Message ID | 20220812191224.15561-1-haitao.huang@linux.intel.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | x86/sgx: handle VA page allocation failure for EAUG on PF. | expand |
Hi Haitao, > -----Original Message----- > From: Haitao Huang <haitao.huang@linux.intel.com> > Sent: Friday, August 12, 2022 12:12 PM > To: linux-sgx@vger.kernel.org; jarkko@kernel.org; Chatre, Reinette > <reinette.chatre@intel.com>; dave.hansen@linux.intel.com; Dhanraj, Vijay > <vijay.dhanraj@intel.com> > Subject: [PATCH] x86/sgx: handle VA page allocation failure for EAUG on PF. > > Return VM_FAULT_NOPAGE to allow the swapping thread to catch up. > > link: https://lore.kernel.org/all/20220804201456.33418-1- > vijay.dhanraj@intel.com/ > > Signed-off-by: Haitao Huang <haitao.huang@linux.intel.com> > --- > arch/x86/kernel/cpu/sgx/encl.c | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/arch/x86/kernel/cpu/sgx/encl.c b/arch/x86/kernel/cpu/sgx/encl.c > index 24c1bb8eb196..1f88f80c4477 100644 > --- a/arch/x86/kernel/cpu/sgx/encl.c > +++ b/arch/x86/kernel/cpu/sgx/encl.c > @@ -344,8 +344,11 @@ static vm_fault_t sgx_encl_eaug_page(struct > vm_area_struct *vma, > } > > va_page = sgx_encl_grow(encl, false); > - if (IS_ERR(va_page)) > - goto err_out_epc; > + if (IS_ERR(va_page)) { > + if (PTR_ERR(va_page) == -EBUSY) > + vmret = VM_FAULT_NOPAGE; > + goto err_out_epc; > + } > > if (va_page) > list_add(&va_page->list, &encl->va_pages); > -- > 2.25.1 After applying this patch, I don't see the allocation failure issue with `augment_via_eaccept_long` test case. Ran the test for about ~60 times and all the time it passes. Tested-by: Vijay Dhanraj <vijay.dhanraj@intel.com>
On Fri, Aug 12, 2022 at 12:12:24PM -0700, Haitao Huang wrote: > Return VM_FAULT_NOPAGE to allow the swapping thread to catch up. > > link: https://lore.kernel.org/all/20220804201456.33418-1-vijay.dhanraj@intel.com/ > > Signed-off-by: Haitao Huang <haitao.huang@linux.intel.com> > --- > arch/x86/kernel/cpu/sgx/encl.c | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/arch/x86/kernel/cpu/sgx/encl.c b/arch/x86/kernel/cpu/sgx/encl.c > index 24c1bb8eb196..1f88f80c4477 100644 > --- a/arch/x86/kernel/cpu/sgx/encl.c > +++ b/arch/x86/kernel/cpu/sgx/encl.c > @@ -344,8 +344,11 @@ static vm_fault_t sgx_encl_eaug_page(struct vm_area_struct *vma, > } > > va_page = sgx_encl_grow(encl, false); > - if (IS_ERR(va_page)) > - goto err_out_epc; > + if (IS_ERR(va_page)) { > + if (PTR_ERR(va_page) == -EBUSY) > + vmret = VM_FAULT_NOPAGE; > + goto err_out_epc; > + } > > if (va_page) > list_add(&va_page->list, &encl->va_pages); > -- > 2.25.1 > Indentation is probably not right, but the fix look legit. You need to use real tab-characters to fix that. Run the patch with scripts/checkpatch.pl and it will report you these issues. Vijay, can you check if this fixes the issue for you? BR, Jarkko
On Fri, Aug 12, 2022 at 10:41:30PM +0000, Dhanraj, Vijay wrote: > Hi Haitao, > > > -----Original Message----- > > From: Haitao Huang <haitao.huang@linux.intel.com> > > Sent: Friday, August 12, 2022 12:12 PM > > To: linux-sgx@vger.kernel.org; jarkko@kernel.org; Chatre, Reinette > > <reinette.chatre@intel.com>; dave.hansen@linux.intel.com; Dhanraj, Vijay > > <vijay.dhanraj@intel.com> > > Subject: [PATCH] x86/sgx: handle VA page allocation failure for EAUG on PF. > > > > Return VM_FAULT_NOPAGE to allow the swapping thread to catch up. > > > > link: https://lore.kernel.org/all/20220804201456.33418-1- > > vijay.dhanraj@intel.com/ > > > > Signed-off-by: Haitao Huang <haitao.huang@linux.intel.com> > > > > arch/x86/kernel/cpu/sgx/encl.c | 7 +++++-- > > 1 file changed, 5 insertions(+), 2 deletions(-) > > > > diff --git a/arch/x86/kernel/cpu/sgx/encl.c b/arch/x86/kernel/cpu/sgx/encl.c > > index 24c1bb8eb196..1f88f80c4477 100644 > > --- a/arch/x86/kernel/cpu/sgx/encl.c > > +++ b/arch/x86/kernel/cpu/sgx/encl.c > > @@ -344,8 +344,11 @@ static vm_fault_t sgx_encl_eaug_page(struct > > vm_area_struct *vma, > > } > > > > va_page = sgx_encl_grow(encl, false); > > - if (IS_ERR(va_page)) > > - goto err_out_epc; > > + if (IS_ERR(va_page)) { > > + if (PTR_ERR(va_page) == -EBUSY) > > + vmret = VM_FAULT_NOPAGE; > > + goto err_out_epc; > > + } > > > > if (va_page) > > list_add(&va_page->list, &encl->va_pages); > > -- > > 2.25.1 > > After applying this patch, I don't see the allocation failure issue with `augment_via_eaccept_long` test case. Ran the test for about ~60 times and all the time it passes. > Tested-by: Vijay Dhanraj <vijay.dhanraj@intel.com> Good to hear. Thank you. BR, Jarkko
On Sun, 14 Aug 2022 14:36:29 -0500, Jarkko Sakkinen <jarkko@kernel.org> wrote: > On Fri, Aug 12, 2022 at 12:12:24PM -0700, Haitao Huang wrote: >> Return VM_FAULT_NOPAGE to allow the swapping thread to catch up. >> >> link: >> https://lore.kernel.org/all/20220804201456.33418-1-vijay.dhanraj@intel.com/ >> >> Signed-off-by: Haitao Huang <haitao.huang@linux.intel.com> >> --- >> arch/x86/kernel/cpu/sgx/encl.c | 7 +++++-- >> 1 file changed, 5 insertions(+), 2 deletions(-) >> >> diff --git a/arch/x86/kernel/cpu/sgx/encl.c >> b/arch/x86/kernel/cpu/sgx/encl.c >> index 24c1bb8eb196..1f88f80c4477 100644 >> --- a/arch/x86/kernel/cpu/sgx/encl.c >> +++ b/arch/x86/kernel/cpu/sgx/encl.c >> @@ -344,8 +344,11 @@ static vm_fault_t sgx_encl_eaug_page(struct >> vm_area_struct *vma, >> } >> >> va_page = sgx_encl_grow(encl, false); >> - if (IS_ERR(va_page)) >> - goto err_out_epc; >> + if (IS_ERR(va_page)) { >> + if (PTR_ERR(va_page) == -EBUSY) >> + vmret = VM_FAULT_NOPAGE; >> + goto err_out_epc; >> + } >> >> if (va_page) >> list_add(&va_page->list, &encl->va_pages); >> -- >> 2.25.1 >> > > Indentation is probably not right, but the fix look legit. Fixed in updated patch. Can I add your "Reviewed-by" tag? Thanks Haitao
diff --git a/arch/x86/kernel/cpu/sgx/encl.c b/arch/x86/kernel/cpu/sgx/encl.c index 24c1bb8eb196..1f88f80c4477 100644 --- a/arch/x86/kernel/cpu/sgx/encl.c +++ b/arch/x86/kernel/cpu/sgx/encl.c @@ -344,8 +344,11 @@ static vm_fault_t sgx_encl_eaug_page(struct vm_area_struct *vma, } va_page = sgx_encl_grow(encl, false); - if (IS_ERR(va_page)) - goto err_out_epc; + if (IS_ERR(va_page)) { + if (PTR_ERR(va_page) == -EBUSY) + vmret = VM_FAULT_NOPAGE; + goto err_out_epc; + } if (va_page) list_add(&va_page->list, &encl->va_pages);
Return VM_FAULT_NOPAGE to allow the swapping thread to catch up. link: https://lore.kernel.org/all/20220804201456.33418-1-vijay.dhanraj@intel.com/ Signed-off-by: Haitao Huang <haitao.huang@linux.intel.com> --- arch/x86/kernel/cpu/sgx/encl.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)