From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by smtp.lore.kernel.org (Postfix) with ESMTP id 6BDEBC30653 for ; Thu, 4 Jul 2024 09:46:33 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id CC94A6B0088; Thu, 4 Jul 2024 05:46:32 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id C2B476B0092; Thu, 4 Jul 2024 05:46:32 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id AA4AA6B00A0; Thu, 4 Jul 2024 05:46:32 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0012.hostedemail.com [216.40.44.12]) by kanga.kvack.org (Postfix) with ESMTP id 881406B0088 for ; Thu, 4 Jul 2024 05:46:32 -0400 (EDT) Received: from smtpin20.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay02.hostedemail.com (Postfix) with ESMTP id 2A6D0120977 for ; Thu, 4 Jul 2024 09:46:32 +0000 (UTC) X-FDA: 82301590224.20.A14629B Received: from foss.arm.com (foss.arm.com [217.140.110.172]) by imf14.hostedemail.com (Postfix) with ESMTP id 6820E10001C for ; Thu, 4 Jul 2024 09:46:30 +0000 (UTC) Authentication-Results: imf14.hostedemail.com; dkim=none; spf=pass (imf14.hostedemail.com: domain of ryan.roberts@arm.com designates 217.140.110.172 as permitted sender) smtp.mailfrom=ryan.roberts@arm.com; dmarc=pass (policy=none) header.from=arm.com ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1720086371; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=DMmGRBMJBrhcF84aqlEGB55WOqIsHIreTteH/1Knh/Q=; b=5Xf4PXbK9jnCAP0CQV4i1Kj6yD8tOp5Jsmu3MrVYTyUSfuasmfxFRnoaC5Fn99vU+twQ/V weJxKDuBmJuKLRuTlIVuHoiF7OWIOqjl3H0JYMrWK3bl7a547LRNHjAgJunnRQ6RFIBtOU bvTJYNzucplXNBz5/NwPUe+HDTiasR4= ARC-Authentication-Results: i=1; imf14.hostedemail.com; dkim=none; spf=pass (imf14.hostedemail.com: domain of ryan.roberts@arm.com designates 217.140.110.172 as permitted sender) smtp.mailfrom=ryan.roberts@arm.com; dmarc=pass (policy=none) header.from=arm.com ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1720086371; a=rsa-sha256; cv=none; b=AEOmIJYUt5kSzVz7p2FmZ/ReEWE/oUD2RcTBe2NbHw4AGZ91aSjDPholFTeT6IS/i+1AnW 7JuxAnT0Bo+NNFX9XbTtUJYc3v6WwuGixHdNJu+AKu11WxRiI3jSc+4evrkRCmVjxPJfPT 0JRpH7p6m2zxx0DNIePEShnx2gfmghw= Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id BB362367; Thu, 4 Jul 2024 02:46:54 -0700 (PDT) Received: from [10.1.29.168] (XHFQ2J9959.cambridge.arm.com [10.1.29.168]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id 7AECC3F766; Thu, 4 Jul 2024 02:46:28 -0700 (PDT) Message-ID: Date: Thu, 4 Jul 2024 10:46:27 +0100 MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH v2] mm: thp: support "THPeligible" semantics for mTHP with anonymous shmem Content-Language: en-GB To: Bang Li , ughd@google.com, akpm@linux-foundation.org Cc: david@redhat.com, wangkefeng.wang@huawei.com, baolin.wang@linux.alibaba.com, ioworker0@gmail.com, ziy@nvidia.com, linux-kernel@vger.kernel.org, linux-mm@kvack.org References: <20240702023401.41553-1-libang.li@antgroup.com> <507da6d0-77c5-46ca-8351-53b405ecb131@arm.com> <8c018e06-74f1-46e9-bc32-b3870342cdc1@antgroup.com> <8d9e501a-0645-4b78-809a-7c9f49f2106d@arm.com> From: Ryan Roberts In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-Stat-Signature: 7k4fdqaaekzkonmpdttyr3si5y3fbhuu X-Rspam-User: X-Rspamd-Queue-Id: 6820E10001C X-Rspamd-Server: rspam02 X-HE-Tag: 1720086390-275769 X-HE-Meta: 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 VR1omLwa fqBlCShw1ashNTvHKGUIGsAUbeBaG2f5CgrpsyQy2V+Ywh3rZg1pd3ncK0cU244GjqmJm+p+r2m/DPncPopYRC0Wjr2qAGixiqYm27B9Ro+Iwmztwe/sHbceCRdb4uztC0bmQNE/c679N+Oz6FJuLUVpx0QcT/WojH7vCSvFT0Uzz8yDBeLs56lD1CMvayv5Z5prCeNaqfY3dQ29UHaZDPLhgo9tpq2zCjHY7VjmnrSEMNL5KNJROVyJjEal6iGoC8YFpUwWG2ujJswjNtBoer0aQJZ9P/13yEjyJ X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: List-Subscribe: List-Unsubscribe: On 03/07/2024 17:02, Bang Li wrote: > On 2024/7/3 18:25, Ryan Roberts wrote: >> On 03/07/2024 08:33, Bang Li wrote: >>> Hi Ryan, >>> >>> Thanks for the review! >>> >>> On 2024/7/2 16:18, Ryan Roberts wrote: >>>> On 02/07/2024 03:34, Bang Li wrote: >>>>> After the commit 7fb1b252afb5 ("mm: shmem: add mTHP support for >>>>> anonymous shmem"), we can configure different policies through >>>>> the multi-size THP sysfs interface for anonymous shmem. But >>>>> currently "THPeligible" indicates only whether the mapping is >>>>> eligible for allocating THP-pages as well as the THP is PMD >>>>> mappable or not for anonymous shmem, we need to support semantics >>>>> for mTHP with anonymous shmem similar to those for mTHP with >>>>> anonymous memory. >>>>> >>>>> Signed-off-by: Bang Li >>>>> --- >>>>> Changes since v1 [1]: >>>>>    - Put anonymous shmem mthp related logic into >>>>>      thp_vma_allowable_orders() (per David) >>>>> >>>>> [1] >>>>> https://lore.kernel.org/linux-mm/20240628104926.34209-1-libang.li@antgroup.com/ >>>>> --- >>>>>    include/linux/huge_mm.h | 11 +++++++++++ >>>>>    mm/huge_memory.c        | 13 +++++++++---- >>>>>    mm/shmem.c              |  9 +-------- >>>>>    3 files changed, 21 insertions(+), 12 deletions(-) >>>>> >>>>> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h >>>>> index 212cca384d7e..f87136f38aa1 100644 >>>>> --- a/include/linux/huge_mm.h >>>>> +++ b/include/linux/huge_mm.h >>>>> @@ -267,6 +267,10 @@ unsigned long thp_vma_allowable_orders(struct >>>>> vm_area_struct *vma, >>>>>        return __thp_vma_allowable_orders(vma, vm_flags, tva_flags, orders); >>>>>    } >>>>>    +unsigned long shmem_allowable_huge_orders(struct inode *inode, >>>>> +                struct vm_area_struct *vma, pgoff_t index, >>>>> +                bool global_huge); >>>>> + >>>>>    struct thpsize { >>>>>        struct kobject kobj; >>>>>        struct list_head node; >>>>> @@ -460,6 +464,13 @@ static inline unsigned long >>>>> thp_vma_allowable_orders(struct vm_area_struct *vma, >>>>>        return 0; >>>>>    } >>>>>    +static inline unsigned long shmem_allowable_huge_orders(struct inode >>>>> *inode, >>>>> +                struct vm_area_struct *vma, pgoff_t index, >>>>> +                bool global_huge) >>>>> +{ >>>>> +    return 0; >>>>> +} >>>>> + >>>>>    #define transparent_hugepage_flags 0UL >>>>>      #define thp_get_unmapped_area    NULL >>>>> diff --git a/mm/huge_memory.c b/mm/huge_memory.c >>>>> index c7ce28f6b7f3..ea377bb4af91 100644 >>>>> --- a/mm/huge_memory.c >>>>> +++ b/mm/huge_memory.c >>>>> @@ -151,10 +151,15 @@ unsigned long __thp_vma_allowable_orders(struct >>>>> vm_area_struct *vma, >>>>>         * Must be done before hugepage flags check since shmem has its >>>>>         * own flags. >>>>>         */ >>>>> -    if (!in_pf && shmem_file(vma->vm_file)) >>>>> -        return shmem_is_huge(file_inode(vma->vm_file), vma->vm_pgoff, >>>>> -                     !enforce_sysfs, vma->vm_mm, vm_flags) >>>>> -            ? orders : 0; >>>>> +    if (!in_pf && shmem_file(vma->vm_file)) { >>>>> +        bool global_huge = shmem_is_huge(file_inode(vma->vm_file), >>>>> vma->vm_pgoff, >>>>> +                            !enforce_sysfs, vma->vm_mm, vm_flags); >>>>> + >>>>> +        if (!vma_is_anon_shmem(vma)) >>>>> +            return global_huge? orders : 0; >>>> >>>> nit: missing space before '?' >>> >>> Yes, thanks. >>> >>>> >>>>> +        return shmem_allowable_huge_orders(file_inode(vma->vm_file), >>>>> +                            vma, vma->vm_pgoff, global_huge); >>>> >>>> What's the rationale for splitting these functions into shmem_is_huge() and >>>> shmem_allowable_huge_orders()? Why not just have a single >>>> shmem_allowable_huge_orders() that tells you the answer? >>>> >>> >>> Currently, shmem_is_huge() is used for all shmem implementations to determine >>> whether the conditions for using THP are met. And shmem_allowable_huge_orders() >>> is currently mainly used for anonymous shmem's mTHP to obtain all orders that >>> meet the conditions. In my opinion, there is no problem in separating these two >>> functions. In the future, when mTHP of other shmem types is also implemented, >>> will shmem_is_huge() be unnecessary? >> >> Personally, I consider shmem_is_huge() to be superfluous; If you have one >> function, shmem_allowable_huge_orders(), that gives you all the information you >> need. If the inode only allows PMD-size, then only return that bit in the field. >> IMHO removing shmem_is_huge() would make things more readable. > > Thank you very much for your opinion. I think there is no big problem in keeping > the current interface status quo. In my opinion, when we implement mTHP for all > shmems, it may be better to remove shmem_is_huge() and use > shmem_allowable_huge_orders() uniformly. I believe that it won't be long before > someone proposes a non-anonymous shared mTHP patchset. Anyway, thank you for > your suggestion :). See [1] for an expanded list of concerns I have about the current state of the interface and implementation. Thanks, Ryan > > Thanks, > Bang > >> >> Thanks, >> Ryan >> >>> >>> Thanks, >>> Bang >>> >>>>> +    } >>>>>          if (!vma_is_anonymous(vma)) { >>>>>            /* >>>>> diff --git a/mm/shmem.c b/mm/shmem.c >>>>> index d495c0701a83..aa85df9c662a 100644 >>>>> --- a/mm/shmem.c >>>>> +++ b/mm/shmem.c >>>>> @@ -1622,7 +1622,7 @@ static gfp_t limit_gfp_mask(gfp_t huge_gfp, gfp_t >>>>> limit_gfp) >>>>>    } >>>>>      #ifdef CONFIG_TRANSPARENT_HUGEPAGE >>>>> -static unsigned long shmem_allowable_huge_orders(struct inode *inode, >>>>> +unsigned long shmem_allowable_huge_orders(struct inode *inode, >>>>>                    struct vm_area_struct *vma, pgoff_t index, >>>>>                    bool global_huge) >>>>>    { >>>>> @@ -1707,13 +1707,6 @@ static unsigned long shmem_suitable_orders(struct >>>>> inode *inode, struct vm_fault >>>>>        return orders; >>>>>    } >>>>>    #else >>>>> -static unsigned long shmem_allowable_huge_orders(struct inode *inode, >>>>> -                struct vm_area_struct *vma, pgoff_t index, >>>>> -                bool global_huge) >>>>> -{ >>>>> -    return 0; >>>>> -} >>>>> - >>>>>    static unsigned long shmem_suitable_orders(struct inode *inode, struct >>>>> vm_fault *vmf, >>>>>                           struct address_space *mapping, pgoff_t index, >>>>>                           unsigned long orders)