On 27.08.2024 14:59, Andrew Cooper wrote:
> On 27/08/2024 1:39 pm, Roger Pau Monne wrote:
>> --- a/xen/arch/x86/dom0_build.c
>> +++ b/xen/arch/x86/dom0_build.c
>> @@ -612,7 +612,24 @@ int __init construct_dom0(struct domain *d, const
>> module_t *image,
>> if ( is_hvm_domain(d) )
>> rc = dom0_construct_pvh(d, image, image_headroom, initrd, cmdline);
>> else if ( is_pv_domain(d) )
>> + {
>> + /*
>> + * Temporarily clear SMAP in CR4 to allow user-accesses in
>> + * construct_dom0(). This saves a large number of corner cases
>> + * interactions with copy_from_user().
>> + */
>> + if ( boot_cpu_has(X86_FEATURE_XEN_SMAP) )
>> + {
>> + cr4_pv32_mask &= ~X86_CR4_SMAP;
>> + write_cr4(read_cr4() & ~X86_CR4_SMAP);
>> + }
>> rc = dom0_construct_pv(d, image, image_headroom, initrd, cmdline);
>> + if ( boot_cpu_has(X86_FEATURE_XEN_SMAP) )
>> + {
>> + write_cr4(read_cr4() | X86_CR4_SMAP);
>> + cr4_pv32_mask |= X86_CR4_SMAP;
>> + }
>> + }
>
> I hate to drag this on further still, but can this logic be move it into
> dom0_construct_pv() itself, rather than here?
Just to mention it: I'm fine with this in principle, as long as this won't
mean a pile of new goto-s in dom0_construct_pv(). If a new wrapper was
introduced (with the present function becoming static), I'd be okay.
Jan
> That way, it won't need moving again to make cr4_pv32_mask exist only in
> PV32 builds. (This step is somewhat tricky, so I'm not suggesting doing
> it in this patch.)
>
> ~Andrew
