xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Roger Pau Monne <roger.pau@citrix.com>, xen-devel@lists.xenproject.org
Cc: Wei Liu <wei.liu2@citrix.com>,
	Ian Jackson <ian.jackson@eu.citrix.com>,
	Ian Campbell <ian.campbell@citrix.com>,
	Jan Beulich <jbeulich@suse.com>,
	Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Subject: Re: [PATCH v5 26/28] libxc/xen: introduce a start info structure for HVMlite guests
Date: Fri, 21 Aug 2015 22:00:20 +0100	[thread overview]
Message-ID: <55D79164.6020205@citrix.com> (raw)
In-Reply-To: <1440176021-18910-27-git-send-email-roger.pau@citrix.com>

On 21/08/15 17:53, Roger Pau Monne wrote:
> This structure contains the physical address of the command line, as well as
> the physical address of the list of loaded modules. The physical address of
> this structure is passed to the guest at boot time in the %ebx register.
>
> Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
> Cc: Ian Jackson <ian.jackson@eu.citrix.com>
> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> Cc: Ian Campbell <ian.campbell@citrix.com>
> Cc: Wei Liu <wei.liu2@citrix.com>
> Cc: Jan Beulich <jbeulich@suse.com>
> Cc: Andrew Cooper <andrew.cooper3@citrix.com>

This looks much nicer than the previous method.  A few comments.

> ---
>  tools/libxc/xc_dom_x86.c | 60 +++++++++++++++++++++++++++++++++++++++++++++++-
>  xen/include/public/xen.h | 19 +++++++++++++++
>  2 files changed, 78 insertions(+), 1 deletion(-)
>
> diff --git a/tools/libxc/xc_dom_x86.c b/tools/libxc/xc_dom_x86.c
> index fb71717..3b40ba1 100644
> --- a/tools/libxc/xc_dom_x86.c
> +++ b/tools/libxc/xc_dom_x86.c
> @@ -560,7 +560,62 @@ static int alloc_magic_pages_hvm(struct xc_dom_image *dom)
>      xc_hvm_param_set(xch, domid, HVM_PARAM_SHARING_RING_PFN,
>                       special_pfn(SPECIALPAGE_SHARING));
>  
> -    if ( dom->device_model )
> +    if ( !dom->device_model )
> +    {
> +        struct xc_dom_seg seg;
> +        struct hvm_start_info *start_info;
> +        char *cmdline;
> +        struct hvm_modlist_entry *modlist;
> +        void *start_page;
> +        size_t cmdline_size = dom->cmdline ? strlen(dom->cmdline) + 1 : 0;
> +        size_t start_info_size = sizeof(*start_info) + cmdline_size +

I would round cmdline_size up to the nearest 8, to align the modlist.

> +                                 dom->ramdisk_blob ? sizeof(*modlist)

sizeof(*modlist) * nr_entries.

This currently works because there is only ever one extra blob, but will
be wrong in the future when we want to pass other blobs.  (I have some
plans to fix the current requirements to build seabios into hvmloader,
and make it possible to pass things like NVRAM blobs.)

>  : 0;

I would suggest:

size_t start_info_size = sizeof(*start_info);
if ( dom->cmdline )
    start_info_size += ROUNDUP(strlen(dom->cmdline) + 1, 8);
if ( dom->ramdisk_blob )
    start_info_size += sizeof(*modlist); /* Only one module at the
moment. */

As a clearer way of expressing the logic.

> +
> +        rc = xc_dom_alloc_segment(dom, &seg, "HVMlite start info", 0,
> +                                  start_info_size);
> +        if ( rc != 0 )
> +        {
> +            DOMPRINTF("Unable to reserve memory for the start info");
> +            goto out;
> +        }
> +
> +        start_page = xc_map_foreign_range(xch, domid, start_info_size,
> +                                          PROT_READ | PROT_WRITE,
> +                                          seg.pfn);
> +        if ( start_page == NULL )
> +        {
> +            DOMPRINTF("Unable to map HVM start info page");
> +            goto error_out;
> +        }
> +
> +        start_info = start_page;
> +        cmdline = start_page + sizeof(*start_info);
> +        modlist = start_page + sizeof(*start_info) + cmdline_size;
> +
> +        if ( dom->cmdline )
> +        {
> +            strncpy(cmdline, dom->cmdline, MAX_GUEST_CMDLINE);
> +            cmdline[MAX_GUEST_CMDLINE - 1] = '\0';
> +            start_info->cmdline_paddr = (seg.pfn << PAGE_SHIFT) +
> +                                        sizeof(*start_info);

As the layout starts to get more complicated, this style of calculation
would be cleaner as

(seg.pfn << PAGE_SHIFT) + (cmdline - start_info);

> +        }
> +
> +        if ( dom->ramdisk_blob )
> +        {
> +            modlist[0].paddr = dom->ramdisk_seg.vstart - dom->parms.virt_base;
> +            modlist[0].size = dom->ramdisk_seg.vend - dom->ramdisk_seg.vstart;
> +            start_info->modlist_paddr = (seg.pfn << PAGE_SHIFT) +
> +                                        sizeof(*start_info) + cmdline_size;
> +            start_info->nr_modules = 1;
> +        }
> +
> +        start_info->magic = HVM_START_MAGIC_VALUE;
> +
> +        munmap(start_page, start_info_size);
> +
> +        dom->start_info_pfn = seg.pfn;
> +    }
> +    else
>      {
>          /*
>           * Allocate and clear additional ioreq server pages. The default
> @@ -916,6 +971,9 @@ static int vcpu_hvm(struct xc_dom_image *dom)
>      /* Set the IP. */
>      bsp_ctx.cpu.rip = dom->parms.phys_entry;
>  
> +    if ( dom->start_info_pfn )
> +        bsp_ctx.cpu.rbx = dom->start_info_pfn << PAGE_SHIFT;
> +
>      /* Set the end descriptor. */
>      bsp_ctx.end_d.typecode = HVM_SAVE_CODE(END);
>      bsp_ctx.end_d.instance = 0;
> diff --git a/xen/include/public/xen.h b/xen/include/public/xen.h
> index ff5547e..ca6a066 100644
> --- a/xen/include/public/xen.h
> +++ b/xen/include/public/xen.h
> @@ -784,6 +784,25 @@ struct start_info {
>  };
>  typedef struct start_info start_info_t;
>  
> +/*
> + * Start of day structure passed to HVMlite guests in %ebx.

As we are planning to rename HVMlite to PVH, I would avoid the use of
HVMlite in committed code.

Also as this is HVM specific, might it be better in public/hvm?  I guess
that depends on whether ARM is likely to use it.

~Andrew

> + */
> +struct hvm_start_info {
> +#define HVM_START_MAGIC_VALUE 0x336ec578
> +    uint32_t magic;             /* Contains the magic value 0x336ec578       */
> +                                /* ("xEn3" with the 0x80 bit of the "E" set).*/
> +    uint32_t flags;             /* SIF_xxx flags.                            */
> +    uint32_t cmdline_paddr;     /* Physical address of the command line.     */
> +    uint32_t nr_modules;        /* Number of modules passed to the kernel.   */
> +    uint32_t modlist_paddr;     /* Physical address of an array of           */
> +                                /* hvm_modlist_entry.                        */
> +};
> +
> +struct hvm_modlist_entry {
> +    uint64_t paddr;             /* Physical address of the module.           */
> +    uint64_t size;              /* Size of the module in bytes.              */
> +};
> +
>  /* New console union for dom0 introduced in 0x00030203. */
>  #if __XEN_INTERFACE_VERSION__ < 0x00030203
>  #define console_mfn    console.domU.mfn


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

  reply	other threads:[~2015-08-21 21:00 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-08-21 16:53 [PATCH v5 00/28] Introduce HVM without dm and new boot ABI Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 01/28] libxc: split x86 HVM setup_guest into smaller logical functions Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 02/28] libxc: unify xc_dom_p2m_{host/guest} Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 03/28] libxc: introduce the notion of a container type Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 04/28] libxc: introduce a domain loader for HVM guest firmware Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 05/28] libxc: make arch_setup_meminit a xc_dom_arch hook Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 06/28] libxc: make arch_setup_boot{init/late} xc_dom_arch hooks Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 07/28] libxc: rework BSP initialization Roger Pau Monne
2015-08-24 18:26   ` Konrad Rzeszutek Wilk
2015-08-25  8:33     ` Roger Pau Monné
2015-08-25  9:02   ` Wei Liu
2015-08-25  9:22     ` Roger Pau Monné
2015-08-25  9:32       ` Wei Liu
2015-08-25  9:35       ` Andrew Cooper
2015-08-21 16:53 ` [PATCH v5 08/28] libxc: introduce a xc_dom_arch for hvm-3.0-x86_32 guests Roger Pau Monne
2015-08-25  9:17   ` Wei Liu
2015-08-21 16:53 ` [PATCH v5 09/28] libxl: switch HVM domain building to use xc_dom_* helpers Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 10/28] libxc: remove dead HVM building code Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 11/28] xen/x86: add bitmap of enabled emulated devices Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 12/28] xen/x86: allow disabling the emulated local apic Roger Pau Monne
2015-08-31 21:55   ` Boris Ostrovsky
2015-08-21 16:53 ` [PATCH v5 13/28] xen/x86: allow disabling the emulated HPET Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 14/28] xen/x86: allow disabling the pmtimer Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 15/28] xen/x86: allow disabling the emulated RTC Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 16/28] xen/x86: allow disabling the emulated IO APIC Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 17/28] xen/x86: allow disabling the emulated PIC Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 18/28] xen/x86: allow disabling the emulated pmu Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 19/28] xen/x86: allow disabling the emulated VGA Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 20/28] xen/x86: allow disabling the emulated IOMMU Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 21/28] xen/x86: allow disabling all emulated devices inside of Xen Roger Pau Monne
2015-08-21 17:19   ` Andrew Cooper
2015-08-21 16:53 ` [PATCH v5 22/28] elfnotes: intorduce a new PHYS_ENTRY elfnote Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 23/28] libxc: allow creating domains without emulated devices Roger Pau Monne
2015-08-21 17:42   ` Andrew Cooper
2015-08-21 16:53 ` [PATCH v5 24/28] xen/x86: allow HVM guests to use hypercalls to bring up vCPUs Roger Pau Monne
2015-08-21 20:36   ` Andrew Cooper
2015-08-24 10:43     ` Roger Pau Monné
2015-08-24 11:57       ` Andrew Cooper
2015-08-21 16:53 ` [PATCH v5 25/28] xenconsole: try to attach to PV console if HVM fails Roger Pau Monne
2015-08-21 16:53 ` [PATCH v5 26/28] libxc/xen: introduce a start info structure for HVMlite guests Roger Pau Monne
2015-08-21 21:00   ` Andrew Cooper [this message]
2015-08-24 15:24     ` Roger Pau Monné
2015-08-21 16:53 ` [PATCH v5 27/28] libxc: switch xc_dom_elfloader to be used with HVMlite domains Roger Pau Monne
2015-08-25  9:32   ` Wei Liu
2015-08-21 16:53 ` [PATCH v5 28/28] libxl: allow the creation of HVM domains without a device model Roger Pau Monne
2015-08-25  9:30   ` Wei Liu

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=55D79164.6020205@citrix.com \
    --to=andrew.cooper3@citrix.com \
    --cc=ian.campbell@citrix.com \
    --cc=ian.jackson@eu.citrix.com \
    --cc=jbeulich@suse.com \
    --cc=roger.pau@citrix.com \
    --cc=stefano.stabellini@eu.citrix.com \
    --cc=wei.liu2@citrix.com \
    --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).