  	spin_lock_init(&qp->rq.consumer_lock);
 
Quoted 4 lines

drivers/infiniband/sw/rxe/rxe_qp.c static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp, · line 181

Quoted 3 lines
drivers/infiniband/sw/rxe/rxe_qp.c +3 -0
+	memset(&qp->req.task, 0, sizeof(struct rxe_task));
+	memset(&qp->comp.task, 0, sizeof(struct rxe_task));
+	memset(&qp->resp.task, 0, sizeof(struct rxe_task));

IMHO QP is already zeroed here.

Sure. Exactly. Here I just confirm that req.task, comp.task and resp.task are zeroed explicitly.

If you think it had better remove these memset functions, I will follow your advice.

Please let me know your advice.

Quoted 1 line

Please don't send patches as reply-to.

Got it. I will follow your advice.

Thanks,

Zhu Yanjun

Quoted 34 lines

Thanks

Quoted 31 lines
drivers/infiniband/sw/rxe/rxe_qp.c +9 -3
+
  	atomic_set(&qp->ssn, 0);
  	atomic_set(&qp->skb_out, 0);
  }
@@ -773,15 +777,20 @@ static void rxe_qp_do_cleanup(struct work_struct *work)
 
  	qp->valid = 0;
  	qp->qp_timeout_jiffies = 0;
-	rxe_cleanup_task(&qp->resp.task);
+
+	if (qp->resp.task.func)
+		rxe_cleanup_task(&qp->resp.task);
 
  	if (qp_type(qp) == IB_QPT_RC) {
  		del_timer_sync(&qp->retrans_timer);
  		del_timer_sync(&qp->rnr_nak_timer);
  	}
 
-	rxe_cleanup_task(&qp->req.task);
-	rxe_cleanup_task(&qp->comp.task);
+	if (qp->req.task.func)
+		rxe_cleanup_task(&qp->req.task);
+
+	if (qp->comp.task.func)
+		rxe_cleanup_task(&qp->comp.task);
 
  	/* flush out any receive wr's or pending requests */
  	if (qp->req.task.func)
Signature · 1 line
2.27.0
#5 lore raw

Re: [PATCH 1/1] RDMA/rxe: Fix the error "trying to register non-static key in rxe_cleanup_task"

Quoted 37 lines · Zhu Yanjun
Quoted 23 lines · Leon Romanovsky
Quoted 22 lines · Zhu Yanjun

From: Zhu Yanjun <[email protected]>

In the function rxe_create_qp(), rxe_qp_from_init() is called to initialize qp, internally things like rxe_init_task are not setup until rxe_qp_init_req().

If an error occures before this point then the unwind will call rxe_cleanup() and eventually to rxe_qp_do_cleanup()/rxe_cleanup_task() which will oops when trying to access the uninitialized spinlock.

If rxe_init_task is not executed, rxe_cleanup_task will not be called.

Fixes 8700e3e7c485 ("Soft RoCE driver")
Fixes 2d4b21e0a291 ("IB/rxe: Prevent from completer to operate on non valid QP")
---
  drivers/infiniband/sw/rxe/rxe_qp.c | 15 ++++++++++++---
  1 file changed, 12 insertions(+), 3 deletions(-)
Quoted 11 lines

drivers/infiniband/sw/rxe/rxe_qp.c static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp, · line 180

Quoted 10 lines
drivers/infiniband/sw/rxe/rxe_qp.c +3 -0
index ab72db68b58f..7856c02c1b46 100644
@@ -176,6 +176,10 @@ static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp,
  	spin_lock_init(&qp->rq.producer_lock);
  	spin_lock_init(&qp->rq.consumer_lock);
+	memset(&qp->req.task, 0, sizeof(struct rxe_task));
+	memset(&qp->comp.task, 0, sizeof(struct rxe_task));
+	memset(&qp->resp.task, 0, sizeof(struct rxe_task));

IMHO QP is already zeroed here.

Quoted 2 lines

Sure. Exactly. Here I just confirm that req.task, comp.task and resp.task are zeroed explicitly.

There is no need to do so. It is quite misleading to read the code and see these memset() functions as they give false impression that QP is not zeroed.

Quoted 3 lines

If you think it had better remove these memset functions, I will follow your advice.

Yes, please.

Thanks

#6 lore raw

Re: [PATCH 1/1] RDMA/rxe: Fix the error "trying to register non-static key in rxe_cleanup_task"

Quoted 36 lines · Leon Romanovsky
Quoted 35 lines · Zhu Yanjun
Quoted 34 lines · Leon Romanovsky
Quoted 27 lines · Zhu Yanjun

From: Zhu Yanjun <[email protected]>

In the function rxe_create_qp(), rxe_qp_from_init() is called to initialize qp, internally things like rxe_init_task are not setup until rxe_qp_init_req().

If an error occures before this point then the unwind will call rxe_cleanup() and eventually to rxe_qp_do_cleanup()/rxe_cleanup_task() which will oops when trying to access the uninitialized spinlock.

If rxe_init_task is not executed, rxe_cleanup_task will not be called.

Fixes 8700e3e7c485 ("Soft RoCE driver")
Fixes 2d4b21e0a291 ("IB/rxe: Prevent from completer to operate on non valid QP")
---
   drivers/infiniband/sw/rxe/rxe_qp.c | 15 ++++++++++++---
   1 file changed, 12 insertions(+), 3 deletions(-)
drivers/infiniband/sw/rxe/rxe_qp.c +0 -0
index ab72db68b58f..7856c02c1b46 100644

drivers/infiniband/sw/rxe/rxe_qp.c static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp, · line 180

Quoted 5 lines
drivers/infiniband/sw/rxe/rxe_qp.c +3 -0
@@ -176,6 +176,10 @@ static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp,
   	spin_lock_init(&qp->rq.producer_lock);
   	spin_lock_init(&qp->rq.consumer_lock);
+	memset(&qp->req.task, 0, sizeof(struct rxe_task));
+	memset(&qp->comp.task, 0, sizeof(struct rxe_task));
+	memset(&qp->resp.task, 0, sizeof(struct rxe_task));

IMHO QP is already zeroed here.

Quoted 5 lines
Quoted 2 lines

Sure. Exactly. Here I just confirm that req.task, comp.task and resp.task are zeroed explicitly.

There is no need to do so. It is quite misleading to read the code and see these memset() functions as they give false impression that QP is not zeroed.

I will remove these memset function in the latest commit.

Thanks,

Zhu Yanjun

Quoted 6 lines
Quoted 3 lines

If you think it had better remove these memset functions, I will follow your advice.

Yes, please.

Thanks