Xen is heavily relying on the DCE stage to remove unused code so the
linker doesn't throw an error because a function is not implemented
yet we defined a prototype for it.

On some GCC versions (such as 9.4 provided by Debian sid), the compiler
DCE stage will not manage to figure that out for
xenmem_add_to_physmap_batch():

ld: ld: prelink.o: in function `xenmem_add_to_physmap_batch':
/xen/xen/common/memory.c:942: undefined reference to `xenmem_add_to_physmap_one'
/xen/xen/common/memory.c:942:(.text+0x22145): relocation truncated
to fit: R_X86_64_PLT32 against undefined symbol `xenmem_add_to_physmap_one'
prelink-efi.o: in function `xenmem_add_to_physmap_batch':
/xen/xen/common/memory.c:942: undefined reference to `xenmem_add_to_physmap_one'
make[2]: *** [Makefile:215: /root/xen/xen/xen.efi] Error 1
make[2]: *** Waiting for unfinished jobs....
ld: /xen/xen/.xen-syms.0: hidden symbol `xenmem_add_to_physmap_one' isn't defined
ld: final link failed: bad value

It is not entirely clear why the compiler DCE is not detecting the
unused code. However, cloning the check introduced by the commit below
into xenmem_add_to_physmap_batch() does the trick.

No functional change intended.

Fixes: d4f699a0df6c ("x86/mm: p2m_add_foreign() is HVM-only")
Reported-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
Signed-off-by: Julien Grall <jgrall@amazon.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Release-Acked-by: Ian Jackson <iwj@xenproject.org>
---
Julien, since I reused most of your patch'es description, I've kept your
S-o-b. Please let me know if you want me to drop it.

--- a/xen/common/memory.c
+++ b/xen/common/memory.c
@@ -904,6 +904,19 @@ static int xenmem_add_to_physmap_batch(s
 {
     union add_to_physmap_extra extra = {};
 
+    /*
+     * While, unlike xenmem_add_to_physmap(), this function is static, there
+     * still have been cases observed where xatp_permission_check(), invoked
+     * by our caller, doesn't lead to elimination of this entire function when
+     * the compile time evaluation of paging_mode_translate(d) is false. Guard
+     * against this be replicating the same check here.
+     */
+    if ( !paging_mode_translate(d) )
+    {
+        ASSERT_UNREACHABLE();
+        return -EACCES;
+    }
+
     if ( unlikely(xatpb->size < extent) )
         return -EILSEQ;
 

On 01/02/2021 16:20, Jan Beulich wrote:
> Xen is heavily relying on the DCE stage to remove unused code so the
> linker doesn't throw an error because a function is not implemented
> yet we defined a prototype for it.
>
> On some GCC versions (such as 9.4 provided by Debian sid), the compiler
> DCE stage will not manage to figure that out for
> xenmem_add_to_physmap_batch():
>
> ld: ld: prelink.o: in function `xenmem_add_to_physmap_batch':
> /xen/xen/common/memory.c:942: undefined reference to `xenmem_add_to_physmap_one'
> /xen/xen/common/memory.c:942:(.text+0x22145): relocation truncated
> to fit: R_X86_64_PLT32 against undefined symbol `xenmem_add_to_physmap_one'
> prelink-efi.o: in function `xenmem_add_to_physmap_batch':
> /xen/xen/common/memory.c:942: undefined reference to `xenmem_add_to_physmap_one'
> make[2]: *** [Makefile:215: /root/xen/xen/xen.efi] Error 1
> make[2]: *** Waiting for unfinished jobs....
> ld: /xen/xen/.xen-syms.0: hidden symbol `xenmem_add_to_physmap_one' isn't defined
> ld: final link failed: bad value
>
> It is not entirely clear why the compiler DCE is not detecting the
> unused code. However, cloning the check introduced by the commit below
> into xenmem_add_to_physmap_batch() does the trick.
>
> No functional change intended.
>
> Fixes: d4f699a0df6c ("x86/mm: p2m_add_foreign() is HVM-only")
> Reported-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
> Signed-off-by: Julien Grall <jgrall@amazon.com>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> Release-Acked-by: Ian Jackson <iwj@xenproject.org>
> ---
> Julien, since I reused most of your patch'es description, I've kept your
> S-o-b. Please let me know if you want me to drop it.
>
> --- a/xen/common/memory.c
> +++ b/xen/common/memory.c
> @@ -904,6 +904,19 @@ static int xenmem_add_to_physmap_batch(s
>  {
>      union add_to_physmap_extra extra = {};
>  
> +    /*
> +     * While, unlike xenmem_add_to_physmap(), this function is static, there
> +     * still have been cases observed where xatp_permission_check(), invoked
> +     * by our caller, doesn't lead to elimination of this entire function when
> +     * the compile time evaluation of paging_mode_translate(d) is false. Guard
> +     * against this be replicating the same check here.
> +     */

Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>, but I feel this
comment can be far more precise/concise.

/* In some configurations, (!HVM, COVERAGE), the
xenmem_add_to_physmap_one() call doesn't succumb to
dead-code-elimination.  Duplicate the short-circut from
xatp_permission_check() to try and help the compiler out. */

~Andrew

On 01.02.2021 19:26, Andrew Cooper wrote:
> On 01/02/2021 16:20, Jan Beulich wrote:
>> Xen is heavily relying on the DCE stage to remove unused code so the
>> linker doesn't throw an error because a function is not implemented
>> yet we defined a prototype for it.
>>
>> On some GCC versions (such as 9.4 provided by Debian sid), the compiler
>> DCE stage will not manage to figure that out for
>> xenmem_add_to_physmap_batch():
>>
>> ld: ld: prelink.o: in function `xenmem_add_to_physmap_batch':
>> /xen/xen/common/memory.c:942: undefined reference to `xenmem_add_to_physmap_one'
>> /xen/xen/common/memory.c:942:(.text+0x22145): relocation truncated
>> to fit: R_X86_64_PLT32 against undefined symbol `xenmem_add_to_physmap_one'
>> prelink-efi.o: in function `xenmem_add_to_physmap_batch':
>> /xen/xen/common/memory.c:942: undefined reference to `xenmem_add_to_physmap_one'
>> make[2]: *** [Makefile:215: /root/xen/xen/xen.efi] Error 1
>> make[2]: *** Waiting for unfinished jobs....
>> ld: /xen/xen/.xen-syms.0: hidden symbol `xenmem_add_to_physmap_one' isn't defined
>> ld: final link failed: bad value
>>
>> It is not entirely clear why the compiler DCE is not detecting the
>> unused code. However, cloning the check introduced by the commit below
>> into xenmem_add_to_physmap_batch() does the trick.
>>
>> No functional change intended.
>>
>> Fixes: d4f699a0df6c ("x86/mm: p2m_add_foreign() is HVM-only")
>> Reported-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
>> Signed-off-by: Julien Grall <jgrall@amazon.com>
>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>> Release-Acked-by: Ian Jackson <iwj@xenproject.org>
>> ---
>> Julien, since I reused most of your patch'es description, I've kept your
>> S-o-b. Please let me know if you want me to drop it.
>>
>> --- a/xen/common/memory.c
>> +++ b/xen/common/memory.c
>> @@ -904,6 +904,19 @@ static int xenmem_add_to_physmap_batch(s
>>  {
>>      union add_to_physmap_extra extra = {};
>>  
>> +    /*
>> +     * While, unlike xenmem_add_to_physmap(), this function is static, there
>> +     * still have been cases observed where xatp_permission_check(), invoked
>> +     * by our caller, doesn't lead to elimination of this entire function when
>> +     * the compile time evaluation of paging_mode_translate(d) is false. Guard
>> +     * against this be replicating the same check here.
>> +     */
> 
> Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>,

Thanks.

> but I feel this
> comment can be far more precise/concise.
> 
> /* In some configurations, (!HVM, COVERAGE), the
> xenmem_add_to_physmap_one() call doesn't succumb to
> dead-code-elimination.  Duplicate the short-circut from
> xatp_permission_check() to try and help the compiler out. */

I'm perfectly fine to take this one. I have to admit though that
I first needed to look up "succumb" ...

Jan

On Mon, Feb 01, 2021 at 05:20:54PM +0100, Jan Beulich wrote:
> Xen is heavily relying on the DCE stage to remove unused code so the
> linker doesn't throw an error because a function is not implemented
> yet we defined a prototype for it.
> 
> On some GCC versions (such as 9.4 provided by Debian sid), the compiler
> DCE stage will not manage to figure that out for
> xenmem_add_to_physmap_batch():
> 
> ld: ld: prelink.o: in function `xenmem_add_to_physmap_batch':
> /xen/xen/common/memory.c:942: undefined reference to `xenmem_add_to_physmap_one'
> /xen/xen/common/memory.c:942:(.text+0x22145): relocation truncated
> to fit: R_X86_64_PLT32 against undefined symbol `xenmem_add_to_physmap_one'
> prelink-efi.o: in function `xenmem_add_to_physmap_batch':
> /xen/xen/common/memory.c:942: undefined reference to `xenmem_add_to_physmap_one'
> make[2]: *** [Makefile:215: /root/xen/xen/xen.efi] Error 1
> make[2]: *** Waiting for unfinished jobs....
> ld: /xen/xen/.xen-syms.0: hidden symbol `xenmem_add_to_physmap_one' isn't defined
> ld: final link failed: bad value
> 
> It is not entirely clear why the compiler DCE is not detecting the
> unused code. However, cloning the check introduced by the commit below
> into xenmem_add_to_physmap_batch() does the trick.
> 
> No functional change intended.
> 
> Fixes: d4f699a0df6c ("x86/mm: p2m_add_foreign() is HVM-only")
> Reported-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@epam.com>
> Signed-off-by: Julien Grall <jgrall@amazon.com>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> Release-Acked-by: Ian Jackson <iwj@xenproject.org>

Reviewed-by: Wei Liu <wl@xen.org>