Re: [Xen-devel] [PATCH 1/2] xenbus: req->body should be updated before req->state

[dongli.zhang@oracle.com]

On 3/3/20 1:40 AM, Julien Grall wrote:
> Hi,
> 
> On 03/03/2020 01:58, Dongli Zhang wrote:
>> The req->body should be updated before req->state is updated and the
>> order should be guaranteed by a barrier.
>>
>> Otherwise, read_reply() might return req->body = NULL.
>>
>> Below is sample callstack when the issue is reproduced on purpose by
>> reordering the updates of req->body and req->state and adding delay in
>> code between updates of req->state and req->body.
>>
>> [   22.356105] general protection fault: 0000 [#1] SMP PTI
>> [   22.361185] CPU: 2 PID: 52 Comm: xenwatch Not tainted 5.5.0xen+ #6
>> [   22.366727] Hardware name: Xen HVM domU, BIOS ...
>> [   22.372245] RIP: 0010:_parse_integer_fixup_radix+0x6/0x60
>> ... ...
>> [   22.392163] RSP: 0018:ffffb2d64023fdf0 EFLAGS: 00010246
>> [   22.395933] RAX: 0000000000000000 RBX: 75746e7562755f6d RCX: 0000000000000000
>> [   22.400871] RDX: 0000000000000000 RSI: ffffb2d64023fdfc RDI: 75746e7562755f6d
>> [   22.405874] RBP: 0000000000000000 R08: 00000000000001e8 R09: 0000000000cdcdcd
>> [   22.410945] R10: ffffb2d6402ffe00 R11: ffff9d95395eaeb0 R12: ffff9d9535935000
>> [   22.417613] R13: ffff9d9526d4a000 R14: ffff9d9526f4f340 R15: ffff9d9537654000
>> [   22.423726] FS:  0000000000000000(0000) GS:ffff9d953bc80000(0000)
>> knlGS:0000000000000000
>> [   22.429898] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [   22.434342] CR2: 000000c4206a9000 CR3: 00000001ea3fc002 CR4: 00000000001606e0
>> [   22.439645] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [   22.444941] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [   22.450342] Call Trace:
>> [   22.452509]  simple_strtoull+0x27/0x70
>> [   22.455572]  xenbus_transaction_start+0x31/0x50
>> [   22.459104]  netback_changed+0x76c/0xcc1 [xen_netfront]
>> [   22.463279]  ? find_watch+0x40/0x40
>> [   22.466156]  xenwatch_thread+0xb4/0x150
>> [   22.469309]  ? wait_woken+0x80/0x80
>> [   22.472198]  kthread+0x10e/0x130
>> [   22.474925]  ? kthread_park+0x80/0x80
>> [   22.477946]  ret_from_fork+0x35/0x40
>> [   22.480968] Modules linked in: xen_kbdfront xen_fbfront(+) xen_netfront
>> xen_blkfront
>> [   22.486783] ---[ end trace a9222030a747c3f7 ]---
>> [   22.490424] RIP: 0010:_parse_integer_fixup_radix+0x6/0x60
>>
>> The "while" is changed to "do while" so that wait_event() is used as a
>> barrier.
> 
> The correct barrier for read_reply() should be virt_rmb(). While on x86, this is
> equivalent to barrier(), on Arm this will be a dmb(ish) to prevent the processor
> re-ordering memory access.
> 
> Therefore the barrier in test_reply() (called by wait_event()) is not going to
> be sufficient for Arm.

Sorry that I just erroneously thought wait_event() would be used as read barrier.

I would change barrier() to virt_rmb() for read_reply() in v2.

Thank you very much!

Dongli Zhang