From patchwork Thu Apr 23 08:37:06 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Jan Beulich X-Patchwork-Id: 11505283 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 5FC87913 for ; Thu, 23 Apr 2020 08:37:27 +0000 (UTC) Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 41C9F20781 for ; Thu, 23 Apr 2020 08:37:27 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 41C9F20781 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXM6-0007Ur-Tj; Thu, 23 Apr 2020 08:37:10 +0000 Received: from all-amaz-eas1.inumbo.com ([34.197.232.57] helo=us1-amaz-eas2.inumbo.com) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXM5-0007Um-PW for xen-devel@lists.xenproject.org; Thu, 23 Apr 2020 08:37:09 +0000 X-Inumbo-ID: 9e667482-853d-11ea-9336-12813bfff9fa Received: from mx2.suse.de (unknown [195.135.220.15]) by us1-amaz-eas2.inumbo.com (Halon) with ESMTPS id 9e667482-853d-11ea-9336-12813bfff9fa; Thu, 23 Apr 2020 08:37:09 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id 63DD3AFC1; Thu, 23 Apr 2020 08:37:07 +0000 (UTC) Subject: [PATCH v2 1/6] x86/mem-paging: fold p2m_mem_paging_prep()'s main if()-s From: Jan Beulich To: "xen-devel@lists.xenproject.org" References: Message-ID: Date: Thu, 23 Apr 2020 10:37:06 +0200 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101 Thunderbird/68.7.0 MIME-Version: 1.0 In-Reply-To: Content-Language: en-US X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Andrew Cooper , George Dunlap , Wei Liu , =?utf-8?q?R?= =?utf-8?q?oger_Pau_Monn=C3=A9?= Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" The condition of the second can be true only if the condition of the first was met; the second half of the condition of the second then also is redundant with an earlier check. Combine them, drop a pointless local variable, and re-flow the affected gdprintk(). Signed-off-by: Jan Beulich Reviewed-by: Roger Pau Monné Acked-by: Andrew Cooper --- a/xen/arch/x86/mm/p2m.c +++ b/xen/arch/x86/mm/p2m.c @@ -1808,6 +1808,8 @@ int p2m_mem_paging_prep(struct domain *d /* Allocate a page if the gfn does not have one yet */ if ( !mfn_valid(mfn) ) { + void *guest_map; + /* If the user did not provide a buffer, we disallow */ ret = -EINVAL; if ( unlikely(user_ptr == NULL) ) @@ -1819,22 +1821,16 @@ int p2m_mem_paging_prep(struct domain *d goto out; mfn = page_to_mfn(page); page_extant = 0; - } - - /* If we were given a buffer, now is the time to use it */ - if ( !page_extant && user_ptr ) - { - void *guest_map; - int rc; ASSERT( mfn_valid(mfn) ); guest_map = map_domain_page(mfn); - rc = copy_from_user(guest_map, user_ptr, PAGE_SIZE); + ret = copy_from_user(guest_map, user_ptr, PAGE_SIZE); unmap_domain_page(guest_map); - if ( rc ) + if ( ret ) { - gdprintk(XENLOG_ERR, "Failed to load paging-in gfn %lx domain %u " - "bytes left %d\n", gfn_l, d->domain_id, rc); + gdprintk(XENLOG_ERR, + "Failed to load paging-in gfn %lx Dom%d bytes left %d\n", + gfn_l, d->domain_id, ret); ret = -EFAULT; put_page(page); /* Don't leak pages */ goto out; From patchwork Thu Apr 23 08:37:46 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Jan Beulich X-Patchwork-Id: 11505285 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 325FE913 for ; Thu, 23 Apr 2020 08:38:36 +0000 (UTC) Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 1827620715 for ; Thu, 23 Apr 2020 08:38:36 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 1827620715 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXMl-0007Yb-7Z; Thu, 23 Apr 2020 08:37:51 +0000 Received: from all-amaz-eas1.inumbo.com ([34.197.232.57] helo=us1-amaz-eas2.inumbo.com) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXMj-0007YP-PF for xen-devel@lists.xenproject.org; Thu, 23 Apr 2020 08:37:49 +0000 X-Inumbo-ID: b62f796a-853d-11ea-9336-12813bfff9fa Received: from mx2.suse.de (unknown [195.135.220.15]) by us1-amaz-eas2.inumbo.com (Halon) with ESMTPS id b62f796a-853d-11ea-9336-12813bfff9fa; Thu, 23 Apr 2020 08:37:48 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id 4DD63AFF7; Thu, 23 Apr 2020 08:37:47 +0000 (UTC) Subject: [PATCH v2 2/6] x86/mem-paging: correct p2m_mem_paging_prep()'s error handling From: Jan Beulich To: "xen-devel@lists.xenproject.org" References: Message-ID: Date: Thu, 23 Apr 2020 10:37:46 +0200 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101 Thunderbird/68.7.0 MIME-Version: 1.0 In-Reply-To: Content-Language: en-US X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Andrew Cooper , George Dunlap , Wei Liu , =?utf-8?q?R?= =?utf-8?q?oger_Pau_Monn=C3=A9?= Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" Communicating errors from p2m_set_entry() to the caller is not enough: Neither the M2P nor the stats updates should occur in such a case. Instead the allocated page needs to be freed again; for cleanliness reasons also properly take into account _PGC_allocated there. Signed-off-by: Jan Beulich Reviewed-by: Roger Pau Monné --- a/xen/arch/x86/mm/p2m.c +++ b/xen/arch/x86/mm/p2m.c @@ -1781,7 +1781,7 @@ void p2m_mem_paging_populate(struct doma */ int p2m_mem_paging_prep(struct domain *d, unsigned long gfn_l, uint64_t buffer) { - struct page_info *page; + struct page_info *page = NULL; p2m_type_t p2mt; p2m_access_t a; gfn_t gfn = _gfn(gfn_l); @@ -1816,9 +1816,19 @@ int p2m_mem_paging_prep(struct domain *d goto out; /* Get a free page */ ret = -ENOMEM; - page = alloc_domheap_page(p2m->domain, 0); + page = alloc_domheap_page(d, 0); if ( unlikely(page == NULL) ) goto out; + if ( unlikely(!get_page(page, d)) ) + { + /* + * The domain can't possibly know about this page yet, so failure + * here is a clear indication of something fishy going on. + */ + domain_crash(d); + page = NULL; + goto out; + } mfn = page_to_mfn(page); page_extant = 0; @@ -1832,7 +1842,6 @@ int p2m_mem_paging_prep(struct domain *d "Failed to load paging-in gfn %lx Dom%d bytes left %d\n", gfn_l, d->domain_id, ret); ret = -EFAULT; - put_page(page); /* Don't leak pages */ goto out; } } @@ -1843,13 +1852,24 @@ int p2m_mem_paging_prep(struct domain *d ret = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, paging_mode_log_dirty(d) ? p2m_ram_logdirty : p2m_ram_rw, a); - set_gpfn_from_mfn(mfn_x(mfn), gfn_l); + if ( !ret ) + { + set_gpfn_from_mfn(mfn_x(mfn), gfn_l); - if ( !page_extant ) - atomic_dec(&d->paged_pages); + if ( !page_extant ) + atomic_dec(&d->paged_pages); + } out: gfn_unlock(p2m, gfn, 0); + + if ( page ) + { + if ( ret ) + put_page_alloc_ref(page); + put_page(page); + } + return ret; } From patchwork Thu Apr 23 08:38:18 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Jan Beulich X-Patchwork-Id: 11505287 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 51FF4112C for ; Thu, 23 Apr 2020 08:39:04 +0000 (UTC) Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 3808821569 for ; Thu, 23 Apr 2020 08:39:04 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 3808821569 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXNH-0007eU-HL; Thu, 23 Apr 2020 08:38:23 +0000 Received: from all-amaz-eas1.inumbo.com ([34.197.232.57] helo=us1-amaz-eas2.inumbo.com) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXNG-0007eK-7K for xen-devel@lists.xenproject.org; Thu, 23 Apr 2020 08:38:22 +0000 X-Inumbo-ID: c956032e-853d-11ea-9336-12813bfff9fa Received: from mx2.suse.de (unknown [195.135.220.15]) by us1-amaz-eas2.inumbo.com (Halon) with ESMTPS id c956032e-853d-11ea-9336-12813bfff9fa; Thu, 23 Apr 2020 08:38:21 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id 5F436AFDB; Thu, 23 Apr 2020 08:38:19 +0000 (UTC) Subject: [PATCH v2 3/6] x86/mem-paging: use guest handle for XENMEM_paging_op_prep From: Jan Beulich To: "xen-devel@lists.xenproject.org" References: Message-ID: <43811c95-aa41-a34a-06ce-7d344cb1411d@suse.com> Date: Thu, 23 Apr 2020 10:38:18 +0200 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101 Thunderbird/68.7.0 MIME-Version: 1.0 In-Reply-To: Content-Language: en-US X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Andrew Cooper , Ian Jackson , George Dunlap , Wei Liu , =?utf-8?q?R?= =?utf-8?q?oger_Pau_Monn=C3=A9?= Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" While it should have been this way from the beginning, not doing so will become an actual problem with PVH Dom0. The interface change is binary compatible, but requires tools side producers to be re-built. Drop the bogus/unnecessary page alignment restriction on the input buffer at the same time. Signed-off-by: Jan Beulich Reviewed-by: Roger Pau Monné Acked-by: Wei Liu --- v2: Use HANDLE_64() instead of HANDLE_PARAM() for function parameter. --- Is there really no way to avoid the buffer copying in libxc? --- a/tools/libxc/xc_mem_paging.c +++ b/tools/libxc/xc_mem_paging.c @@ -26,15 +26,33 @@ static int xc_mem_paging_memop(xc_interf unsigned int op, uint64_t gfn, void *buffer) { xen_mem_paging_op_t mpo; + DECLARE_HYPERCALL_BOUNCE(buffer, XC_PAGE_SIZE, + XC_HYPERCALL_BUFFER_BOUNCE_IN); + int rc; memset(&mpo, 0, sizeof(mpo)); mpo.op = op; mpo.domain = domain_id; mpo.gfn = gfn; - mpo.buffer = (unsigned long) buffer; - return do_memory_op(xch, XENMEM_paging_op, &mpo, sizeof(mpo)); + if ( buffer ) + { + if ( xc_hypercall_bounce_pre(xch, buffer) ) + { + PERROR("Could not bounce memory for XENMEM_paging_op %u", op); + return -1; + } + + set_xen_guest_handle(mpo.buffer, buffer); + } + + rc = do_memory_op(xch, XENMEM_paging_op, &mpo, sizeof(mpo)); + + if ( buffer ) + xc_hypercall_bounce_post(xch, buffer); + + return rc; } int xc_mem_paging_enable(xc_interface *xch, uint32_t domain_id, @@ -92,28 +110,13 @@ int xc_mem_paging_prep(xc_interface *xch int xc_mem_paging_load(xc_interface *xch, uint32_t domain_id, uint64_t gfn, void *buffer) { - int rc, old_errno; - errno = EINVAL; if ( !buffer ) return -1; - if ( ((unsigned long) buffer) & (XC_PAGE_SIZE - 1) ) - return -1; - - if ( mlock(buffer, XC_PAGE_SIZE) ) - return -1; - - rc = xc_mem_paging_memop(xch, domain_id, - XENMEM_paging_op_prep, - gfn, buffer); - - old_errno = errno; - munlock(buffer, XC_PAGE_SIZE); - errno = old_errno; - - return rc; + return xc_mem_paging_memop(xch, domain_id, XENMEM_paging_op_prep, + gfn, buffer); } --- a/xen/arch/x86/mm/p2m.c +++ b/xen/arch/x86/mm/p2m.c @@ -1779,7 +1779,8 @@ void p2m_mem_paging_populate(struct doma * mfn if populate was called for gfn which was nominated but not evicted. In * this case only the p2mt needs to be forwarded. */ -int p2m_mem_paging_prep(struct domain *d, unsigned long gfn_l, uint64_t buffer) +int p2m_mem_paging_prep(struct domain *d, unsigned long gfn_l, + XEN_GUEST_HANDLE_64(const_uint8) buffer) { struct page_info *page = NULL; p2m_type_t p2mt; @@ -1788,13 +1789,9 @@ int p2m_mem_paging_prep(struct domain *d mfn_t mfn; struct p2m_domain *p2m = p2m_get_hostp2m(d); int ret, page_extant = 1; - const void *user_ptr = (const void *) buffer; - if ( user_ptr ) - /* Sanity check the buffer and bail out early if trouble */ - if ( (buffer & (PAGE_SIZE - 1)) || - (!access_ok(user_ptr, PAGE_SIZE)) ) - return -EINVAL; + if ( !guest_handle_okay(buffer, PAGE_SIZE) ) + return -EINVAL; gfn_lock(p2m, gfn, 0); @@ -1812,7 +1809,7 @@ int p2m_mem_paging_prep(struct domain *d /* If the user did not provide a buffer, we disallow */ ret = -EINVAL; - if ( unlikely(user_ptr == NULL) ) + if ( unlikely(guest_handle_is_null(buffer)) ) goto out; /* Get a free page */ ret = -ENOMEM; @@ -1834,7 +1831,7 @@ int p2m_mem_paging_prep(struct domain *d ASSERT( mfn_valid(mfn) ); guest_map = map_domain_page(mfn); - ret = copy_from_user(guest_map, user_ptr, PAGE_SIZE); + ret = copy_from_guest(guest_map, buffer, PAGE_SIZE); unmap_domain_page(guest_map); if ( ret ) { --- a/xen/include/asm-x86/p2m.h +++ b/xen/include/asm-x86/p2m.h @@ -741,7 +741,8 @@ void p2m_mem_paging_drop_page(struct dom /* Start populating a paged out frame */ void p2m_mem_paging_populate(struct domain *d, unsigned long gfn); /* Prepare the p2m for paging a frame in */ -int p2m_mem_paging_prep(struct domain *d, unsigned long gfn, uint64_t buffer); +int p2m_mem_paging_prep(struct domain *d, unsigned long gfn, + XEN_GUEST_HANDLE_64(const_uint8) buffer); /* Resume normal operation (in case a domain was paused) */ struct vm_event_st; void p2m_mem_paging_resume(struct domain *d, struct vm_event_st *rsp); --- a/xen/include/public/memory.h +++ b/xen/include/public/memory.h @@ -396,10 +396,10 @@ struct xen_mem_paging_op { uint8_t op; /* XENMEM_paging_op_* */ domid_t domain; - /* PAGING_PREP IN: buffer to immediately fill page in */ - uint64_aligned_t buffer; - /* Other OPs */ - uint64_aligned_t gfn; /* IN: gfn of page being operated on */ + /* IN: (XENMEM_paging_op_prep) buffer to immediately fill page from */ + XEN_GUEST_HANDLE_64(const_uint8) buffer; + /* IN: gfn of page being operated on */ + uint64_aligned_t gfn; }; typedef struct xen_mem_paging_op xen_mem_paging_op_t; DEFINE_XEN_GUEST_HANDLE(xen_mem_paging_op_t); From patchwork Thu Apr 23 08:38:44 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Jan Beulich X-Patchwork-Id: 11505289 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 4701C913 for ; Thu, 23 Apr 2020 08:39:34 +0000 (UTC) Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 2D50B20781 for ; Thu, 23 Apr 2020 08:39:34 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 2D50B20781 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXNf-0007j4-R8; Thu, 23 Apr 2020 08:38:47 +0000 Received: from all-amaz-eas1.inumbo.com ([34.197.232.57] helo=us1-amaz-eas2.inumbo.com) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXNf-0007is-2X for xen-devel@lists.xenproject.org; Thu, 23 Apr 2020 08:38:47 +0000 X-Inumbo-ID: d89c5f9a-853d-11ea-9336-12813bfff9fa Received: from mx2.suse.de (unknown [195.135.220.15]) by us1-amaz-eas2.inumbo.com (Halon) with ESMTPS id d89c5f9a-853d-11ea-9336-12813bfff9fa; Thu, 23 Apr 2020 08:38:46 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id DF9E3AFC1; Thu, 23 Apr 2020 08:38:44 +0000 (UTC) Subject: [PATCH v2 4/6] x86/mem-paging: add minimal lock order enforcement to p2m_mem_paging_prep() From: Jan Beulich To: "xen-devel@lists.xenproject.org" References: Message-ID: <4af1f459-fe7a-cd61-43cb-fb3fa4f15c00@suse.com> Date: Thu, 23 Apr 2020 10:38:44 +0200 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101 Thunderbird/68.7.0 MIME-Version: 1.0 In-Reply-To: Content-Language: en-US X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Andrew Cooper , George Dunlap , Wei Liu , =?utf-8?q?R?= =?utf-8?q?oger_Pau_Monn=C3=A9?= Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" While full checking is impossible (as the lock is being acquired/ released down the call tree), perform at least a lock level check. Signed-off-by: Jan Beulich Reviewed-by: Roger Pau Monné --- a/xen/arch/x86/mm/p2m.c +++ b/xen/arch/x86/mm/p2m.c @@ -1813,6 +1813,7 @@ int p2m_mem_paging_prep(struct domain *d goto out; /* Get a free page */ ret = -ENOMEM; + page_alloc_mm_pre_lock(d); page = alloc_domheap_page(d, 0); if ( unlikely(page == NULL) ) goto out; From patchwork Thu Apr 23 08:39:17 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Jan Beulich X-Patchwork-Id: 11505295 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 52277913 for ; Thu, 23 Apr 2020 08:41:02 +0000 (UTC) Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 2B0322084D for ; Thu, 23 Apr 2020 08:41:02 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 2B0322084D Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXOE-0007pU-9E; Thu, 23 Apr 2020 08:39:22 +0000 Received: from all-amaz-eas1.inumbo.com ([34.197.232.57] helo=us1-amaz-eas2.inumbo.com) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXOD-0007pI-7K for xen-devel@lists.xenproject.org; Thu, 23 Apr 2020 08:39:21 +0000 X-Inumbo-ID: eb8ad23b-853d-11ea-9336-12813bfff9fa Received: from mx2.suse.de (unknown [195.135.220.15]) by us1-amaz-eas2.inumbo.com (Halon) with ESMTPS id eb8ad23b-853d-11ea-9336-12813bfff9fa; Thu, 23 Apr 2020 08:39:19 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id C30BBAF4F; Thu, 23 Apr 2020 08:39:17 +0000 (UTC) Subject: [PATCH v2 5/6] x86/mem-paging: move code to its dedicated source file From: Jan Beulich To: "xen-devel@lists.xenproject.org" References: Message-ID: Date: Thu, 23 Apr 2020 10:39:17 +0200 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101 Thunderbird/68.7.0 MIME-Version: 1.0 In-Reply-To: Content-Language: en-US X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Andrew Cooper , George Dunlap , Wei Liu , =?utf-8?q?R?= =?utf-8?q?oger_Pau_Monn=C3=A9?= Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" Do a little bit of style adjustment along the way, and drop the "p2m_mem_paging_" prefixes from the now static functions. Signed-off-by: Jan Beulich Reviewed-by: Roger Pau Monné Acked-by: Andrew Cooper --- a/xen/arch/x86/mm/mem_paging.c +++ b/xen/arch/x86/mm/mem_paging.c @@ -25,6 +25,421 @@ #include #include +#include "mm-locks.h" + +/* + * p2m_mem_paging_drop_page - Tell pager to drop its reference to a paged page + * @d: guest domain + * @gfn: guest page to drop + * + * p2m_mem_paging_drop_page() will notify the pager that a paged-out gfn was + * released by the guest. The pager is supposed to drop its reference of the + * gfn. + */ +void p2m_mem_paging_drop_page(struct domain *d, unsigned long gfn, + p2m_type_t p2mt) +{ + vm_event_request_t req = { + .reason = VM_EVENT_REASON_MEM_PAGING, + .u.mem_paging.gfn = gfn + }; + + /* + * We allow no ring in this unique case, because it won't affect + * correctness of the guest execution at this point. If this is the only + * page that happens to be paged-out, we'll be okay.. but it's likely the + * guest will crash shortly anyways. + */ + int rc = vm_event_claim_slot(d, d->vm_event_paging); + + if ( rc < 0 ) + return; + + /* Send release notification to pager */ + req.u.mem_paging.flags = MEM_PAGING_DROP_PAGE; + + /* Update stats unless the page hasn't yet been evicted */ + if ( p2mt != p2m_ram_paging_out ) + atomic_dec(&d->paged_pages); + else + /* Evict will fail now, tag this request for pager */ + req.u.mem_paging.flags |= MEM_PAGING_EVICT_FAIL; + + vm_event_put_request(d, d->vm_event_paging, &req); +} + +/* + * p2m_mem_paging_populate - Tell pager to populate a paged page + * @d: guest domain + * @gfn: guest page in paging state + * + * p2m_mem_paging_populate() will notify the pager that a page in any of the + * paging states needs to be written back into the guest. + * This function needs to be called whenever gfn_to_mfn() returns any of the p2m + * paging types because the gfn may not be backed by a mfn. + * + * The gfn can be in any of the paging states, but the pager needs only be + * notified when the gfn is in the paging-out path (paging_out or paged). This + * function may be called more than once from several vcpus. If the vcpu belongs + * to the guest, the vcpu must be stopped and the pager notified that the vcpu + * was stopped. The pager needs to handle several requests for the same gfn. + * + * If the gfn is not in the paging-out path and the vcpu does not belong to the + * guest, nothing needs to be done and the function assumes that a request was + * already sent to the pager. In this case the caller has to try again until the + * gfn is fully paged in again. + */ +void p2m_mem_paging_populate(struct domain *d, unsigned long gfn_l) +{ + struct vcpu *v = current; + vm_event_request_t req = { + .reason = VM_EVENT_REASON_MEM_PAGING, + .u.mem_paging.gfn = gfn_l + }; + p2m_type_t p2mt; + p2m_access_t a; + gfn_t gfn = _gfn(gfn_l); + mfn_t mfn; + struct p2m_domain *p2m = p2m_get_hostp2m(d); + int rc = vm_event_claim_slot(d, d->vm_event_paging); + + /* We're paging. There should be a ring. */ + if ( rc == -EOPNOTSUPP ) + { + gdprintk(XENLOG_ERR, "Dom%d paging gfn %lx yet no ring in place\n", + d->domain_id, gfn_l); + /* Prevent the vcpu from faulting repeatedly on the same gfn */ + if ( v->domain == d ) + vcpu_pause_nosync(v); + domain_crash(d); + return; + } + else if ( rc < 0 ) + return; + + /* Fix p2m mapping */ + gfn_lock(p2m, gfn, 0); + mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); + /* Allow only nominated or evicted pages to enter page-in path */ + if ( p2mt == p2m_ram_paging_out || p2mt == p2m_ram_paged ) + { + /* Evict will fail now, tag this request for pager */ + if ( p2mt == p2m_ram_paging_out ) + req.u.mem_paging.flags |= MEM_PAGING_EVICT_FAIL; + + rc = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, p2m_ram_paging_in, a); + } + gfn_unlock(p2m, gfn, 0); + if ( rc < 0 ) + goto out_cancel; + + /* Pause domain if request came from guest and gfn has paging type */ + if ( p2m_is_paging(p2mt) && v->domain == d ) + { + vm_event_vcpu_pause(v); + req.flags |= VM_EVENT_FLAG_VCPU_PAUSED; + } + /* No need to inform pager if the gfn is not in the page-out path */ + else if ( p2mt != p2m_ram_paging_out && p2mt != p2m_ram_paged ) + { + /* gfn is already on its way back and vcpu is not paused */ + out_cancel: + vm_event_cancel_slot(d, d->vm_event_paging); + return; + } + + /* Send request to pager */ + req.u.mem_paging.p2mt = p2mt; + req.vcpu_id = v->vcpu_id; + + vm_event_put_request(d, d->vm_event_paging, &req); +} + +/* + * p2m_mem_paging_resume - Resume guest gfn + * @d: guest domain + * @rsp: vm_event response received + * + * p2m_mem_paging_resume() will forward the p2mt of a gfn to ram_rw. It is + * called by the pager. + * + * The gfn was previously either evicted and populated, or nominated and + * populated. If the page was evicted the p2mt will be p2m_ram_paging_in. If + * the page was just nominated the p2mt will be p2m_ram_paging_in_start because + * the pager did not call prepare(). + * + * If the gfn was dropped the vcpu needs to be unpaused. + */ +void p2m_mem_paging_resume(struct domain *d, vm_event_response_t *rsp) +{ + struct p2m_domain *p2m = p2m_get_hostp2m(d); + p2m_type_t p2mt; + p2m_access_t a; + mfn_t mfn; + + /* Fix p2m entry if the page was not dropped */ + if ( !(rsp->u.mem_paging.flags & MEM_PAGING_DROP_PAGE) ) + { + gfn_t gfn = _gfn(rsp->u.mem_access.gfn); + + gfn_lock(p2m, gfn, 0); + mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); + /* + * Allow only pages which were prepared properly, or pages which + * were nominated but not evicted. + */ + if ( mfn_valid(mfn) && (p2mt == p2m_ram_paging_in) ) + { + int rc = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, + paging_mode_log_dirty(d) ? p2m_ram_logdirty + : p2m_ram_rw, a); + + if ( !rc ) + set_gpfn_from_mfn(mfn_x(mfn), gfn_x(gfn)); + } + gfn_unlock(p2m, gfn, 0); + } +} + +/* + * nominate - Mark a guest page as to-be-paged-out + * @d: guest domain + * @gfn: guest page to nominate + * + * Returns 0 for success or negative errno values if gfn is not pageable. + * + * nominate() is called by the pager and checks if a guest page can be paged + * out. If the following conditions are met the p2mt will be changed: + * - the gfn is backed by a mfn + * - the p2mt of the gfn is pageable + * - the mfn is not used for IO + * - the mfn has exactly one user and has no special meaning + * + * Once the p2mt is changed the page is readonly for the guest. On success the + * pager can write the page contents to disk and later evict the page. + */ +static int nominate(struct domain *d, unsigned long gfn_l) +{ + struct page_info *page; + struct p2m_domain *p2m = p2m_get_hostp2m(d); + p2m_type_t p2mt; + p2m_access_t a; + gfn_t gfn = _gfn(gfn_l); + mfn_t mfn; + int ret = -EBUSY; + + gfn_lock(p2m, gfn, 0); + + mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); + + /* Check if mfn is valid */ + if ( !mfn_valid(mfn) ) + goto out; + + /* Check p2m type */ + if ( !p2m_is_pageable(p2mt) ) + goto out; + + /* Check for io memory page */ + if ( is_iomem_page(mfn) ) + goto out; + + /* Check page count and type */ + page = mfn_to_page(mfn); + if ( (page->count_info & (PGC_count_mask | PGC_allocated)) != + (1 | PGC_allocated) ) + goto out; + + if ( (page->u.inuse.type_info & PGT_count_mask) != 0 ) + goto out; + + /* Fix p2m entry */ + ret = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, p2m_ram_paging_out, a); + + out: + gfn_unlock(p2m, gfn, 0); + return ret; +} + +/* + * evict - Mark a guest page as paged-out + * @d: guest domain + * @gfn: guest page to evict + * + * Returns 0 for success or negative errno values if eviction is not possible. + * + * evict() is called by the pager and will free a guest page and release it + * back to Xen. If the following conditions are met the page can be freed: + * - the gfn is backed by a mfn + * - the gfn was nominated + * - the mfn has still exactly one user and has no special meaning + * + * After successful nomination some other process could have mapped the page. In + * this case eviction can not be done. If the gfn was populated before the pager + * could evict it, eviction can not be done either. In this case the gfn is + * still backed by a mfn. + */ +static int evict(struct domain *d, unsigned long gfn_l) +{ + struct page_info *page; + p2m_type_t p2mt; + p2m_access_t a; + gfn_t gfn = _gfn(gfn_l); + mfn_t mfn; + struct p2m_domain *p2m = p2m_get_hostp2m(d); + int ret = -EBUSY; + + gfn_lock(p2m, gfn, 0); + + /* Get mfn */ + mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); + if ( unlikely(!mfn_valid(mfn)) ) + goto out; + + /* Allow only nominated pages */ + if ( p2mt != p2m_ram_paging_out ) + goto out; + + /* Get the page so it doesn't get modified under Xen's feet */ + page = mfn_to_page(mfn); + if ( unlikely(!get_page(page, d)) ) + goto out; + + /* Check page count and type once more */ + if ( (page->count_info & (PGC_count_mask | PGC_allocated)) != + (2 | PGC_allocated) ) + goto out_put; + + if ( (page->u.inuse.type_info & PGT_count_mask) != 0 ) + goto out_put; + + /* Decrement guest domain's ref count of the page */ + put_page_alloc_ref(page); + + /* Remove mapping from p2m table */ + ret = p2m_set_entry(p2m, gfn, INVALID_MFN, PAGE_ORDER_4K, + p2m_ram_paged, a); + + /* Clear content before returning the page to Xen */ + scrub_one_page(page); + + /* Track number of paged gfns */ + atomic_inc(&d->paged_pages); + + out_put: + /* Put the page back so it gets freed */ + put_page(page); + + out: + gfn_unlock(p2m, gfn, 0); + return ret; +} + +/* + * prepare - Allocate a new page for the guest + * @d: guest domain + * @gfn: guest page in paging state + * + * prepare() will allocate a new page for the guest if the gfn is not backed + * by a mfn. It is called by the pager. + * It is required that the gfn was already populated. The gfn may already have a + * mfn if populate was called for gfn which was nominated but not evicted. In + * this case only the p2mt needs to be forwarded. + */ +static int prepare(struct domain *d, unsigned long gfn_l, + XEN_GUEST_HANDLE_64(const_uint8) buffer) +{ + struct page_info *page = NULL; + p2m_type_t p2mt; + p2m_access_t a; + gfn_t gfn = _gfn(gfn_l); + mfn_t mfn; + struct p2m_domain *p2m = p2m_get_hostp2m(d); + int ret, page_extant = 1; + + if ( !guest_handle_okay(buffer, PAGE_SIZE) ) + return -EINVAL; + + gfn_lock(p2m, gfn, 0); + + mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); + + ret = -ENOENT; + /* Allow missing pages */ + if ( (p2mt != p2m_ram_paging_in) && (p2mt != p2m_ram_paged) ) + goto out; + + /* Allocate a page if the gfn does not have one yet */ + if ( !mfn_valid(mfn) ) + { + void *guest_map; + + /* If the user did not provide a buffer, we disallow */ + ret = -EINVAL; + if ( unlikely(guest_handle_is_null(buffer)) ) + goto out; + /* Get a free page */ + ret = -ENOMEM; + page_alloc_mm_pre_lock(d); + page = alloc_domheap_page(d, 0); + if ( unlikely(page == NULL) ) + goto out; + if ( unlikely(!get_page(page, d)) ) + { + /* + * The domain can't possibly know about this page yet, so failure + * here is a clear indication of something fishy going on. + */ + domain_crash(d); + page = NULL; + goto out; + } + mfn = page_to_mfn(page); + page_extant = 0; + + ASSERT( mfn_valid(mfn) ); + guest_map = map_domain_page(mfn); + ret = copy_from_guest(guest_map, buffer, PAGE_SIZE); + unmap_domain_page(guest_map); + if ( ret ) + { + gdprintk(XENLOG_ERR, + "Failed to load paging-in gfn %lx Dom%d bytes left %d\n", + gfn_l, d->domain_id, ret); + ret = -EFAULT; + goto out; + } + } + + /* + * Make the page already guest-accessible. If the pager still has a + * pending resume operation, it will be idempotent p2m entry-wise, but + * will unpause the vcpu. + */ + ret = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, + paging_mode_log_dirty(d) ? p2m_ram_logdirty + : p2m_ram_rw, a); + if ( !ret ) + { + set_gpfn_from_mfn(mfn_x(mfn), gfn_l); + + if ( !page_extant ) + atomic_dec(&d->paged_pages); + } + + out: + gfn_unlock(p2m, gfn, 0); + + if ( page ) + { + if ( ret ) + put_page_alloc_ref(page); + put_page(page); + } + + return ret; +} + int mem_paging_memop(XEN_GUEST_HANDLE_PARAM(xen_mem_paging_op_t) arg) { int rc; @@ -50,15 +465,15 @@ int mem_paging_memop(XEN_GUEST_HANDLE_PA switch( mpo.op ) { case XENMEM_paging_op_nominate: - rc = p2m_mem_paging_nominate(d, mpo.gfn); + rc = nominate(d, mpo.gfn); break; case XENMEM_paging_op_evict: - rc = p2m_mem_paging_evict(d, mpo.gfn); + rc = evict(d, mpo.gfn); break; case XENMEM_paging_op_prep: - rc = p2m_mem_paging_prep(d, mpo.gfn, mpo.buffer); + rc = prepare(d, mpo.gfn, mpo.buffer); if ( !rc ) copyback = 1; break; --- a/xen/arch/x86/mm/p2m.c +++ b/xen/arch/x86/mm/p2m.c @@ -23,7 +23,6 @@ * along with this program; If not, see . */ -#include /* copy_from_guest() */ #include #include #include @@ -1506,418 +1505,6 @@ int set_shared_p2m_entry(struct domain * return rc; } -/** - * p2m_mem_paging_nominate - Mark a guest page as to-be-paged-out - * @d: guest domain - * @gfn: guest page to nominate - * - * Returns 0 for success or negative errno values if gfn is not pageable. - * - * p2m_mem_paging_nominate() is called by the pager and checks if a guest page - * can be paged out. If the following conditions are met the p2mt will be - * changed: - * - the gfn is backed by a mfn - * - the p2mt of the gfn is pageable - * - the mfn is not used for IO - * - the mfn has exactly one user and has no special meaning - * - * Once the p2mt is changed the page is readonly for the guest. On success the - * pager can write the page contents to disk and later evict the page. - */ -int p2m_mem_paging_nominate(struct domain *d, unsigned long gfn_l) -{ - struct page_info *page; - struct p2m_domain *p2m = p2m_get_hostp2m(d); - p2m_type_t p2mt; - p2m_access_t a; - gfn_t gfn = _gfn(gfn_l); - mfn_t mfn; - int ret = -EBUSY; - - gfn_lock(p2m, gfn, 0); - - mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); - - /* Check if mfn is valid */ - if ( !mfn_valid(mfn) ) - goto out; - - /* Check p2m type */ - if ( !p2m_is_pageable(p2mt) ) - goto out; - - /* Check for io memory page */ - if ( is_iomem_page(mfn) ) - goto out; - - /* Check page count and type */ - page = mfn_to_page(mfn); - if ( (page->count_info & (PGC_count_mask | PGC_allocated)) != - (1 | PGC_allocated) ) - goto out; - - if ( (page->u.inuse.type_info & PGT_count_mask) != 0 ) - goto out; - - /* Fix p2m entry */ - ret = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, p2m_ram_paging_out, a); - - out: - gfn_unlock(p2m, gfn, 0); - return ret; -} - -/** - * p2m_mem_paging_evict - Mark a guest page as paged-out - * @d: guest domain - * @gfn: guest page to evict - * - * Returns 0 for success or negative errno values if eviction is not possible. - * - * p2m_mem_paging_evict() is called by the pager and will free a guest page and - * release it back to Xen. If the following conditions are met the page can be - * freed: - * - the gfn is backed by a mfn - * - the gfn was nominated - * - the mfn has still exactly one user and has no special meaning - * - * After successful nomination some other process could have mapped the page. In - * this case eviction can not be done. If the gfn was populated before the pager - * could evict it, eviction can not be done either. In this case the gfn is - * still backed by a mfn. - */ -int p2m_mem_paging_evict(struct domain *d, unsigned long gfn_l) -{ - struct page_info *page; - p2m_type_t p2mt; - p2m_access_t a; - gfn_t gfn = _gfn(gfn_l); - mfn_t mfn; - struct p2m_domain *p2m = p2m_get_hostp2m(d); - int ret = -EBUSY; - - gfn_lock(p2m, gfn, 0); - - /* Get mfn */ - mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); - if ( unlikely(!mfn_valid(mfn)) ) - goto out; - - /* Allow only nominated pages */ - if ( p2mt != p2m_ram_paging_out ) - goto out; - - /* Get the page so it doesn't get modified under Xen's feet */ - page = mfn_to_page(mfn); - if ( unlikely(!get_page(page, d)) ) - goto out; - - /* Check page count and type once more */ - if ( (page->count_info & (PGC_count_mask | PGC_allocated)) != - (2 | PGC_allocated) ) - goto out_put; - - if ( (page->u.inuse.type_info & PGT_count_mask) != 0 ) - goto out_put; - - /* Decrement guest domain's ref count of the page */ - put_page_alloc_ref(page); - - /* Remove mapping from p2m table */ - ret = p2m_set_entry(p2m, gfn, INVALID_MFN, PAGE_ORDER_4K, - p2m_ram_paged, a); - - /* Clear content before returning the page to Xen */ - scrub_one_page(page); - - /* Track number of paged gfns */ - atomic_inc(&d->paged_pages); - - out_put: - /* Put the page back so it gets freed */ - put_page(page); - - out: - gfn_unlock(p2m, gfn, 0); - return ret; -} - -/** - * p2m_mem_paging_drop_page - Tell pager to drop its reference to a paged page - * @d: guest domain - * @gfn: guest page to drop - * - * p2m_mem_paging_drop_page() will notify the pager that a paged-out gfn was - * released by the guest. The pager is supposed to drop its reference of the - * gfn. - */ -void p2m_mem_paging_drop_page(struct domain *d, unsigned long gfn, - p2m_type_t p2mt) -{ - vm_event_request_t req = { - .reason = VM_EVENT_REASON_MEM_PAGING, - .u.mem_paging.gfn = gfn - }; - - /* We allow no ring in this unique case, because it won't affect - * correctness of the guest execution at this point. If this is the only - * page that happens to be paged-out, we'll be okay.. but it's likely the - * guest will crash shortly anyways. */ - int rc = vm_event_claim_slot(d, d->vm_event_paging); - if ( rc < 0 ) - return; - - /* Send release notification to pager */ - req.u.mem_paging.flags = MEM_PAGING_DROP_PAGE; - - /* Update stats unless the page hasn't yet been evicted */ - if ( p2mt != p2m_ram_paging_out ) - atomic_dec(&d->paged_pages); - else - /* Evict will fail now, tag this request for pager */ - req.u.mem_paging.flags |= MEM_PAGING_EVICT_FAIL; - - vm_event_put_request(d, d->vm_event_paging, &req); -} - -/** - * p2m_mem_paging_populate - Tell pager to populate a paged page - * @d: guest domain - * @gfn: guest page in paging state - * - * p2m_mem_paging_populate() will notify the pager that a page in any of the - * paging states needs to be written back into the guest. - * This function needs to be called whenever gfn_to_mfn() returns any of the p2m - * paging types because the gfn may not be backed by a mfn. - * - * The gfn can be in any of the paging states, but the pager needs only be - * notified when the gfn is in the paging-out path (paging_out or paged). This - * function may be called more than once from several vcpus. If the vcpu belongs - * to the guest, the vcpu must be stopped and the pager notified that the vcpu - * was stopped. The pager needs to handle several requests for the same gfn. - * - * If the gfn is not in the paging-out path and the vcpu does not belong to the - * guest, nothing needs to be done and the function assumes that a request was - * already sent to the pager. In this case the caller has to try again until the - * gfn is fully paged in again. - */ -void p2m_mem_paging_populate(struct domain *d, unsigned long gfn_l) -{ - struct vcpu *v = current; - vm_event_request_t req = { - .reason = VM_EVENT_REASON_MEM_PAGING, - .u.mem_paging.gfn = gfn_l - }; - p2m_type_t p2mt; - p2m_access_t a; - gfn_t gfn = _gfn(gfn_l); - mfn_t mfn; - struct p2m_domain *p2m = p2m_get_hostp2m(d); - - /* We're paging. There should be a ring */ - int rc = vm_event_claim_slot(d, d->vm_event_paging); - - if ( rc == -EOPNOTSUPP ) - { - gdprintk(XENLOG_ERR, "Domain %hu paging gfn %lx yet no ring " - "in place\n", d->domain_id, gfn_l); - /* Prevent the vcpu from faulting repeatedly on the same gfn */ - if ( v->domain == d ) - vcpu_pause_nosync(v); - domain_crash(d); - return; - } - else if ( rc < 0 ) - return; - - /* Fix p2m mapping */ - gfn_lock(p2m, gfn, 0); - mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); - /* Allow only nominated or evicted pages to enter page-in path */ - if ( p2mt == p2m_ram_paging_out || p2mt == p2m_ram_paged ) - { - /* Evict will fail now, tag this request for pager */ - if ( p2mt == p2m_ram_paging_out ) - req.u.mem_paging.flags |= MEM_PAGING_EVICT_FAIL; - - rc = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, p2m_ram_paging_in, a); - } - gfn_unlock(p2m, gfn, 0); - if ( rc < 0 ) - goto out_cancel; - - /* Pause domain if request came from guest and gfn has paging type */ - if ( p2m_is_paging(p2mt) && v->domain == d ) - { - vm_event_vcpu_pause(v); - req.flags |= VM_EVENT_FLAG_VCPU_PAUSED; - } - /* No need to inform pager if the gfn is not in the page-out path */ - else if ( p2mt != p2m_ram_paging_out && p2mt != p2m_ram_paged ) - { - /* gfn is already on its way back and vcpu is not paused */ - out_cancel: - vm_event_cancel_slot(d, d->vm_event_paging); - return; - } - - /* Send request to pager */ - req.u.mem_paging.p2mt = p2mt; - req.vcpu_id = v->vcpu_id; - - vm_event_put_request(d, d->vm_event_paging, &req); -} - -/** - * p2m_mem_paging_prep - Allocate a new page for the guest - * @d: guest domain - * @gfn: guest page in paging state - * - * p2m_mem_paging_prep() will allocate a new page for the guest if the gfn is - * not backed by a mfn. It is called by the pager. - * It is required that the gfn was already populated. The gfn may already have a - * mfn if populate was called for gfn which was nominated but not evicted. In - * this case only the p2mt needs to be forwarded. - */ -int p2m_mem_paging_prep(struct domain *d, unsigned long gfn_l, - XEN_GUEST_HANDLE_64(const_uint8) buffer) -{ - struct page_info *page = NULL; - p2m_type_t p2mt; - p2m_access_t a; - gfn_t gfn = _gfn(gfn_l); - mfn_t mfn; - struct p2m_domain *p2m = p2m_get_hostp2m(d); - int ret, page_extant = 1; - - if ( !guest_handle_okay(buffer, PAGE_SIZE) ) - return -EINVAL; - - gfn_lock(p2m, gfn, 0); - - mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); - - ret = -ENOENT; - /* Allow missing pages */ - if ( (p2mt != p2m_ram_paging_in) && (p2mt != p2m_ram_paged) ) - goto out; - - /* Allocate a page if the gfn does not have one yet */ - if ( !mfn_valid(mfn) ) - { - void *guest_map; - - /* If the user did not provide a buffer, we disallow */ - ret = -EINVAL; - if ( unlikely(guest_handle_is_null(buffer)) ) - goto out; - /* Get a free page */ - ret = -ENOMEM; - page_alloc_mm_pre_lock(d); - page = alloc_domheap_page(d, 0); - if ( unlikely(page == NULL) ) - goto out; - if ( unlikely(!get_page(page, d)) ) - { - /* - * The domain can't possibly know about this page yet, so failure - * here is a clear indication of something fishy going on. - */ - domain_crash(d); - page = NULL; - goto out; - } - mfn = page_to_mfn(page); - page_extant = 0; - - ASSERT( mfn_valid(mfn) ); - guest_map = map_domain_page(mfn); - ret = copy_from_guest(guest_map, buffer, PAGE_SIZE); - unmap_domain_page(guest_map); - if ( ret ) - { - gdprintk(XENLOG_ERR, - "Failed to load paging-in gfn %lx Dom%d bytes left %d\n", - gfn_l, d->domain_id, ret); - ret = -EFAULT; - goto out; - } - } - - /* Make the page already guest-accessible. If the pager still has a - * pending resume operation, it will be idempotent p2m entry-wise, - * but will unpause the vcpu */ - ret = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, - paging_mode_log_dirty(d) ? p2m_ram_logdirty - : p2m_ram_rw, a); - if ( !ret ) - { - set_gpfn_from_mfn(mfn_x(mfn), gfn_l); - - if ( !page_extant ) - atomic_dec(&d->paged_pages); - } - - out: - gfn_unlock(p2m, gfn, 0); - - if ( page ) - { - if ( ret ) - put_page_alloc_ref(page); - put_page(page); - } - - return ret; -} - -/** - * p2m_mem_paging_resume - Resume guest gfn - * @d: guest domain - * @rsp: vm_event response received - * - * p2m_mem_paging_resume() will forward the p2mt of a gfn to ram_rw. It is - * called by the pager. - * - * The gfn was previously either evicted and populated, or nominated and - * populated. If the page was evicted the p2mt will be p2m_ram_paging_in. If - * the page was just nominated the p2mt will be p2m_ram_paging_in_start because - * the pager did not call p2m_mem_paging_prep(). - * - * If the gfn was dropped the vcpu needs to be unpaused. - */ - -void p2m_mem_paging_resume(struct domain *d, vm_event_response_t *rsp) -{ - struct p2m_domain *p2m = p2m_get_hostp2m(d); - p2m_type_t p2mt; - p2m_access_t a; - mfn_t mfn; - - /* Fix p2m entry if the page was not dropped */ - if ( !(rsp->u.mem_paging.flags & MEM_PAGING_DROP_PAGE) ) - { - gfn_t gfn = _gfn(rsp->u.mem_access.gfn); - - gfn_lock(p2m, gfn, 0); - mfn = p2m->get_entry(p2m, gfn, &p2mt, &a, 0, NULL, NULL); - /* - * Allow only pages which were prepared properly, or pages which - * were nominated but not evicted. - */ - if ( mfn_valid(mfn) && (p2mt == p2m_ram_paging_in) ) - { - int rc = p2m_set_entry(p2m, gfn, mfn, PAGE_ORDER_4K, - paging_mode_log_dirty(d) ? p2m_ram_logdirty : - p2m_ram_rw, a); - - if ( !rc ) - set_gpfn_from_mfn(mfn_x(mfn), gfn_x(gfn)); - } - gfn_unlock(p2m, gfn, 0); - } -} - #ifdef CONFIG_HVM static struct p2m_domain * p2m_getlru_nestedp2m(struct domain *d, struct p2m_domain *p2m) --- a/xen/include/asm-x86/p2m.h +++ b/xen/include/asm-x86/p2m.h @@ -731,18 +731,11 @@ static inline void p2m_pod_init(struct p /* Modify p2m table for shared gfn */ int set_shared_p2m_entry(struct domain *d, unsigned long gfn, mfn_t mfn); -/* Check if a nominated gfn is valid to be paged out */ -int p2m_mem_paging_nominate(struct domain *d, unsigned long gfn); -/* Evict a frame */ -int p2m_mem_paging_evict(struct domain *d, unsigned long gfn); /* Tell xenpaging to drop a paged out frame */ void p2m_mem_paging_drop_page(struct domain *d, unsigned long gfn, p2m_type_t p2mt); /* Start populating a paged out frame */ void p2m_mem_paging_populate(struct domain *d, unsigned long gfn); -/* Prepare the p2m for paging a frame in */ -int p2m_mem_paging_prep(struct domain *d, unsigned long gfn, - XEN_GUEST_HANDLE_64(const_uint8) buffer); /* Resume normal operation (in case a domain was paused) */ struct vm_event_st; void p2m_mem_paging_resume(struct domain *d, struct vm_event_st *rsp); From patchwork Thu Apr 23 08:39:39 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Jan Beulich X-Patchwork-Id: 11505293 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 82024112C for ; Thu, 23 Apr 2020 08:40:34 +0000 (UTC) Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 683B721569 for ; Thu, 23 Apr 2020 08:40:34 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 683B721569 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXOZ-0007v8-Ml; Thu, 23 Apr 2020 08:39:43 +0000 Received: from us1-rack-iad1.inumbo.com ([172.99.69.81]) by lists.xenproject.org with esmtp (Exim 4.92) (envelope-from ) id 1jRXOZ-0007uu-8S for xen-devel@lists.xenproject.org; Thu, 23 Apr 2020 08:39:43 +0000 X-Inumbo-ID: f9ae4d1a-853d-11ea-b58d-bc764e2007e4 Received: from mx2.suse.de (unknown [195.135.220.15]) by us1-rack-iad1.inumbo.com (Halon) with ESMTPS id f9ae4d1a-853d-11ea-b58d-bc764e2007e4; Thu, 23 Apr 2020 08:39:42 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id 68E07AFC1; Thu, 23 Apr 2020 08:39:40 +0000 (UTC) Subject: [PATCH v2 6/6] x86/mem-paging: consistently use gfn_t From: Jan Beulich To: "xen-devel@lists.xenproject.org" References: Message-ID: Date: Thu, 23 Apr 2020 10:39:39 +0200 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101 Thunderbird/68.7.0 MIME-Version: 1.0 In-Reply-To: Content-Language: en-US X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Andrew Cooper , George Dunlap , Wei Liu , =?utf-8?q?R?= =?utf-8?q?oger_Pau_Monn=C3=A9?= Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" Where gprintk()s get touched anyway to switch to PRI_gfn, also switch to %pd for the domain logged. Signed-off-by: Jan Beulich Reviewed-by: Roger Pau Monné --- v2: Use gfn_t for local var in mod_l1_entry(). Use PRI_gfn when printk() arg is gfn_x(). Re-base. --- a/xen/arch/x86/hvm/dm.c +++ b/xen/arch/x86/hvm/dm.c @@ -278,7 +278,7 @@ static int set_mem_type(struct domain *d if ( p2m_is_paging(t) ) { put_gfn(d, pfn); - p2m_mem_paging_populate(d, pfn); + p2m_mem_paging_populate(d, _gfn(pfn)); return -EAGAIN; } --- a/xen/arch/x86/hvm/hvm.c +++ b/xen/arch/x86/hvm/hvm.c @@ -1988,7 +1988,7 @@ int hvm_hap_nested_page_fault(paddr_t gp * locks in such circumstance. */ if ( paged ) - p2m_mem_paging_populate(currd, gfn); + p2m_mem_paging_populate(currd, _gfn(gfn)); if ( sharing_enomem ) { @@ -3222,7 +3222,7 @@ enum hvm_translation_result hvm_translat if ( p2m_is_paging(p2mt) ) { put_page(page); - p2m_mem_paging_populate(v->domain, gfn_x(gfn)); + p2m_mem_paging_populate(v->domain, gfn); return HVMTRANS_gfn_paged_out; } if ( p2m_is_shared(p2mt) ) --- a/xen/arch/x86/mm.c +++ b/xen/arch/x86/mm.c @@ -2120,16 +2120,17 @@ static int mod_l1_entry(l1_pgentry_t *pl paging_mode_translate(pg_dom) ) { p2m_type_t p2mt; + gfn_t gfn = _gfn(l1e_get_pfn(nl1e)); p2m_query_t q = l1e_get_flags(nl1e) & _PAGE_RW ? P2M_ALLOC | P2M_UNSHARE : P2M_ALLOC; - page = get_page_from_gfn(pg_dom, l1e_get_pfn(nl1e), &p2mt, q); + page = get_page_from_gfn(pg_dom, gfn_x(gfn), &p2mt, q); if ( p2m_is_paged(p2mt) ) { if ( page ) put_page(page); - p2m_mem_paging_populate(pg_dom, l1e_get_pfn(nl1e)); + p2m_mem_paging_populate(pg_dom, gfn); return -ENOENT; } @@ -3951,7 +3952,7 @@ long do_mmu_update( put_page(page); if ( p2m_is_paged(p2mt) ) { - p2m_mem_paging_populate(pt_owner, gmfn); + p2m_mem_paging_populate(pt_owner, _gfn(gmfn)); rc = -ENOENT; } else --- a/xen/arch/x86/mm/hap/guest_walk.c +++ b/xen/arch/x86/mm/hap/guest_walk.c @@ -68,7 +68,7 @@ unsigned long hap_p2m_ga_to_gfn(GUEST_PA *pfec = PFEC_page_paged; if ( top_page ) put_page(top_page); - p2m_mem_paging_populate(p2m->domain, cr3 >> PAGE_SHIFT); + p2m_mem_paging_populate(p2m->domain, _gfn(PFN_DOWN(cr3))); return gfn_x(INVALID_GFN); } if ( p2m_is_shared(p2mt) ) @@ -110,7 +110,7 @@ unsigned long hap_p2m_ga_to_gfn(GUEST_PA { ASSERT(p2m_is_hostp2m(p2m)); *pfec = PFEC_page_paged; - p2m_mem_paging_populate(p2m->domain, gfn_x(gfn)); + p2m_mem_paging_populate(p2m->domain, gfn); return gfn_x(INVALID_GFN); } if ( p2m_is_shared(p2mt) ) --- a/xen/arch/x86/mm/mem_paging.c +++ b/xen/arch/x86/mm/mem_paging.c @@ -36,12 +36,11 @@ * released by the guest. The pager is supposed to drop its reference of the * gfn. */ -void p2m_mem_paging_drop_page(struct domain *d, unsigned long gfn, - p2m_type_t p2mt) +void p2m_mem_paging_drop_page(struct domain *d, gfn_t gfn, p2m_type_t p2mt) { vm_event_request_t req = { .reason = VM_EVENT_REASON_MEM_PAGING, - .u.mem_paging.gfn = gfn + .u.mem_paging.gfn = gfn_x(gfn) }; /* @@ -89,16 +88,15 @@ void p2m_mem_paging_drop_page(struct dom * already sent to the pager. In this case the caller has to try again until the * gfn is fully paged in again. */ -void p2m_mem_paging_populate(struct domain *d, unsigned long gfn_l) +void p2m_mem_paging_populate(struct domain *d, gfn_t gfn) { struct vcpu *v = current; vm_event_request_t req = { .reason = VM_EVENT_REASON_MEM_PAGING, - .u.mem_paging.gfn = gfn_l + .u.mem_paging.gfn = gfn_x(gfn) }; p2m_type_t p2mt; p2m_access_t a; - gfn_t gfn = _gfn(gfn_l); mfn_t mfn; struct p2m_domain *p2m = p2m_get_hostp2m(d); int rc = vm_event_claim_slot(d, d->vm_event_paging); @@ -106,8 +104,8 @@ void p2m_mem_paging_populate(struct doma /* We're paging. There should be a ring. */ if ( rc == -EOPNOTSUPP ) { - gdprintk(XENLOG_ERR, "Dom%d paging gfn %lx yet no ring in place\n", - d->domain_id, gfn_l); + gdprintk(XENLOG_ERR, "%pd paging gfn %"PRI_gfn" yet no ring in place\n", + d, gfn_x(gfn)); /* Prevent the vcpu from faulting repeatedly on the same gfn */ if ( v->domain == d ) vcpu_pause_nosync(v); @@ -218,13 +216,12 @@ void p2m_mem_paging_resume(struct domain * Once the p2mt is changed the page is readonly for the guest. On success the * pager can write the page contents to disk and later evict the page. */ -static int nominate(struct domain *d, unsigned long gfn_l) +static int nominate(struct domain *d, gfn_t gfn) { struct page_info *page; struct p2m_domain *p2m = p2m_get_hostp2m(d); p2m_type_t p2mt; p2m_access_t a; - gfn_t gfn = _gfn(gfn_l); mfn_t mfn; int ret = -EBUSY; @@ -279,12 +276,11 @@ static int nominate(struct domain *d, un * could evict it, eviction can not be done either. In this case the gfn is * still backed by a mfn. */ -static int evict(struct domain *d, unsigned long gfn_l) +static int evict(struct domain *d, gfn_t gfn) { struct page_info *page; p2m_type_t p2mt; p2m_access_t a; - gfn_t gfn = _gfn(gfn_l); mfn_t mfn; struct p2m_domain *p2m = p2m_get_hostp2m(d); int ret = -EBUSY; @@ -346,13 +342,12 @@ static int evict(struct domain *d, unsig * mfn if populate was called for gfn which was nominated but not evicted. In * this case only the p2mt needs to be forwarded. */ -static int prepare(struct domain *d, unsigned long gfn_l, +static int prepare(struct domain *d, gfn_t gfn, XEN_GUEST_HANDLE_64(const_uint8) buffer) { struct page_info *page = NULL; p2m_type_t p2mt; p2m_access_t a; - gfn_t gfn = _gfn(gfn_l); mfn_t mfn; struct p2m_domain *p2m = p2m_get_hostp2m(d); int ret, page_extant = 1; @@ -404,8 +399,8 @@ static int prepare(struct domain *d, uns if ( ret ) { gdprintk(XENLOG_ERR, - "Failed to load paging-in gfn %lx Dom%d bytes left %d\n", - gfn_l, d->domain_id, ret); + "Failed to load paging-in %pd gfn %"PRI_gfn" bytes left %d\n", + d, gfn_x(gfn), ret); ret = -EFAULT; goto out; } @@ -421,7 +416,7 @@ static int prepare(struct domain *d, uns : p2m_ram_rw, a); if ( !ret ) { - set_gpfn_from_mfn(mfn_x(mfn), gfn_l); + set_gpfn_from_mfn(mfn_x(mfn), gfn_x(gfn)); if ( !page_extant ) atomic_dec(&d->paged_pages); @@ -465,15 +460,15 @@ int mem_paging_memop(XEN_GUEST_HANDLE_PA switch( mpo.op ) { case XENMEM_paging_op_nominate: - rc = nominate(d, mpo.gfn); + rc = nominate(d, _gfn(mpo.gfn)); break; case XENMEM_paging_op_evict: - rc = evict(d, mpo.gfn); + rc = evict(d, _gfn(mpo.gfn)); break; case XENMEM_paging_op_prep: - rc = prepare(d, mpo.gfn, mpo.buffer); + rc = prepare(d, _gfn(mpo.gfn), mpo.buffer); if ( !rc ) copyback = 1; break; --- a/xen/arch/x86/mm/p2m.c +++ b/xen/arch/x86/mm/p2m.c @@ -1835,7 +1835,7 @@ void *map_domain_gfn(struct p2m_domain * ASSERT(p2m_is_hostp2m(p2m)); if ( page ) put_page(page); - p2m_mem_paging_populate(p2m->domain, gfn_x(gfn)); + p2m_mem_paging_populate(p2m->domain, gfn); *pfec = PFEC_page_paged; return NULL; } --- a/xen/arch/x86/traps.c +++ b/xen/arch/x86/traps.c @@ -848,7 +848,7 @@ int guest_wrmsr_xen(struct vcpu *v, uint if ( p2m_is_paging(t) ) { - p2m_mem_paging_populate(d, gmfn); + p2m_mem_paging_populate(d, _gfn(gmfn)); return X86EMUL_RETRY; } --- a/xen/common/memory.c +++ b/xen/common/memory.c @@ -322,7 +322,7 @@ int guest_remove_page(struct domain *d, put_gfn(d, gmfn); - p2m_mem_paging_drop_page(d, gmfn, p2mt); + p2m_mem_paging_drop_page(d, _gfn(gmfn), p2mt); return 0; } @@ -1667,7 +1667,7 @@ int check_get_page_from_gfn(struct domai if ( page ) put_page(page); - p2m_mem_paging_populate(d, gfn_x(gfn)); + p2m_mem_paging_populate(d, gfn); return -EAGAIN; } #endif --- a/xen/include/asm-x86/p2m.h +++ b/xen/include/asm-x86/p2m.h @@ -732,10 +732,9 @@ static inline void p2m_pod_init(struct p int set_shared_p2m_entry(struct domain *d, unsigned long gfn, mfn_t mfn); /* Tell xenpaging to drop a paged out frame */ -void p2m_mem_paging_drop_page(struct domain *d, unsigned long gfn, - p2m_type_t p2mt); +void p2m_mem_paging_drop_page(struct domain *d, gfn_t gfn, p2m_type_t p2mt); /* Start populating a paged out frame */ -void p2m_mem_paging_populate(struct domain *d, unsigned long gfn); +void p2m_mem_paging_populate(struct domain *d, gfn_t gfn); /* Resume normal operation (in case a domain was paused) */ struct vm_event_st; void p2m_mem_paging_resume(struct domain *d, struct vm_event_st *rsp);