xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Stefano Stabellini <sstabellini@kernel.org>
To: Oleksandr Tyshchenko <olekstysh@gmail.com>
Cc: xen-devel@lists.xenproject.org,
	"Oleksandr Tyshchenko" <oleksandr_tyshchenko@epam.com>,
	"Stefano Stabellini" <sstabellini@kernel.org>,
	"Julien Grall" <julien@xen.org>,
	"Volodymyr Babchuk" <Volodymyr_Babchuk@epam.com>,
	"Andrew Cooper" <andrew.cooper3@citrix.com>,
	"George Dunlap" <george.dunlap@citrix.com>,
	"Ian Jackson" <iwj@xenproject.org>,
	"Jan Beulich" <jbeulich@suse.com>, "Wei Liu" <wl@xen.org>,
	"Roger Pau Monné" <roger.pau@citrix.com>,
	"Julien Grall" <julien.grall@arm.com>
Subject: Re: [PATCH V3 16/23] xen/mm: Handle properly reference in set_foreign_p2m_entry() on Arm
Date: Wed, 9 Dec 2020 15:49:27 -0800 (PST)	[thread overview]
Message-ID: <alpine.DEB.2.21.2012091549140.20986@sstabellini-ThinkPad-T480s> (raw)
In-Reply-To: <1606732298-22107-17-git-send-email-olekstysh@gmail.com>

[-- Attachment #1: Type: text/plain, Size: 9299 bytes --]

On Mon, 30 Nov 2020, Oleksandr Tyshchenko wrote:
> From: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
> 
> This patch implements reference counting of foreign entries in
> in set_foreign_p2m_entry() on Arm. This is a mandatory action if
> we want to run emulator (IOREQ server) in other than dom0 domain,
> as we can't trust it to do the right thing if it is not running
> in dom0. So we need to grab a reference on the page to avoid it
> disappearing.
> 
> It is valid to always pass "p2m_map_foreign_rw" type to
> guest_physmap_add_entry() since the current and foreign domains
> would be always different. A case when they are equal would be
> rejected by rcu_lock_remote_domain_by_id(). Besides the similar
> comment in the code put a respective ASSERT() to catch incorrect
> usage in future.
> 
> It was tested with IOREQ feature to confirm that all the pages given
> to this function belong to a domain, so we can use the same approach
> as for XENMAPSPACE_gmfn_foreign handling in xenmem_add_to_physmap_one().
> 
> This involves adding an extra parameter for the foreign domain to
> set_foreign_p2m_entry() and a helper to indicate whether the arch
> supports the reference counting of foreign entries and the restriction
> for the hardware domain in the common code can be skipped for it.
> 
> Signed-off-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
> CC: Julien Grall <julien.grall@arm.com>

The arm side looks OK to me


> ---
> Please note, this is a split/cleanup/hardening of Julien's PoC:
> "Add support for Guest IO forwarding to a device emulator"
> 
> Changes RFC -> V1:
>    - new patch, was split from:
>      "[RFC PATCH V1 04/12] xen/arm: Introduce arch specific bits for IOREQ/DM features"
>    - rewrite a logic to handle properly reference in set_foreign_p2m_entry()
>      instead of treating foreign entries as p2m_ram_rw
> 
> Changes V1 -> V2:
>    - rebase according to the recent changes to acquire_resource()
>    - update patch description
>    - introduce arch_refcounts_p2m()
>    - add an explanation why p2m_map_foreign_rw is valid
>    - move set_foreign_p2m_entry() to p2m-common.h
>    - add const to new parameter
> 
> Changes V2 -> V3:
>    - update patch description
>    - rename arch_refcounts_p2m() to arch_acquire_resource_check()
>    - move comment to x86’s arch_acquire_resource_check()
>    - return rc in Arm's set_foreign_p2m_entry()
>    - put a respective ASSERT() into Arm's set_foreign_p2m_entry()
> ---
> ---
>  xen/arch/arm/p2m.c           | 24 ++++++++++++++++++++++++
>  xen/arch/x86/mm/p2m.c        |  5 +++--
>  xen/common/memory.c          | 10 +++-------
>  xen/include/asm-arm/p2m.h    | 19 +++++++++----------
>  xen/include/asm-x86/p2m.h    | 16 +++++++++++++---
>  xen/include/xen/p2m-common.h |  4 ++++
>  6 files changed, 56 insertions(+), 22 deletions(-)
> 
> diff --git a/xen/arch/arm/p2m.c b/xen/arch/arm/p2m.c
> index 4eeb867..5b8d494 100644
> --- a/xen/arch/arm/p2m.c
> +++ b/xen/arch/arm/p2m.c
> @@ -1380,6 +1380,30 @@ int guest_physmap_remove_page(struct domain *d, gfn_t gfn, mfn_t mfn,
>      return p2m_remove_mapping(d, gfn, (1 << page_order), mfn);
>  }
>  
> +int set_foreign_p2m_entry(struct domain *d, const struct domain *fd,
> +                          unsigned long gfn, mfn_t mfn)
> +{
> +    struct page_info *page = mfn_to_page(mfn);
> +    int rc;
> +
> +    if ( !get_page(page, fd) )
> +        return -EINVAL;
> +
> +    /*
> +     * It is valid to always use p2m_map_foreign_rw here as if this gets
> +     * called then d != fd. A case when d == fd would be rejected by
> +     * rcu_lock_remote_domain_by_id() earlier. Put a respective ASSERT()
> +     * to catch incorrect usage in future.
> +     */
> +    ASSERT(d != fd);
> +
> +    rc = guest_physmap_add_entry(d, _gfn(gfn), mfn, 0, p2m_map_foreign_rw);
> +    if ( rc )
> +        put_page(page);
> +
> +    return rc;
> +}
> +
>  static struct page_info *p2m_allocate_root(void)
>  {
>      struct page_info *page;
> diff --git a/xen/arch/x86/mm/p2m.c b/xen/arch/x86/mm/p2m.c
> index 7a2ba82..4772c86 100644
> --- a/xen/arch/x86/mm/p2m.c
> +++ b/xen/arch/x86/mm/p2m.c
> @@ -1321,7 +1321,8 @@ static int set_typed_p2m_entry(struct domain *d, unsigned long gfn_l,
>  }
>  
>  /* Set foreign mfn in the given guest's p2m table. */
> -int set_foreign_p2m_entry(struct domain *d, unsigned long gfn, mfn_t mfn)
> +int set_foreign_p2m_entry(struct domain *d, const struct domain *fd,
> +                          unsigned long gfn, mfn_t mfn)
>  {
>      return set_typed_p2m_entry(d, gfn, mfn, PAGE_ORDER_4K, p2m_map_foreign,
>                                 p2m_get_hostp2m(d)->default_access);
> @@ -2621,7 +2622,7 @@ int p2m_add_foreign(struct domain *tdom, unsigned long fgfn,
>       * will update the m2p table which will result in  mfn -> gpfn of dom0
>       * and not fgfn of domU.
>       */
> -    rc = set_foreign_p2m_entry(tdom, gpfn, mfn);
> +    rc = set_foreign_p2m_entry(tdom, fdom, gpfn, mfn);
>      if ( rc )
>          gdprintk(XENLOG_WARNING, "set_foreign_p2m_entry failed. "
>                   "gpfn:%lx mfn:%lx fgfn:%lx td:%d fd:%d\n",
> diff --git a/xen/common/memory.c b/xen/common/memory.c
> index 3363c06..49e3001 100644
> --- a/xen/common/memory.c
> +++ b/xen/common/memory.c
> @@ -1134,12 +1134,8 @@ static int acquire_resource(
>      xen_pfn_t mfn_list[32];
>      int rc;
>  
> -    /*
> -     * FIXME: Until foreign pages inserted into the P2M are properly
> -     *        reference counted, it is unsafe to allow mapping of
> -     *        resource pages unless the caller is the hardware domain.
> -     */
> -    if ( paging_mode_translate(currd) && !is_hardware_domain(currd) )
> +    if ( paging_mode_translate(currd) && !is_hardware_domain(currd) &&
> +         !arch_acquire_resource_check() )
>          return -EACCES;
>  
>      if ( copy_from_guest(&xmar, arg, 1) )
> @@ -1207,7 +1203,7 @@ static int acquire_resource(
>  
>          for ( i = 0; !rc && i < xmar.nr_frames; i++ )
>          {
> -            rc = set_foreign_p2m_entry(currd, gfn_list[i],
> +            rc = set_foreign_p2m_entry(currd, d, gfn_list[i],
>                                         _mfn(mfn_list[i]));
>              /* rc should be -EIO for any iteration other than the first */
>              if ( rc && i )
> diff --git a/xen/include/asm-arm/p2m.h b/xen/include/asm-arm/p2m.h
> index 28ca9a8..4f8056e 100644
> --- a/xen/include/asm-arm/p2m.h
> +++ b/xen/include/asm-arm/p2m.h
> @@ -161,6 +161,15 @@ typedef enum {
>  #endif
>  #include <xen/p2m-common.h>
>  
> +static inline bool arch_acquire_resource_check(void)
> +{
> +    /*
> +     * The reference counting of foreign entries in set_foreign_p2m_entry()
> +     * is supported on Arm.
> +     */
> +    return true;
> +}
> +
>  static inline
>  void p2m_altp2m_check(struct vcpu *v, uint16_t idx)
>  {
> @@ -392,16 +401,6 @@ static inline gfn_t gfn_next_boundary(gfn_t gfn, unsigned int order)
>      return gfn_add(gfn, 1UL << order);
>  }
>  
> -static inline int set_foreign_p2m_entry(struct domain *d, unsigned long gfn,
> -                                        mfn_t mfn)
> -{
> -    /*
> -     * NOTE: If this is implemented then proper reference counting of
> -     *       foreign entries will need to be implemented.
> -     */
> -    return -EOPNOTSUPP;
> -}
> -
>  /*
>   * A vCPU has cache enabled only when the MMU is enabled and data cache
>   * is enabled.
> diff --git a/xen/include/asm-x86/p2m.h b/xen/include/asm-x86/p2m.h
> index 4603560..8d2dc22 100644
> --- a/xen/include/asm-x86/p2m.h
> +++ b/xen/include/asm-x86/p2m.h
> @@ -382,6 +382,19 @@ struct p2m_domain {
>  #endif
>  #include <xen/p2m-common.h>
>  
> +static inline bool arch_acquire_resource_check(void)
> +{
> +    /*
> +     * The reference counting of foreign entries in set_foreign_p2m_entry()
> +     * is not supported on x86.
> +     *
> +     * FIXME: Until foreign pages inserted into the P2M are properly
> +     * reference counted, it is unsafe to allow mapping of
> +     * resource pages unless the caller is the hardware domain.
> +     */
> +    return false;
> +}
> +
>  /*
>   * Updates vCPU's n2pm to match its np2m_base in VMCx12 and returns that np2m.
>   */
> @@ -647,9 +660,6 @@ int p2m_finish_type_change(struct domain *d,
>  int p2m_is_logdirty_range(struct p2m_domain *, unsigned long start,
>                            unsigned long end);
>  
> -/* Set foreign entry in the p2m table (for priv-mapping) */
> -int set_foreign_p2m_entry(struct domain *d, unsigned long gfn, mfn_t mfn);
> -
>  /* Set mmio addresses in the p2m table (for pass-through) */
>  int set_mmio_p2m_entry(struct domain *d, gfn_t gfn, mfn_t mfn,
>                         unsigned int order);
> diff --git a/xen/include/xen/p2m-common.h b/xen/include/xen/p2m-common.h
> index 58031a6..b4bc709 100644
> --- a/xen/include/xen/p2m-common.h
> +++ b/xen/include/xen/p2m-common.h
> @@ -3,6 +3,10 @@
>  
>  #include <xen/mm.h>
>  
> +/* Set foreign entry in the p2m table */
> +int set_foreign_p2m_entry(struct domain *d, const struct domain *fd,
> +                          unsigned long gfn, mfn_t mfn);
> +
>  /* Remove a page from a domain's p2m table */
>  int __must_check
>  guest_physmap_remove_page(struct domain *d, gfn_t gfn, mfn_t mfn,
> -- 
> 2.7.4
> 

  parent reply	other threads:[~2020-12-09 23:49 UTC|newest]

Thread overview: 127+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-11-30 10:31 Oleksandr Tyshchenko
2020-11-30 10:31 ` [PATCH V3 01/23] x86/ioreq: Prepare IOREQ feature for making it common Oleksandr Tyshchenko
2020-12-01 11:03   ` Alex Bennée
2020-12-01 18:53     ` Oleksandr
2020-12-01 19:36       ` Alex Bennée
2020-12-02  8:00       ` Jan Beulich
2020-12-02 11:19         ` Oleksandr
2020-12-07 11:13   ` Jan Beulich
2020-12-07 15:27     ` Oleksandr
2020-12-07 16:29       ` Jan Beulich
2020-12-07 17:21         ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 02/23] x86/ioreq: Add IOREQ_STATUS_* #define-s and update code for moving Oleksandr Tyshchenko
2020-12-01 11:07   ` Alex Bennée
2020-12-07 11:19   ` Jan Beulich
2020-12-07 15:37     ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 03/23] x86/ioreq: Provide out-of-line wrapper for the handle_mmio() Oleksandr Tyshchenko
2020-12-07 11:27   ` Jan Beulich
2020-12-07 15:39     ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 04/23] xen/ioreq: Make x86's IOREQ feature common Oleksandr Tyshchenko
2020-12-07 11:41   ` Jan Beulich
2020-12-07 19:43     ` Oleksandr
2020-12-08  9:21       ` Jan Beulich
2020-12-08 13:56         ` Oleksandr
2020-12-08 15:02           ` Jan Beulich
2020-12-08 17:24             ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 05/23] xen/ioreq: Make x86's hvm_ioreq_needs_completion() common Oleksandr Tyshchenko
2020-12-07 11:47   ` Jan Beulich
2020-11-30 10:31 ` [PATCH V3 06/23] xen/ioreq: Make x86's hvm_mmio_first(last)_byte() common Oleksandr Tyshchenko
2020-12-07 11:48   ` Jan Beulich
2020-11-30 10:31 ` [PATCH V3 07/23] xen/ioreq: Make x86's hvm_ioreq_(page/vcpu/server) structs common Oleksandr Tyshchenko
2020-12-07 11:54   ` Jan Beulich
2020-11-30 10:31 ` [PATCH V3 08/23] xen/ioreq: Move x86's ioreq_server to struct domain Oleksandr Tyshchenko
2020-12-07 12:04   ` Jan Beulich
2020-12-07 12:12     ` Paul Durrant
2020-12-07 19:52     ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 09/23] xen/dm: Make x86's DM feature common Oleksandr Tyshchenko
2020-12-07 12:08   ` Jan Beulich
2020-12-07 20:23     ` Oleksandr
2020-12-08  9:30       ` Jan Beulich
2020-12-08 14:54         ` Oleksandr
2021-01-07 14:38           ` Oleksandr
2021-01-07 15:01             ` Jan Beulich
2021-01-07 16:49               ` Oleksandr
2021-01-12 22:23                 ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 10/23] xen/mm: Make x86's XENMEM_resource_ioreq_server handling common Oleksandr Tyshchenko
2020-12-07 11:35   ` Jan Beulich
2020-12-07 12:11     ` Jan Beulich
2020-12-07 21:06       ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 11/23] xen/ioreq: Move x86's io_completion/io_req fields to struct vcpu Oleksandr Tyshchenko
2020-12-07 12:32   ` Jan Beulich
2020-12-07 20:59     ` Oleksandr
2020-12-08  7:52       ` Paul Durrant
2020-12-08  9:35         ` Jan Beulich
2020-12-08 18:21         ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 12/23] xen/ioreq: Remove "hvm" prefixes from involved function names Oleksandr Tyshchenko
2020-12-07 12:45   ` Jan Beulich
2020-12-07 20:28     ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 13/23] xen/ioreq: Use guest_cmpxchg64() instead of cmpxchg() Oleksandr Tyshchenko
2020-12-09 21:32   ` Stefano Stabellini
2020-12-09 22:34     ` Oleksandr
2020-12-10  2:30       ` Stefano Stabellini
2020-11-30 10:31 ` [PATCH V3 14/23] arm/ioreq: Introduce arch specific bits for IOREQ/DM features Oleksandr Tyshchenko
2020-12-09 22:04   ` Stefano Stabellini
2020-12-09 22:49     ` Oleksandr
2020-12-10  2:30       ` Stefano Stabellini
2020-11-30 10:31 ` [PATCH V3 15/23] xen/arm: Stick around in leave_hypervisor_to_guest until I/O has completed Oleksandr Tyshchenko
2020-11-30 20:51   ` Volodymyr Babchuk
2020-12-01 12:46     ` Julien Grall
2020-12-09 23:18   ` Stefano Stabellini
2020-12-09 23:35     ` Stefano Stabellini
2020-12-09 23:47       ` Julien Grall
2020-12-10  2:30         ` Stefano Stabellini
2020-12-10 13:17           ` Julien Grall
2020-12-10 13:21           ` Oleksandr
2020-12-09 23:38     ` Julien Grall
2020-11-30 10:31 ` [PATCH V3 16/23] xen/mm: Handle properly reference in set_foreign_p2m_entry() on Arm Oleksandr Tyshchenko
2020-12-08 14:24   ` Jan Beulich
2020-12-08 16:41     ` Oleksandr
2020-12-09 23:49   ` Stefano Stabellini [this message]
2021-01-15  1:18   ` Stefano Stabellini
2020-11-30 10:31 ` [PATCH V3 17/23] xen/ioreq: Introduce domain_has_ioreq_server() Oleksandr Tyshchenko
2020-12-08 15:11   ` Jan Beulich
2020-12-08 15:33     ` Oleksandr
2020-12-08 16:56       ` Oleksandr
2020-12-08 19:43         ` Paul Durrant
2020-12-08 20:16           ` Oleksandr
2020-12-09  9:01             ` Paul Durrant
2020-12-09 18:58               ` Julien Grall
2020-12-09 21:05                 ` Oleksandr
2020-12-09 20:36               ` Oleksandr
2020-12-10  8:38                 ` Paul Durrant
2020-12-10 16:57                   ` Oleksandr
2020-11-30 10:31 ` [PATCH V3 18/23] xen/dm: Introduce xendevicemodel_set_irq_level DM op Oleksandr Tyshchenko
2020-12-10  2:21   ` Stefano Stabellini
2020-12-10 12:58     ` Oleksandr
2020-12-10 13:38     ` Julien Grall
2020-11-30 10:31 ` [PATCH V3 19/23] xen/arm: io: Abstract sign-extension Oleksandr Tyshchenko
2020-11-30 21:03   ` Volodymyr Babchuk
2020-11-30 23:27     ` Oleksandr
2020-12-01  7:55       ` Jan Beulich
2020-12-01 10:30         ` Julien Grall
2020-12-01 10:42           ` Oleksandr
2020-12-01 12:13             ` Julien Grall
2020-12-01 12:24               ` Oleksandr
2020-12-01 12:28                 ` Julien Grall
2020-12-01 10:49           ` Jan Beulich
2020-12-01 10:23       ` Julien Grall
2020-11-30 10:31 ` [PATCH V3 20/23] xen/ioreq: Make x86's send_invalidate_req() common Oleksandr Tyshchenko
2020-12-08 15:24   ` Jan Beulich
2020-12-08 16:49     ` Oleksandr
2020-12-09  8:21       ` Jan Beulich
2020-11-30 10:31 ` [PATCH V3 21/23] xen/arm: Add mapcache invalidation handling Oleksandr Tyshchenko
2020-12-10  2:30   ` Stefano Stabellini
2020-12-10 18:50     ` Julien Grall
2020-12-11  1:28       ` Stefano Stabellini
2020-12-11 11:21         ` Oleksandr
2020-12-11 19:07           ` Stefano Stabellini
2020-12-11 19:37             ` Julien Grall
2020-12-11 19:27         ` Julien Grall
2020-11-30 10:31 ` [PATCH V3 22/23] libxl: Introduce basic virtio-mmio support on Arm Oleksandr Tyshchenko
2020-11-30 10:31 ` [PATCH V3 23/23] [RFC] libxl: Add support for virtio-disk configuration Oleksandr Tyshchenko
2020-11-30 11:22 ` [PATCH V3 00/23] IOREQ feature (+ virtio-mmio) on Arm Oleksandr
2020-12-07 13:03   ` Wei Chen
2020-12-07 21:03     ` Oleksandr
2020-11-30 16:21 ` Alex Bennée
2020-11-30 22:22   ` [PATCH V3 00/23] IOREQ feature (+ virtio-mmio) on Arm Oleksandr
2020-12-29 15:32   ` Roger Pau Monné

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=alpine.DEB.2.21.2012091549140.20986@sstabellini-ThinkPad-T480s \
    --to=sstabellini@kernel.org \
    --cc=Volodymyr_Babchuk@epam.com \
    --cc=andrew.cooper3@citrix.com \
    --cc=george.dunlap@citrix.com \
    --cc=iwj@xenproject.org \
    --cc=jbeulich@suse.com \
    --cc=julien.grall@arm.com \
    --cc=julien@xen.org \
    --cc=oleksandr_tyshchenko@epam.com \
    --cc=olekstysh@gmail.com \
    --cc=roger.pau@citrix.com \
    --cc=wl@xen.org \
    --cc=xen-devel@lists.xenproject.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).