[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5173673f-9aeb-4753-1e74-4ef442e06e6a@de.ibm.com>
Date: Thu, 30 Apr 2020 22:03:23 +0200
From: Christian Borntraeger <borntraeger@...ibm.com>
To: Dave Hansen <dave.hansen@...el.com>,
Claudio Imbrenda <imbrenda@...ux.ibm.com>,
akpm@...ux-foundation.org, jack@...e.cz, kirill@...temov.name
Cc: david@...hat.com, aarcange@...hat.com, linux-mm@...ck.org,
frankja@...ux.ibm.com, sfr@...b.auug.org.au, jhubbard@...dia.com,
linux-kernel@...r.kernel.org, linux-s390@...r.kernel.org,
peterz@...radead.org, sean.j.christopherson@...el.com
Subject: Re: [PATCH v1 1/1] fs/splice: add missing callback for inaccessible
pages
On 30.04.20 22:01, Dave Hansen wrote:
> On 4/30/20 12:38 PM, Christian Borntraeger wrote:
>>>
>>>> What do you see when you also do the SetPageAccessible(page);
>>>> in the else page of prep_new_page (order == 0).
>>>> (I do get > 10000 of these non compound page allocs just during boot).
>>> Yes, I see the same thing.
>>>
>>> I updated the patch and double-checked that it triggers properly with a
>>> socket-based sendfile().
>> Do you have a calltrace?
>
> It triggers with the thread from this patch _not_ applied. I just
> wanted to point out that it was able to find the real bug and that the
> patch in question squashed this instance.
Ah now it makes sense. Thanks.
>
> Here's the call trace I see:
>
>> [ 199.566150] WARNING: CPU: 0 PID: 878 at mm/page_alloc.c:8860 check_page_accessible+0x5f/0xb0
>> [ 199.567813] Modules linked in:
>> [ 199.568447] CPU: 0 PID: 878 Comm: server Not tainted 5.7.0-rc3-dirty #6544
>> [ 199.569948] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014
>> [ 199.571980] RIP: 0010:check_page_accessible+0x5f/0xb0
>> [ 199.572737] Code: 4c 01 48 85 db 74 18 48 8b 03 48 8b 7b 08 48 83 c3 18 48 89 ee ff d0 48 8b 03 48 85 c0 75 eb 48 8b 45 00 a9 00 00 20 00 75 bb <0f> 0b 0f 1f 44 00 00 3e 80 4d 02 20 5b 5d c3 65 8b 05 db 72 d3 7e
>> [ 199.576514] RSP: 0018:ffffc900003f7810 EFLAGS: 00010246
>> [ 199.577556] RAX: 000ffff800020016 RBX: ffff88800ff5b068 RCX: 0000000000000000
>> [ 199.578964] RDX: 0000000000000003 RSI: 0000000000000008 RDI: ffff88800fc18400
>> [ 199.580352] RBP: ffffea0001d5ca40 R08: 00000000000001b5 R09: ffff88800fe74000
>> [ 199.581784] R10: ffff88800fe74170 R11: ffff88800fc18400 R12: 0000000000001000
>> [ 199.583179] R13: 0000000000000000 R14: 0000000000000000 R15: ffffc900005e52d0
>> [ 199.584587] FS: 00007ffff7fe8700(0000) GS:ffff88807ce00000(0000) knlGS:0000000000000000
>> [ 199.586045] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [ 199.587200] CR2: 00007ffff7b042b0 CR3: 000000007ab5a000 CR4: 00000000003406f0
>> [ 199.588655] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [ 199.591784] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [ 199.593217] Call Trace:
>> [ 199.593755] ? e1000_xmit_frame+0x50c/0x1040
>> [ 199.594626] ? dev_hard_start_xmit+0x8d/0x1e0
>> [ 199.596126] ? sch_direct_xmit+0xe8/0x220
>> [ 199.597676] ? __qdisc_run+0x13a/0x4e0
>> [ 199.598476] ? __dev_queue_xmit+0x2d7/0x710
>> [ 199.599767] ? ip_finish_output2+0x2a8/0x550
>> [ 199.601195] ? skb_gso_validate_network_len+0x11/0x70
>> [ 199.602272] ? ip_output+0x6d/0xe0
>> [ 199.602955] ? ip_forward_options.cold.9+0x27/0x27
>> [ 199.603918] ? __ip_queue_xmit+0x14f/0x370
>> [ 199.604734] ? __tcp_transmit_skb+0x54b/0xad0
>> [ 199.605608] ? tcp_write_xmit+0x379/0x10c0
>> [ 199.606354] ? do_tcp_sendpages+0x2bc/0x5f0
>> [ 199.607206] ? tcp_sendpage_locked+0x44/0x60
>> [ 199.608066] ? tcp_sendpage+0x37/0x50
>> [ 199.608801] ? inet_sendpage+0x4f/0x80
>> [ 199.609554] ? kernel_sendpage+0x17/0x20
>> [ 199.610341] ? sock_sendpage+0x20/0x30
>> [ 199.610989] ? pipe_to_sendpage+0x60/0xa0
>> [ 199.611658] ? __splice_from_pipe+0x9f/0x180
>> [ 199.612567] ? generic_pipe_buf_nosteal+0x10/0x10
>> [ 199.613509] ? generic_pipe_buf_nosteal+0x10/0x10
>> [ 199.614452] ? splice_from_pipe+0x5d/0x90
>> [ 199.615258] ? direct_splice_actor+0x32/0x40
>> [ 199.616102] ? splice_direct_to_actor+0x101/0x220
>> [ 199.617054] ? pipe_to_sendpage+0xa0/0xa0
>> [ 199.617858] ? do_splice_direct+0x9a/0xd0
>> [ 199.618660] ? do_sendfile+0x1ce/0x3d0
>> [ 199.619411] ? __x64_sys_sendfile64+0x5c/0xc0
>> [ 199.620343] ? do_syscall_64+0x4a/0x130
>> [ 199.621042] ? entry_SYSCALL_64_after_hwframe+0x49/0xb3
>> [ 199.622119] ---[ end trace 19796ac5d41cc1f4 ]---
Powered by blists - more mailing lists