| Message ID | 20220826160503.1576966-3-zhiquan1.li@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | x86/sgx: fine grained SGX MCA behavior | expand |
On Sat, Aug 27, 2022 at 12:05:02AM +0800, Zhiquan Li wrote: > When a page triggers a machine check, it only reports the PFN. But in > order to inject #MC into hypervisor, the virtual address is required. > The 'encl_owner' field is useless in virtualization case, then > repurpose it as 'vepc_vaddr' - the virtual address of the virtual EPC > page for such case so that arch_memory_failure() can easily retrieve it. > > Introduce a union to prevent adding a new dedicated structure to > track the virtual address of virtual EPC page. And it can also prevent > playing the casting games while using it. > > Add a new EPC page flag - SGX_EPC_PAGE_KVM_GUEST to interpret the > meaning of the field. > > Co-developed-by: Cathy Zhang <cathy.zhang@intel.com> > Signed-off-by: Cathy Zhang <cathy.zhang@intel.com> > Signed-off-by: Zhiquan Li <zhiquan1.li@intel.com> > Acked-by: Kai Huang <kai.huang@intel.com> > Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org> > > --- > Changes since V5: > - To prevent casting the 'encl_owner' field, introduce a union with > another field - 'vepc_vaddr', sugguested by Dave Hansen. > - Add Reviewed-by from Jarkko. > Link: https://lore.kernel.org/linux-sgx/Yrf27fugD7lkyaek@kernel.org/T/#m379d00fc7f1d43726a42b3884637532061a8c0d1 > > Changes since V4: > - Add Co-developed-by and Signed-off-by from Cathy Zhang, as she had > fully discussed the flag name with Jarkko. > Link: https://lore.kernel.org/all/df92395ade424401ac3c6322de568720@intel.com/ > - Add Acked-by from Kai Huang > Link: https://lore.kernel.org/linux-sgx/0676cd4e-d94b-e904-81ae-ca1c05d37070@intel.com/T/#mccfb11df30698dbd060f2b6f06383cda7f154ef3 > > Changes since V3: > - Take the definition of EPC page flag SGX_EPC_PAGE_KVM_GUEST from > Cathy Zhang's third patch of SGX rebootless recovery patch set but > discard irrelevant portion, since it might need some time to re-forge > and these are two different features. > Link: https://lore.kernel.org/linux-sgx/41704e5d4c03b49fcda12e695595211d950cfb08.camel@kernel.org/T/#m9782d23496cacecb7da07a67daa79f4b322ae170 > > Changes since V2: > - Remove struct sgx_vepc_page and relevant code. > - Rework the patch suggested by Jarkko. > - Remove new EPC page flag SGX_EPC_PAGE_IS_VEPC definition as it is > duplicated to SGX_EPC_PAGE_KVM_GUEST. > Link: https://lore.kernel.org/linux-sgx/eb95b32ecf3d44a695610cf7f2816785@intel.com/T/#u > > Changes since V1: > - Add documentation suggested by Jarkko. > --- > arch/x86/kernel/cpu/sgx/main.c | 4 ++++ > arch/x86/kernel/cpu/sgx/sgx.h | 8 +++++++- > arch/x86/kernel/cpu/sgx/virt.c | 4 +++- > 3 files changed, 14 insertions(+), 2 deletions(-) > > diff --git a/arch/x86/kernel/cpu/sgx/main.c b/arch/x86/kernel/cpu/sgx/main.c > index 1315c69a733e..b319bedcaf1e 100644 > --- a/arch/x86/kernel/cpu/sgx/main.c > +++ b/arch/x86/kernel/cpu/sgx/main.c > @@ -549,6 +549,10 @@ int sgx_unmark_page_reclaimable(struct sgx_epc_page *page) > * Finally, wake up ksgxd when the number of pages goes below the watermark > * before returning back to the caller. > * > + * When an EPC page is assigned to KVM guest, repurpose the 'encl_owner' field > + * as the virtual address of virtual EPC page, since it is useless in such > + * scenario, so 'owner' is assigned to 'vepc_vaddr'. > + * > * Return: > * an EPC page, > * -errno on error > diff --git a/arch/x86/kernel/cpu/sgx/sgx.h b/arch/x86/kernel/cpu/sgx/sgx.h > index 4d88abccd12e..d16a8baa28d4 100644 > --- a/arch/x86/kernel/cpu/sgx/sgx.h > +++ b/arch/x86/kernel/cpu/sgx/sgx.h > @@ -28,12 +28,18 @@ > > /* Pages on free list */ > #define SGX_EPC_PAGE_IS_FREE BIT(1) > +/* Pages allocated for KVM guest */ > +#define SGX_EPC_PAGE_KVM_GUEST BIT(2) > > struct sgx_epc_page { > unsigned int section; > u16 flags; > u16 poison; > - struct sgx_encl_page *encl_owner; > + union { > + struct sgx_encl_page *encl_owner; > + /* Use when SGX_EPC_PAGE_KVM_GUEST set in ->flags: */ > + void __user *vepc_vaddr; Revisiting this, I'd tend to think that simple "void *owner" would be less complicated option than this. Thoughts? BR, Jarkko > + }; > struct list_head list; > }; > > diff --git a/arch/x86/kernel/cpu/sgx/virt.c b/arch/x86/kernel/cpu/sgx/virt.c > index 6a77a14eee38..776ae5c1c032 100644 > --- a/arch/x86/kernel/cpu/sgx/virt.c > +++ b/arch/x86/kernel/cpu/sgx/virt.c > @@ -46,10 +46,12 @@ static int __sgx_vepc_fault(struct sgx_vepc *vepc, > if (epc_page) > return 0; > > - epc_page = sgx_alloc_epc_page(vepc, false); > + epc_page = sgx_alloc_epc_page((void *)addr, false); > if (IS_ERR(epc_page)) > return PTR_ERR(epc_page); > > + epc_page->flags |= SGX_EPC_PAGE_KVM_GUEST; > + > ret = xa_err(xa_store(&vepc->page_array, index, epc_page, GFP_KERNEL)); > if (ret) > goto err_free; > -- > 2.25.1 >
On 2022/8/28 05:47, Jarkko Sakkinen wrote: >> - struct sgx_encl_page *encl_owner; >> + union { >> + struct sgx_encl_page *encl_owner; >> + /* Use when SGX_EPC_PAGE_KVM_GUEST set in ->flags: */ >> + void __user *vepc_vaddr; > Revisiting this, I'd tend to think that simple > "void *owner" would be less complicated option > than this. > > Thoughts? > > BR, Jarkko > Thanks for your thought, Jarkko. The reason is in the reply of patch 01, let's discuss it there. Best Regards, Zhiquan
On Mon, Aug 29, 2022 at 09:27:08AM +0800, Zhiquan Li wrote: > > On 2022/8/28 05:47, Jarkko Sakkinen wrote: > >> - struct sgx_encl_page *encl_owner; > >> + union { > >> + struct sgx_encl_page *encl_owner; > >> + /* Use when SGX_EPC_PAGE_KVM_GUEST set in ->flags: */ > >> + void __user *vepc_vaddr; > > Revisiting this, I'd tend to think that simple > > "void *owner" would be less complicated option > > than this. > > > > Thoughts? > > > > BR, Jarkko > > > > Thanks for your thought, Jarkko. > > The reason is in the reply of patch 01, let's discuss it there. > > Best Regards, > Zhiquan OK, it's cool, the problem was in the commit message. BR, Jarkko
diff --git a/arch/x86/kernel/cpu/sgx/main.c b/arch/x86/kernel/cpu/sgx/main.c index 1315c69a733e..b319bedcaf1e 100644 --- a/arch/x86/kernel/cpu/sgx/main.c +++ b/arch/x86/kernel/cpu/sgx/main.c @@ -549,6 +549,10 @@ int sgx_unmark_page_reclaimable(struct sgx_epc_page *page) * Finally, wake up ksgxd when the number of pages goes below the watermark * before returning back to the caller. * + * When an EPC page is assigned to KVM guest, repurpose the 'encl_owner' field + * as the virtual address of virtual EPC page, since it is useless in such + * scenario, so 'owner' is assigned to 'vepc_vaddr'. + * * Return: * an EPC page, * -errno on error diff --git a/arch/x86/kernel/cpu/sgx/sgx.h b/arch/x86/kernel/cpu/sgx/sgx.h index 4d88abccd12e..d16a8baa28d4 100644 --- a/arch/x86/kernel/cpu/sgx/sgx.h +++ b/arch/x86/kernel/cpu/sgx/sgx.h @@ -28,12 +28,18 @@ /* Pages on free list */ #define SGX_EPC_PAGE_IS_FREE BIT(1) +/* Pages allocated for KVM guest */ +#define SGX_EPC_PAGE_KVM_GUEST BIT(2) struct sgx_epc_page { unsigned int section; u16 flags; u16 poison; - struct sgx_encl_page *encl_owner; + union { + struct sgx_encl_page *encl_owner; + /* Use when SGX_EPC_PAGE_KVM_GUEST set in ->flags: */ + void __user *vepc_vaddr; + }; struct list_head list; }; diff --git a/arch/x86/kernel/cpu/sgx/virt.c b/arch/x86/kernel/cpu/sgx/virt.c index 6a77a14eee38..776ae5c1c032 100644 --- a/arch/x86/kernel/cpu/sgx/virt.c +++ b/arch/x86/kernel/cpu/sgx/virt.c @@ -46,10 +46,12 @@ static int __sgx_vepc_fault(struct sgx_vepc *vepc, if (epc_page) return 0; - epc_page = sgx_alloc_epc_page(vepc, false); + epc_page = sgx_alloc_epc_page((void *)addr, false); if (IS_ERR(epc_page)) return PTR_ERR(epc_page); + epc_page->flags |= SGX_EPC_PAGE_KVM_GUEST; + ret = xa_err(xa_store(&vepc->page_array, index, epc_page, GFP_KERNEL)); if (ret) goto err_free;