[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH v2 2/2] xen/efi: optionally call SetVirtualAddressMap()


  • To: Marek Marczykowski-Górecki <marmarek@xxxxxxxxxxxxxxxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxxx>
  • From: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
  • Date: Sat, 12 Oct 2019 17:29:34 +0100
  • Authentication-results: esa5.hc3370-68.iphmx.com; dkim=none (message not signed) header.i=none; spf=None smtp.pra=andrew.cooper3@xxxxxxxxxx; spf=Pass smtp.mailfrom=Andrew.Cooper3@xxxxxxxxxx; spf=None smtp.helo=postmaster@xxxxxxxxxxxxxxx
  • Autocrypt: addr=andrew.cooper3@xxxxxxxxxx; prefer-encrypt=mutual; keydata= mQINBFLhNn8BEADVhE+Hb8i0GV6mihnnr/uiQQdPF8kUoFzCOPXkf7jQ5sLYeJa0cQi6Penp VtiFYznTairnVsN5J+ujSTIb+OlMSJUWV4opS7WVNnxHbFTPYZVQ3erv7NKc2iVizCRZ2Kxn srM1oPXWRic8BIAdYOKOloF2300SL/bIpeD+x7h3w9B/qez7nOin5NzkxgFoaUeIal12pXSR Q354FKFoy6Vh96gc4VRqte3jw8mPuJQpfws+Pb+swvSf/i1q1+1I4jsRQQh2m6OTADHIqg2E ofTYAEh7R5HfPx0EXoEDMdRjOeKn8+vvkAwhviWXTHlG3R1QkbE5M/oywnZ83udJmi+lxjJ5 YhQ5IzomvJ16H0Bq+TLyVLO/VRksp1VR9HxCzItLNCS8PdpYYz5TC204ViycobYU65WMpzWe LFAGn8jSS25XIpqv0Y9k87dLbctKKA14Ifw2kq5OIVu2FuX+3i446JOa2vpCI9GcjCzi3oHV e00bzYiHMIl0FICrNJU0Kjho8pdo0m2uxkn6SYEpogAy9pnatUlO+erL4LqFUO7GXSdBRbw5 gNt25XTLdSFuZtMxkY3tq8MFss5QnjhehCVPEpE6y9ZjI4XB8ad1G4oBHVGK5LMsvg22PfMJ ISWFSHoF/B5+lHkCKWkFxZ0gZn33ju5n6/FOdEx4B8cMJt+cWwARAQABtClBbmRyZXcgQ29v cGVyIDxhbmRyZXcuY29vcGVyM0BjaXRyaXguY29tPokCOgQTAQgAJAIbAwULCQgHAwUVCgkI CwUWAgMBAAIeAQIXgAUCWKD95wIZAQAKCRBlw/kGpdefoHbdD/9AIoR3k6fKl+RFiFpyAhvO 59ttDFI7nIAnlYngev2XUR3acFElJATHSDO0ju+hqWqAb8kVijXLops0gOfqt3VPZq9cuHlh IMDquatGLzAadfFx2eQYIYT+FYuMoPZy/aTUazmJIDVxP7L383grjIkn+7tAv+qeDfE+txL4 SAm1UHNvmdfgL2/lcmL3xRh7sub3nJilM93RWX1Pe5LBSDXO45uzCGEdst6uSlzYR/MEr+5Z JQQ32JV64zwvf/aKaagSQSQMYNX9JFgfZ3TKWC1KJQbX5ssoX/5hNLqxMcZV3TN7kU8I3kjK mPec9+1nECOjjJSO/h4P0sBZyIUGfguwzhEeGf4sMCuSEM4xjCnwiBwftR17sr0spYcOpqET ZGcAmyYcNjy6CYadNCnfR40vhhWuCfNCBzWnUW0lFoo12wb0YnzoOLjvfD6OL3JjIUJNOmJy RCsJ5IA/Iz33RhSVRmROu+TztwuThClw63g7+hoyewv7BemKyuU6FTVhjjW+XUWmS/FzknSi dAG+insr0746cTPpSkGl3KAXeWDGJzve7/SBBfyznWCMGaf8E2P1oOdIZRxHgWj0zNr1+ooF /PzgLPiCI4OMUttTlEKChgbUTQ+5o0P080JojqfXwbPAyumbaYcQNiH1/xYbJdOFSiBv9rpt TQTBLzDKXok86LkCDQRS4TZ/ARAAkgqudHsp+hd82UVkvgnlqZjzz2vyrYfz7bkPtXaGb9H4 Rfo7mQsEQavEBdWWjbga6eMnDqtu+FC+qeTGYebToxEyp2lKDSoAsvt8w82tIlP/EbmRbDVn 7bhjBlfRcFjVYw8uVDPptT0TV47vpoCVkTwcyb6OltJrvg/QzV9f07DJswuda1JH3/qvYu0p vjPnYvCq4NsqY2XSdAJ02HrdYPFtNyPEntu1n1KK+gJrstjtw7KsZ4ygXYrsm/oCBiVW/OgU g/XIlGErkrxe4vQvJyVwg6YH653YTX5hLLUEL1NS4TCo47RP+wi6y+TnuAL36UtK/uFyEuPy wwrDVcC4cIFhYSfsO0BumEI65yu7a8aHbGfq2lW251UcoU48Z27ZUUZd2Dr6O/n8poQHbaTd 6bJJSjzGGHZVbRP9UQ3lkmkmc0+XCHmj5WhwNNYjgbbmML7y0fsJT5RgvefAIFfHBg7fTY/i kBEimoUsTEQz+N4hbKwo1hULfVxDJStE4sbPhjbsPCrlXf6W9CxSyQ0qmZ2bXsLQYRj2xqd1 bpA+1o1j2N4/au1R/uSiUFjewJdT/LX1EklKDcQwpk06Af/N7VZtSfEJeRV04unbsKVXWZAk uAJyDDKN99ziC0Wz5kcPyVD1HNf8bgaqGDzrv3TfYjwqayRFcMf7xJaL9xXedMcAEQEAAYkC HwQYAQgACQUCUuE2fwIbDAAKCRBlw/kGpdefoG4XEACD1Qf/er8EA7g23HMxYWd3FXHThrVQ HgiGdk5Yh632vjOm9L4sd/GCEACVQKjsu98e8o3ysitFlznEns5EAAXEbITrgKWXDDUWGYxd pnjj2u+GkVdsOAGk0kxczX6s+VRBhpbBI2PWnOsRJgU2n10PZ3mZD4Xu9kU2IXYmuW+e5KCA vTArRUdCrAtIa1k01sPipPPw6dfxx2e5asy21YOytzxuWFfJTGnVxZZSCyLUO83sh6OZhJkk b9rxL9wPmpN/t2IPaEKoAc0FTQZS36wAMOXkBh24PQ9gaLJvfPKpNzGD8XWR5HHF0NLIJhgg 4ZlEXQ2fVp3XrtocHqhu4UZR4koCijgB8sB7Tb0GCpwK+C4UePdFLfhKyRdSXuvY3AHJd4CP 4JzW0Bzq/WXY3XMOzUTYApGQpnUpdOmuQSfpV9MQO+/jo7r6yPbxT7CwRS5dcQPzUiuHLK9i nvjREdh84qycnx0/6dDroYhp0DFv4udxuAvt1h4wGwTPRQZerSm4xaYegEFusyhbZrI0U9tJ B8WrhBLXDiYlyJT6zOV2yZFuW47VrLsjYnHwn27hmxTC/7tvG3euCklmkn9Sl9IAKFu29RSo d5bD8kMSCYsTqtTfT6W4A3qHGvIDta3ptLYpIAOD2sY3GYq2nf3Bbzx81wZK14JdDDHUX2Rs 6+ahAA==
  • Cc: Stefano Stabellini <sstabellini@xxxxxxxxxx>, Wei Liu <wl@xxxxxxx>, Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>, George Dunlap <George.Dunlap@xxxxxxxxxxxxx>, Tim Deegan <tim@xxxxxxx>, Ian Jackson <ian.jackson@xxxxxxxxxxxxx>, Julien Grall <julien.grall@xxxxxxx>, Jan Beulich <jbeulich@xxxxxxxx>
  • Delivery-date: Sat, 12 Oct 2019 16:29:59 +0000
  • Ironport-sdr: RbIZ+tH2IKm29uuIstWYd3Nb8O/zHlwCOKpIBvKeeQNgCsonY3ElgJl8ifpaNrFje9dq4Brcb3 /N/FuRjsnln4kpmu5TZe/Ce0Q702aBEHdBc0rcR2gtDgLN5DyKZQRxo2GerIOwaH6W4LYvk1iA YkdJFeTaT1KGr9hnWJ5n27+z3fdFnS8GEHZYKioS6Wme8Kp+JupTcNuOqHikCPDXjadVVPcCwU FeijKYV30r3kirhRQRUuO4Zw9rdW1/gP4hXQ2PjsEu/+TLpZ+7mMrXQTUSgm78TM5dDLnS09hA 5M0=
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
  • Openpgp: preference=signencrypt

On 12/10/2019 15:36, Marek Marczykowski-Górecki wrote:
> SetVirtualAddressMap() can be called only once,

True.

> hence it's incompatible with kexec.

Most certainly not.

Linux unconditionally enters virtual mode, citing a huge slew of EFI
firmware bugs, and is perfectly capable of kexec-ing on the resulting
systems.

This is how Xen should behave as well, and I suspect it will have a
marked improvement on our ability to actually boot on EFI systems.


Now - it may be true that Xen is missing some piece of plumbing to allow
kexec in virtual mode to work, and that is a fine reason to leave a note
in the text of an EXPERT option noting what what is/isn't expected to
work (and what may or may not have been tested).

> For this reason, make it an optional feature, depending on
> !KEXEC.

This presupposes (at Xen's build time) that a kexec'd kernel is going to
want/need to use runtime services.  I'm not convinced this is
universally true, or a reasonable restriction to make, as kexec is the
action of last resort to try and get something useful out.  (However,
given the 4.13 timeline, and that this is off-by-default, lets not waste
time arguing, so it can stay as it is.)

> And to not inflate number of supported configurations, hide it
> behind EXPERT.

"number of supported configurations" isn't a relevant argument.  We will
have as few or as many as are appropriate to present to user, given a
baseline competency of "able to at read and comprehend the descriptions
given".

A valid reason for putting this behind EXPERT is because it is an
interim bit of duct tape, trying to work around other breakages in Xen,
and its late in the 4.13 dev cycle, and use of this option might cause
other things to explode in weird and wonderful ways.

> diff --git a/xen/common/Kconfig b/xen/common/Kconfig
> index 16829f6..fe98f8a 100644
> --- a/xen/common/Kconfig
> +++ b/xen/common/Kconfig
> @@ -88,6 +88,19 @@ config KEXEC
>  
>         If unsure, say Y.
>  
> +config SET_VIRTUAL_ADDRESS_MAP
> +    bool "EFI: call SetVirtualAddressMap()" if EXPERT = "y"
> +    default n
> +    depends on !KEXEC
> +    ---help---
> +      Call EFI SetVirtualAddressMap() runtime service to setup memory map for
> +      further runtime services. According to UEFI spec, it isn't strictly
> +      necessary, but many UEFI implementations misbehave when this call is
> +      missing. On the other hand, this call can be made only once, which 
> makes
> +      it incompatible with kexec (kexec-ing this Xen from other Xen or 
> Linux).
> +
> +      If unsuser, say N.

"unsure".

> +
>  config XENOPROF
>       def_bool y
>       prompt "Xen Oprofile Support" if EXPERT = "y"
> diff --git a/xen/common/efi/boot.c b/xen/common/efi/boot.c
> index cddf3de..6eaabd4 100644
> --- a/xen/common/efi/boot.c
> +++ b/xen/common/efi/boot.c
> @@ -1056,11 +1056,17 @@ static void __init 
> efi_set_gop_mode(EFI_GRAPHICS_OUTPUT_PROTOCOL *gop, UINTN gop
>          efi_arch_video_init(gop, info_size, mode_info);
>  }
>  
> +#define INVALID_VIRTUAL_ADDRESS (0xBAAADUL << \
> +                                 (EFI_PAGE_SHIFT + BITS_PER_LONG - 32))
> +
>  static void __init efi_exit_boot(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE 
> *SystemTable)
>  {
>      EFI_STATUS status;
>      UINTN info_size = 0, map_key;
>      bool retry;
> +#ifdef CONFIG_SET_VIRTUAL_ADDRESS_MAP
> +    unsigned int i;
> +#endif
>  
>      efi_bs->GetMemoryMap(&info_size, NULL, &map_key,
>                           &efi_mdesc_size, &mdesc_ver);
> @@ -1094,6 +1100,26 @@ static void __init efi_exit_boot(EFI_HANDLE 
> ImageHandle, EFI_SYSTEM_TABLE *Syste
>      if ( EFI_ERROR(status) )
>          PrintErrMesg(L"Cannot exit boot services", status);

Use this example...

>  
> +#ifdef CONFIG_SET_VIRTUAL_ADDRESS_MAP
> +    for ( i = 0; i < efi_memmap_size; i += efi_mdesc_size )
> +    {
> +        EFI_MEMORY_DESCRIPTOR *desc = efi_memmap + i;
> +
> +        if ( desc->Attribute & EFI_MEMORY_RUNTIME )
> +            desc->VirtualStart = desc->PhysicalStart;
> +        else
> +            desc->VirtualStart = INVALID_VIRTUAL_ADDRESS;
> +    }
> +    status = efi_rs->SetVirtualAddressMap(efi_memmap_size, efi_mdesc_size,
> +                                          mdesc_ver, efi_memmap);
> +    if ( status != EFI_SUCCESS )
> +    {
> +        printk(XENLOG_ERR "EFI: SetVirtualAddressMap() failed (%#lx), 
> disabling runtime services\n",
> +               status);

... here.  printk() isn't set up, and won't go anywhere useful.

With this, and a bit of rewording of the commit message and Kconfig
text, I think this is fine for inclusion into 4.13.  It is off by
default, so will not interfere with existing configuration, and it
clearly improves the status quo on others.

~Andrew

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.