Received-SPF: pass (google.com: domain of linux-kernel+bounces-173660-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Message-ID: <1b4c50b6-2371-4e1b-aef3-d70c32888054@oracle.com>
Date: Wed, 8 May 2024 10:45:15 -0700
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH 3/3] mm/memory-failure: send SIGBUS in the event of thp
 split fail
To: Miaohe Lin <linmiaohe@huawei.com>, nao.horiguchi@gmail.com,
        akpm@linux-foundation.org, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
References: <20240501232458.3919593-1-jane.chu@oracle.com>
 <20240501232458.3919593-4-jane.chu@oracle.com>
 <038cffc0-e027-b518-460f-40099819c588@huawei.com>
 <c172fa3d-d4a4-4605-8f39-df0536718bd5@oracle.com>
 <b6c1b513-4470-4721-120c-1b1c813b2680@huawei.com>
Content-Language: en-US
From: Jane Chu <jane.chu@oracle.com>
In-Reply-To: <b6c1b513-4470-4721-120c-1b1c813b2680@huawei.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk
MIME-Version: 1.0
X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-MessageData-0: 
	=?utf-8?B?SXorY3ZZYlRvZ29QcVpVa2djQ1lpV3VjSjZ3SmZUdU9BUlhQTUFGOHc0V080?=
 =?utf-8?B?R0JKODM5S3Y4YWx0dGhFQWIvLzRqczJGdzNJTVAvYmljR3EyVjUvYy9lVStD?=
 =?utf-8?B?ZmxPdy9URStwMzR0bEdDZzlEVlA4b2IrZTI4cEFlVlE0N3dGbTdtamlrRWVx?=
 =?utf-8?B?TElwZHRHRmdsbEE1Vk9Sb1lvekpwQkUzOTNGMGgyOGFDTHZsbTVDT3JnTHZ6?=
 =?utf-8?B?aEZNbENPWVZMNUUwTDcwd2QzUE9XVEdPWTB2ZXpKbHRGWDVRNGF4cGFNZzN1?=
 =?utf-8?B?VitreE1SK05WaVFXWE94dFBDWU5mS3BhVDcxNDNYYmtOVGRxOG95VmlzOTVB?=
 =?utf-8?B?bTA2Z3kzTFp0VGZjMXlqNXptMU5tNUVlNFpWWEFhQ0JBNklOaFFwUW5NRThI?=
 =?utf-8?B?RHBBQjV0cGNyaDhSTGwwQUlDVmdnaUNPaEloZHJaV3YvNXBUZll1N1hTcnNW?=
 =?utf-8?B?ZTA1d3JTL0Fpd2ZCUkpzSVY1TW52QUl2bVB0LzUzbUU5dWZwN2VpNnRDelZy?=
 =?utf-8?B?S0NxeS9reVBjZ0RrRUp0SGtvUWNYT3RuOHcrUVRZWHpuOVVCa2F2L0Vsb1dT?=
 =?utf-8?B?WmJ0T1pBdVdUaUhrc0J6c01RTmRoL3Z1bEhiTHd6eklTbndMQWxtN1lObWUz?=
 =?utf-8?B?ZHVkNnNIMjhXS2EyRnZZOUNrQzFsRklJWi9OdDhwak5LKzg3ZCtkSTYrNlJz?=
 =?utf-8?B?U1g1Vkh0eThYd1hnTDZpa0lvUzZhRFJpY0wydUlnUjdtdUpkeW4rNVdlenZS?=
 =?utf-8?B?eWhySlN5alJXYThQRzBSOWJMdTMrTW9DenpQV2o0K1R5aXhRUEoxUkJSQWN6?=
 =?utf-8?B?ZVpWbHhaUERxckxyNkYvejdLb053RWhxTUMxL0tqaTB3c1hQZWtObVRqT1RG?=
 =?utf-8?B?bGRaL1E4WlBwTGN3ZktlR3YvUlJ1a3gwK0d3cDZzNWplamlnTy9Ka21kRjN3?=
 =?utf-8?B?c05EMDBhTUxpWmEyRXFhakNlTmZ3Zkl6Q0UrdWVxV1JUYjBlaUQrZFh2UTZK?=
 =?utf-8?B?RURqOEtjcEluVjJZVlZVVkptYXV3RVMrcjVTL0NqUEhQelI1YWFqV3d5S1c5?=
 =?utf-8?B?OW5HU29qeFZ3WHhLSjFjaWdVaE1FUGRJWkVjR0t5aW51MGlhQjBocjB3enlw?=
 =?utf-8?B?b3o2K3Rna0c1bkJLMkU5dkttT3JXV2tiT2xiYngrMlFNZExOMGpJbEYxRWkz?=
 =?utf-8?B?WGdTaHJpRSt1RjZObHF5U2RrUEhheDBjRE9VTkN4bXc5UGk0aGlqMnd5TnJW?=
 =?utf-8?B?Nm9XTUxRcDM1RUl1eUpUSDhRT0tvbks1MEpwajBPditON3JST1ViTGRjNWwy?=
 =?utf-8?B?d3RVb2xHaXhRV3RUdmFZSWo5SEhrRGhNNGw0R2lGOEFrc3BwNUlhWlZtWmlM?=
 =?utf-8?B?R0EzYjB2bU85MWhUd0I0TG80UTd6aEsxR051MERpYWdZcTlSRXRsd1g3YVBk?=
 =?utf-8?B?eDY5d3hQRGtyNlZONXpZTWNwTDJ1bE5nTkZmbmhzY0o3R0pHd2MrQXIwUWRk?=
 =?utf-8?B?U1V0UFBQcWwrZkhLcGo0dEt0QW9rTTVrSmRBV0MvNmNmeHVqMjl2TXdVU2tv?=
 =?utf-8?B?UWpEcml3K3FqYlhqZkgxQ0gyZTdFY1h2eUtGM2ExQ0FJNmJOZWNwb0p0ZGF1?=
 =?utf-8?B?bXZpY3ArRFpQNldidk83ZjFaY01qTzBDQkRDV3kxeUtWWHBpTER0QmhobHdK?=
 =?utf-8?B?U0dpNC9PK1BFRS9NeEVEc1JHdldTUDZBWG9abCtWYmk3bFFmMkVCK3lwaWQ0?=
 =?utf-8?B?WWpNL1JTMS9GdWNMUDBBWlVIVThlK1d1cDFVMHN2bGFrdVNVVXUzQmZDZE80?=
 =?utf-8?B?RnlRYzRSOER5TzJ5YjZaVklGMjV3VXNQQmtrQUlReW11RFVQaklod0sxNzFD?=
 =?utf-8?B?S2J6djRJczVqcDVpeE0wTFFUZmNnWWpMT2NoUjNtbWF6N2x1MXI1RmVoZ2hB?=
 =?utf-8?B?N0wvMW9mUitmVmY5T0pLV1lMOXlVSjF1Rm9Ya2YvVytHZkx3MFZyTEdxQ1dD?=
 =?utf-8?B?UExjN2g4OTl4bVpKTGFFYzhnTG5tbEp6QWtzVXB3bDFkYXZVamtMUEFONlVh?=
 =?utf-8?B?UVcybGxYUWVFZitLQm90eEw0V3d6VjR4K0w3S1ZPN3NGdGdtTUhFeGlBVFBr?=
 =?utf-8?Q?jSkKHm90/1AmLRPEY7bRDTi8o?=
X-MS-Exchange-AntiSpam-ExternalHop-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-ExternalHop-MessageData-0: 
	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
X-OriginatorOrg: oracle.com
X-MS-Exchange-CrossTenant-Network-Message-Id: 8a21792b-cabc-4ace-cc37-08dc6f86a05e
X-MS-Exchange-CrossTenant-AuthSource: SJ0PR10MB4429.namprd10.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 08 May 2024 17:45:18.8079
 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 4e2c6054-71cb-48f1-bd6c-3a9705aca71b
X-MS-Exchange-CrossTenant-MailboxType: HOSTED
X-MS-Exchange-CrossTenant-UserPrincipalName: A+OXUjeaXD9uqdNeh2eEio2VWvEdNUF3akqLzxgeMCGkuPa6V+Qly7/dOyNI9FKrJAnkwssz3nN8geWBTzZEow==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: PH8PR10MB6316
X-Proofpoint-Virus-Version: vendor=baseguard
 engine=ICAP:2.0.293,Aquarius:18.0.1039,Hydra:6.0.650,FMLib:17.11.176.26
 definitions=2024-05-08_09,2024-05-08_01,2023-05-22_02
X-Proofpoint-Spam-Details: rule=notspam policy=default score=0 bulkscore=0 mlxlogscore=999
 adultscore=0 suspectscore=0 malwarescore=0 phishscore=0 spamscore=0
 mlxscore=0 classifier=spam adjust=0 reason=mlx scancount=1
 engine=8.12.0-2405010000 definitions=main-2405080129
X-Proofpoint-GUID: l7w8dImPO2RWVUOnuwBVWuSccGw08xiv
X-Proofpoint-ORIG-GUID: l7w8dImPO2RWVUOnuwBVWuSccGw08xiv

On 5/8/2024 1:08 AM, Miaohe Lin wrote:

> On 2024/5/7 4:26, Jane Chu wrote:
>> On 5/5/2024 12:00 AM, Miaohe Lin wrote:
>>
>>> On 2024/5/2 7:24, Jane Chu wrote:
>>>> When handle hwpoison in a GUP longterm pin'ed thp page,
>>>> try_to_split_thp_page() will fail. And at this point, there is little else
>>>> the kernel could do except sending a SIGBUS to the user process, thus
>>>> give it a chance to recover.
>>>>
>>>> Signed-off-by: Jane Chu <jane.chu@oracle.com>
>>> Thanks for your patch. Some comments below.
>>>
>>>> ---
>>>>    mm/memory-failure.c | 36 ++++++++++++++++++++++++++++++++++++
>>>>    1 file changed, 36 insertions(+)
>>>>
>>>> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
>>>> index 7fcf182abb96..67f4d24a98e7 100644
>>>> --- a/mm/memory-failure.c
>>>> +++ b/mm/memory-failure.c
>>>> @@ -2168,6 +2168,37 @@ static int memory_failure_dev_pagemap(unsigned long pfn, int flags,
>>>>        return rc;
>>>>    }
>>>>    +/*
>>>> + * The calling condition is as such: thp split failed, page might have
>>>> + * been GUP longterm pinned, not much can be done for recovery.
>>>> + * But a SIGBUS should be delivered with vaddr provided so that the user
>>>> + * application has a chance to recover. Also, application processes'
>>>> + * election for MCE early killed will be honored.
>>>> + */
>>>> +static int kill_procs_now(struct page *p, unsigned long pfn, int flags,
>>>> +            struct page *hpage)
>>>> +{
>>>> +    struct folio *folio = page_folio(hpage);
>>>> +    LIST_HEAD(tokill);
>>>> +    int res = -EHWPOISON;
>>>> +
>>>> +    /* deal with user pages only */
>>>> +    if (PageReserved(p) || PageSlab(p) || PageTable(p) || PageOffline(p))
>>>> +        res = -EBUSY;
>>>> +    if (!(PageLRU(hpage) || PageHuge(p)))
>>>> +        res = -EBUSY;
>>> Above checks seems unneeded. We already know it's thp?
>> Agreed.
>>
>> I  lifted these checks from hwpoison_user_mapping() with a hope to make kill_procs_now() more generic,
>>
>> such as, potentially replacing kill_accessing_processes() for re-accessing hwpoisoned page.
>>
>> But I backed out at last, due to concerns that my tests might not have covered sufficient number of scenarios.
>>
>>>> +
>>>> +    if (res == -EHWPOISON) {
>>>> +        collect_procs(folio, p, &tokill, flags & MF_ACTION_REQUIRED);
>>>> +        kill_procs(&tokill, true, pfn, flags);
>>>> +    }
>>>> +
>>>> +    if (flags & MF_COUNT_INCREASED)
>>>> +        put_page(p);
>>> This if block is broken. put_page() has been done when try_to_split_thp_page() fails?
>> put_page() has not been done if try_to_split_thp_page() fails, and I think it should.
> In try_to_split_thp_page(), if split_huge_page fails, i.e. ret != 0, put_page() is called. See below:
>
> static int try_to_split_thp_page(struct page *page)
> {
> 	int ret;
>
> 	lock_page(page);
> 	ret = split_huge_page(page);
> 	unlock_page(page);
>
> 	if (unlikely(ret))
> 		put_page(page);
> 	^^^^^^^^^^^^^^^^^^^^^^^
> 	return ret;
> }
>
> Or am I miss something?

I think you caught a bug in my code, thanks!

How about moving put_page() outside try_to_split_thp_page() ?

>
>> I will revise the code so that put_page() is called regardless MF_ACTION_REQUIRED is set or not.
>>
>>>> +
>>> action_result is missing?
>> Indeed,  action_result() isn't always called, referring to the re-accessing hwpoison scenarios.
>>
>> In this case, I think the reason  is that, we just killed the process and there is nothing
>>
>> else to do or to report.
>>
>>>> +    return res;
>>>> +}
>>>> +
>>>>    /**
>>>>     * memory_failure - Handle memory failure of a page.
>>>>     * @pfn: Page Number of the corrupted page
>>>> @@ -2297,6 +2328,11 @@ int memory_failure(unsigned long pfn, int flags)
>>>>             */
>>>>            SetPageHasHWPoisoned(hpage);
>>>>            if (try_to_split_thp_page(p) < 0) {
>>> Should hwpoison_filter() be called in this case?
>> Yes, it should. I will add the hwpoison_filter check.
>>>> +            if (flags & MF_ACTION_REQUIRED) {
> Only in MF_ACTION_REQUIRED case, SIGBUS is sent to processes when thp split failed. Any reson under it?

I took a clue from kill_accessing_process() which is invoked only if 
MF_ACTION_REQUIRED is set.

The usual code path for delivery signal is

if page-is-dirty or MF_MUST_KILL-is-set or umap-failed, then

- send SIGKILL if vaddr is -EFAULT

- send SIGBUS with BUS_MCEERR_AR if MF_ACTION_REQUIRED is set

- send SIGBUS with BUS_MCEERR_AO if MF_ACTION_REQUIRED is not set and 
process elected for MCE-early-kill

So, if kill_procs_now() is invoked only if MF_ACTION_REQUIRED (as it is 
in the patch), one can argue that

the MCE-early-kill request is not honored which deviates from the 
existing behavior.

Perhaps I should remove the

+ if (flags & MF_ACTION_REQUIRED) {

check.

thanks!

-jane


>
> Thanks.
> .