From patchwork Wed Jun 5 13:32:58 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Yunsheng Lin X-Patchwork-Id: 13686918 X-Patchwork-Delegate: kuba@kernel.org Received: from szxga01-in.huawei.com (szxga01-in.huawei.com [45.249.212.187]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id D85D4198E7D; Wed, 5 Jun 2024 13:36:22 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=45.249.212.187 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1717594584; cv=none; b=aO2cfANvl2XDxdM02U2Y+H6xVRC/pM33ww7sZ59JoWMnfVtkD2+N8YNZqLollXRRHiTO539RTBCOcsLwTwxClq2eDMumrsx7LIuW8wWi9YILFTA4OpIeCXvOFv9FbtMnyc092gwI9IybhPUYuq7ffEoswiK4qhy9QhNZ3nIf1ps= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1717594584; c=relaxed/simple; bh=/GQc1xCGODuPlWbdb7NB50XX8Lr4CO1eTPpwc2q/JQA=; h=From:To:CC:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=AbUM0n5DC6F87mL6TWTfOPMCQ9zi5kJu4SDWvltoQ2HGOLgAscFf58qcAvNL5AIkWiPsycNw+HueNDMOAp03sX9pXfxiL2ULM+4V4GDAckQLhRbUcgp2r8ggKjIT6J6KEXKsqQrXpii6FkmR9S3F5ULEag2OUFvitWZF6SpdyQo= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=huawei.com; spf=pass smtp.mailfrom=huawei.com; arc=none smtp.client-ip=45.249.212.187 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=huawei.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=huawei.com Received: from mail.maildlp.com (unknown [172.19.88.194]) by szxga01-in.huawei.com (SkyGuard) with ESMTP id 4VvT0S1b8zzwRqm; Wed, 5 Jun 2024 21:32:24 +0800 (CST) Received: from dggpemm500005.china.huawei.com (unknown [7.185.36.74]) by mail.maildlp.com (Postfix) with ESMTPS id 07CB214037F; Wed, 5 Jun 2024 21:36:21 +0800 (CST) Received: from localhost.localdomain (10.69.192.56) by dggpemm500005.china.huawei.com (7.185.36.74) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.1.2507.39; Wed, 5 Jun 2024 21:36:20 +0800 From: Yunsheng Lin To: , , CC: , , Yunsheng Lin , Alexander Duyck , Andrew Morton , Subject: [PATCH net-next v6 08/15] mm: page_frag: reuse existing space for 'size' and 'pfmemalloc' Date: Wed, 5 Jun 2024 21:32:58 +0800 Message-ID: <20240605133306.11272-9-linyunsheng@huawei.com> X-Mailer: git-send-email 2.33.0 In-Reply-To: <20240605133306.11272-1-linyunsheng@huawei.com> References: <20240605133306.11272-1-linyunsheng@huawei.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-ClientProxiedBy: dggems703-chm.china.huawei.com (10.3.19.180) To dggpemm500005.china.huawei.com (7.185.36.74) X-Patchwork-Delegate: kuba@kernel.org Currently there is one 'struct page_frag' for every 'struct sock' and 'struct task_struct', we are about to replace the 'struct page_frag' with 'struct page_frag_cache' for them. Before begin the replacing, we need to ensure the size of 'struct page_frag_cache' is not bigger than the size of 'struct page_frag', as there may be tens of thousands of 'struct sock' and 'struct task_struct' instances in the system. By or'ing the page order & pfmemalloc with lower bits of 'va' instead of using 'u16' or 'u32' for page size and 'u8' for pfmemalloc, we are able to avoid 3 or 5 bytes space waste. And page address & pfmemalloc & order is unchanged for the same page in the same 'page_frag_cache' instance, it makes sense to fit them together. Also, it is better to replace 'offset' with 'remaining', which is the remaining size for the cache in a 'page_frag_cache' instance, we are able to do a single 'fragsz > remaining' checking for the case of cache not being enough, which should be the fast path if we ensure size is zoro when 'va' == NULL by memset'ing 'struct page_frag_cache' in page_frag_cache_init() and page_frag_cache_drain(). After this patch, the size of 'struct page_frag_cache' should be the same as the size of 'struct page_frag'. CC: Alexander Duyck Signed-off-by: Yunsheng Lin --- include/linux/page_frag_cache.h | 76 +++++++++++++++++++++++----- mm/page_frag_cache.c | 90 ++++++++++++++++++++------------- 2 files changed, 118 insertions(+), 48 deletions(-) diff --git a/include/linux/page_frag_cache.h b/include/linux/page_frag_cache.h index 6ac3a25089d1..b33904d4494f 100644 --- a/include/linux/page_frag_cache.h +++ b/include/linux/page_frag_cache.h @@ -8,29 +8,81 @@ #define PAGE_FRAG_CACHE_MAX_SIZE __ALIGN_MASK(32768, ~PAGE_MASK) #define PAGE_FRAG_CACHE_MAX_ORDER get_order(PAGE_FRAG_CACHE_MAX_SIZE) -struct page_frag_cache { - void *va; +/* + * struct encoded_va - a nonexistent type marking this pointer + * + * An 'encoded_va' pointer is a pointer to a aligned virtual address, which is + * at least aligned to PAGE_SIZE, that means there are at least 12 lower bits + * space available for other purposes. + * + * Currently we use the lower 8 bits and bit 9 for the order and PFMEMALLOC + * flag of the page this 'va' is corresponding to. + * + * Use the supplied helper functions to endcode/decode the pointer and bits. + */ +struct encoded_va; + +#define PAGE_FRAG_CACHE_ORDER_MASK GENMASK(7, 0) +#define PAGE_FRAG_CACHE_PFMEMALLOC_BIT BIT(8) +#define PAGE_FRAG_CACHE_PFMEMALLOC_SHIFT 8 + +static inline struct encoded_va *encode_aligned_va(void *va, + unsigned int order, + bool pfmemalloc) +{ #if (PAGE_SIZE < PAGE_FRAG_CACHE_MAX_SIZE) - __u16 offset; - __u16 size; + return (struct encoded_va *)((unsigned long)va | order | + pfmemalloc << PAGE_FRAG_CACHE_PFMEMALLOC_SHIFT); #else - __u32 offset; + return (struct encoded_va *)((unsigned long)va | + pfmemalloc << PAGE_FRAG_CACHE_PFMEMALLOC_SHIFT); +#endif +} + +static inline unsigned long encoded_page_order(struct encoded_va *encoded_va) +{ +#if (PAGE_SIZE < PAGE_FRAG_CACHE_MAX_SIZE) + return PAGE_FRAG_CACHE_ORDER_MASK & (unsigned long)encoded_va; +#else + return 0; +#endif +} + +static inline bool encoded_page_pfmemalloc(struct encoded_va *encoded_va) +{ + return PAGE_FRAG_CACHE_PFMEMALLOC_BIT & (unsigned long)encoded_va; +} + +static inline void *encoded_page_address(struct encoded_va *encoded_va) +{ + return (void *)((unsigned long)encoded_va & PAGE_MASK); +} + +struct page_frag_cache { + struct encoded_va *encoded_va; + +#if (PAGE_SIZE < PAGE_FRAG_CACHE_MAX_SIZE) && (BITS_PER_LONG <= 32) + u16 pagecnt_bias; + u16 remaining; +#else + u32 pagecnt_bias; + u32 remaining; #endif - /* we maintain a pagecount bias, so that we dont dirty cache line - * containing page->_refcount every time we allocate a fragment. - */ - unsigned int pagecnt_bias; - bool pfmemalloc; }; static inline void page_frag_cache_init(struct page_frag_cache *nc) { - nc->va = NULL; + memset(nc, 0, sizeof(*nc)); } static inline bool page_frag_cache_is_pfmemalloc(struct page_frag_cache *nc) { - return !!nc->pfmemalloc; + return encoded_page_pfmemalloc(nc->encoded_va); +} + +static inline unsigned int page_frag_cache_page_size(struct encoded_va *encoded_va) +{ + return PAGE_SIZE << encoded_page_order(encoded_va); } void page_frag_cache_drain(struct page_frag_cache *nc); diff --git a/mm/page_frag_cache.c b/mm/page_frag_cache.c index e5fd4acae003..756ef68cce21 100644 --- a/mm/page_frag_cache.c +++ b/mm/page_frag_cache.c @@ -18,11 +18,20 @@ #include #include "internal.h" +static void *page_frag_cache_current_va(struct page_frag_cache *nc) +{ + struct encoded_va *encoded_va = nc->encoded_va; + + return (void *)(((unsigned long)encoded_va & PAGE_MASK) | + (page_frag_cache_page_size(encoded_va) - nc->remaining)); +} + static struct page *__page_frag_cache_refill(struct page_frag_cache *nc, gfp_t gfp_mask) { struct page *page = NULL; gfp_t gfp = gfp_mask; + unsigned int order; #if (PAGE_SIZE < PAGE_FRAG_CACHE_MAX_SIZE) /* Ensure free_unref_page() can be used to free the page fragment */ @@ -32,23 +41,41 @@ static struct page *__page_frag_cache_refill(struct page_frag_cache *nc, __GFP_NOWARN | __GFP_NORETRY | __GFP_NOMEMALLOC; page = alloc_pages_node(NUMA_NO_NODE, gfp_mask, PAGE_FRAG_CACHE_MAX_ORDER); - nc->size = page ? PAGE_FRAG_CACHE_MAX_SIZE : PAGE_SIZE; #endif - if (unlikely(!page)) + if (unlikely(!page)) { page = alloc_pages_node(NUMA_NO_NODE, gfp, 0); + if (unlikely(!page)) { + memset(nc, 0, sizeof(*nc)); + return NULL; + } + + order = 0; + nc->remaining = PAGE_SIZE; + } else { + order = PAGE_FRAG_CACHE_MAX_ORDER; + nc->remaining = PAGE_FRAG_CACHE_MAX_SIZE; + } - nc->va = page ? page_address(page) : NULL; + /* Even if we own the page, we do not use atomic_set(). + * This would break get_page_unless_zero() users. + */ + page_ref_add(page, PAGE_FRAG_CACHE_MAX_SIZE); + /* reset page count bias of new frag */ + nc->pagecnt_bias = PAGE_FRAG_CACHE_MAX_SIZE + 1; + nc->encoded_va = encode_aligned_va(page_address(page), order, + page_is_pfmemalloc(page)); return page; } void page_frag_cache_drain(struct page_frag_cache *nc) { - if (!nc->va) + if (!nc->encoded_va) return; - __page_frag_cache_drain(virt_to_head_page(nc->va), nc->pagecnt_bias); - nc->va = NULL; + __page_frag_cache_drain(virt_to_head_page(nc->encoded_va), + nc->pagecnt_bias); + memset(nc, 0, sizeof(*nc)); } EXPORT_SYMBOL(page_frag_cache_drain); @@ -65,51 +92,41 @@ void *__page_frag_alloc_va_align(struct page_frag_cache *nc, unsigned int fragsz, gfp_t gfp_mask, unsigned int align_mask) { - unsigned int size = PAGE_SIZE; + struct encoded_va *encoded_va = nc->encoded_va; struct page *page; - int offset; + int remaining; + void *va; - if (unlikely(!nc->va)) { + if (unlikely(!encoded_va)) { refill: - page = __page_frag_cache_refill(nc, gfp_mask); - if (!page) + if (unlikely(!__page_frag_cache_refill(nc, gfp_mask))) return NULL; - /* Even if we own the page, we do not use atomic_set(). - * This would break get_page_unless_zero() users. - */ - page_ref_add(page, PAGE_FRAG_CACHE_MAX_SIZE); - - /* reset page count bias and offset to start of new frag */ - nc->pfmemalloc = page_is_pfmemalloc(page); - nc->pagecnt_bias = PAGE_FRAG_CACHE_MAX_SIZE + 1; - nc->offset = 0; + encoded_va = nc->encoded_va; } -#if (PAGE_SIZE < PAGE_FRAG_CACHE_MAX_SIZE) - /* if size can vary use size else just use PAGE_SIZE */ - size = nc->size; -#endif - - offset = __ALIGN_KERNEL_MASK(nc->offset, ~align_mask); - if (unlikely(offset + fragsz > size)) { - page = virt_to_page(nc->va); - + remaining = nc->remaining & align_mask; + remaining -= fragsz; + if (unlikely(remaining < 0)) { + page = virt_to_page(encoded_va); if (!page_ref_sub_and_test(page, nc->pagecnt_bias)) goto refill; - if (unlikely(nc->pfmemalloc)) { - free_unref_page(page, compound_order(page)); + if (unlikely(encoded_page_pfmemalloc(encoded_va))) { + VM_BUG_ON(compound_order(page) != + encoded_page_order(encoded_va)); + free_unref_page(page, encoded_page_order(encoded_va)); goto refill; } /* OK, page count is 0, we can safely set it */ set_page_count(page, PAGE_FRAG_CACHE_MAX_SIZE + 1); - /* reset page count bias and offset to start of new frag */ + /* reset page count bias and remaining of new frag */ nc->pagecnt_bias = PAGE_FRAG_CACHE_MAX_SIZE + 1; - offset = 0; - if (unlikely(fragsz > PAGE_SIZE)) { + nc->remaining = remaining = page_frag_cache_page_size(encoded_va); + remaining -= fragsz; + if (unlikely(remaining < 0)) { /* * The caller is trying to allocate a fragment * with fragsz > PAGE_SIZE but the cache isn't big @@ -123,10 +140,11 @@ void *__page_frag_alloc_va_align(struct page_frag_cache *nc, } } + va = page_frag_cache_current_va(nc); nc->pagecnt_bias--; - nc->offset = offset + fragsz; + nc->remaining = remaining; - return nc->va + offset; + return va; } EXPORT_SYMBOL(__page_frag_alloc_va_align);