| Message ID | 20210312231632.5666-1-sstabellini@kernel.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v3] xen: introduce XENFEAT_direct_mapped and XENFEAT_not_direct_mapped | expand |
Hi Stefano, On 12/03/2021 23:16, Stefano Stabellini wrote: > Introduce two feature flags to tell the domain whether it is > direct-mapped or not. It allows the guest kernel to make informed > decisions on things such as swiotlb-xen enablement. > > The introduction of both flags (XENFEAT_direct_mapped and > XENFEAT_not_direct_mapped) allows the guest kernel to avoid any > guesswork if one of the two is present, or fallback to the current > checks if neither of them is present. > > XENFEAT_direct_mapped is always set for not auto-translated guests. > > For auto-translated guests, only Dom0 on ARM is direct-mapped. Also, > see is_domain_direct_mapped() which refers to auto-translated guests: > xen/include/asm-arm/domain.h:is_domain_direct_mapped > xen/include/asm-x86/domain.h:is_domain_direct_mapped > > Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com> Reviewed-by: Julien Grall <jgrall@amazon.com> Cheers,
On 13.03.2021 00:16, Stefano Stabellini wrote: > Introduce two feature flags to tell the domain whether it is > direct-mapped or not. It allows the guest kernel to make informed > decisions on things such as swiotlb-xen enablement. > > The introduction of both flags (XENFEAT_direct_mapped and > XENFEAT_not_direct_mapped) allows the guest kernel to avoid any > guesswork if one of the two is present, or fallback to the current > checks if neither of them is present. > > XENFEAT_direct_mapped is always set for not auto-translated guests. > > For auto-translated guests, only Dom0 on ARM is direct-mapped. Also, > see is_domain_direct_mapped() which refers to auto-translated guests: > xen/include/asm-arm/domain.h:is_domain_direct_mapped > xen/include/asm-x86/domain.h:is_domain_direct_mapped > > Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com> > CC: jbeulich@suse.com > CC: andrew.cooper3@citrix.com > CC: julien@xen.org Any particular reason my previously given R-b isn't here? Jan
On Mon, 15 Mar 2021, Jan Beulich wrote: > On 13.03.2021 00:16, Stefano Stabellini wrote: > > Introduce two feature flags to tell the domain whether it is > > direct-mapped or not. It allows the guest kernel to make informed > > decisions on things such as swiotlb-xen enablement. > > > > The introduction of both flags (XENFEAT_direct_mapped and > > XENFEAT_not_direct_mapped) allows the guest kernel to avoid any > > guesswork if one of the two is present, or fallback to the current > > checks if neither of them is present. > > > > XENFEAT_direct_mapped is always set for not auto-translated guests. > > > > For auto-translated guests, only Dom0 on ARM is direct-mapped. Also, > > see is_domain_direct_mapped() which refers to auto-translated guests: > > xen/include/asm-arm/domain.h:is_domain_direct_mapped > > xen/include/asm-x86/domain.h:is_domain_direct_mapped > > > > Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com> > > CC: jbeulich@suse.com > > CC: andrew.cooper3@citrix.com > > CC: julien@xen.org > > Any particular reason my previously given R-b isn't here? Hi Jan, I reworded part of the comment in the public header, and I decided to err on the side of caution and not add your R-b given this change compared to the previous version.
On 15.03.2021 21:01, Stefano Stabellini wrote: > On Mon, 15 Mar 2021, Jan Beulich wrote: >> On 13.03.2021 00:16, Stefano Stabellini wrote: >>> Introduce two feature flags to tell the domain whether it is >>> direct-mapped or not. It allows the guest kernel to make informed >>> decisions on things such as swiotlb-xen enablement. >>> >>> The introduction of both flags (XENFEAT_direct_mapped and >>> XENFEAT_not_direct_mapped) allows the guest kernel to avoid any >>> guesswork if one of the two is present, or fallback to the current >>> checks if neither of them is present. >>> >>> XENFEAT_direct_mapped is always set for not auto-translated guests. >>> >>> For auto-translated guests, only Dom0 on ARM is direct-mapped. Also, >>> see is_domain_direct_mapped() which refers to auto-translated guests: >>> xen/include/asm-arm/domain.h:is_domain_direct_mapped >>> xen/include/asm-x86/domain.h:is_domain_direct_mapped >>> >>> Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com> >>> CC: jbeulich@suse.com >>> CC: andrew.cooper3@citrix.com >>> CC: julien@xen.org >> >> Any particular reason my previously given R-b isn't here? > > I reworded part of the comment in the public header, and I decided to > err on the side of caution and not add your R-b given this change > compared to the previous version. I see. FAOD, despite me not being overly happy with the "older Xen assumptions" part of the comment, feel free to add it back. Jan
On Tue, 16 Mar 2021, Jan Beulich wrote: > On 15.03.2021 21:01, Stefano Stabellini wrote: > > On Mon, 15 Mar 2021, Jan Beulich wrote: > >> On 13.03.2021 00:16, Stefano Stabellini wrote: > >>> Introduce two feature flags to tell the domain whether it is > >>> direct-mapped or not. It allows the guest kernel to make informed > >>> decisions on things such as swiotlb-xen enablement. > >>> > >>> The introduction of both flags (XENFEAT_direct_mapped and > >>> XENFEAT_not_direct_mapped) allows the guest kernel to avoid any > >>> guesswork if one of the two is present, or fallback to the current > >>> checks if neither of them is present. > >>> > >>> XENFEAT_direct_mapped is always set for not auto-translated guests. > >>> > >>> For auto-translated guests, only Dom0 on ARM is direct-mapped. Also, > >>> see is_domain_direct_mapped() which refers to auto-translated guests: > >>> xen/include/asm-arm/domain.h:is_domain_direct_mapped > >>> xen/include/asm-x86/domain.h:is_domain_direct_mapped > >>> > >>> Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com> > >>> CC: jbeulich@suse.com > >>> CC: andrew.cooper3@citrix.com > >>> CC: julien@xen.org > >> > >> Any particular reason my previously given R-b isn't here? > > > > I reworded part of the comment in the public header, and I decided to > > err on the side of caution and not add your R-b given this change > > compared to the previous version. > > I see. FAOD, despite me not being overly happy with the "older > Xen assumptions" part of the comment, feel free to add it back. Thank you! Julien, please go ahead and commit it to your for-next/4.16 branch when it is convenient.
Hi Stefano, On 17/03/2021 00:04, Stefano Stabellini wrote: > On Tue, 16 Mar 2021, Jan Beulich wrote: >> On 15.03.2021 21:01, Stefano Stabellini wrote: >>> On Mon, 15 Mar 2021, Jan Beulich wrote: >>>> On 13.03.2021 00:16, Stefano Stabellini wrote: >>>>> Introduce two feature flags to tell the domain whether it is >>>>> direct-mapped or not. It allows the guest kernel to make informed >>>>> decisions on things such as swiotlb-xen enablement. >>>>> >>>>> The introduction of both flags (XENFEAT_direct_mapped and >>>>> XENFEAT_not_direct_mapped) allows the guest kernel to avoid any >>>>> guesswork if one of the two is present, or fallback to the current >>>>> checks if neither of them is present. >>>>> >>>>> XENFEAT_direct_mapped is always set for not auto-translated guests. >>>>> >>>>> For auto-translated guests, only Dom0 on ARM is direct-mapped. Also, >>>>> see is_domain_direct_mapped() which refers to auto-translated guests: >>>>> xen/include/asm-arm/domain.h:is_domain_direct_mapped >>>>> xen/include/asm-x86/domain.h:is_domain_direct_mapped >>>>> >>>>> Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com> >>>>> CC: jbeulich@suse.com >>>>> CC: andrew.cooper3@citrix.com >>>>> CC: julien@xen.org >>>> >>>> Any particular reason my previously given R-b isn't here? >>> >>> I reworded part of the comment in the public header, and I decided to >>> err on the side of caution and not add your R-b given this change >>> compared to the previous version. >> >> I see. FAOD, despite me not being overly happy with the "older >> Xen assumptions" part of the comment, feel free to add it back. > > Thank you! > > Julien, please go ahead and commit it to your for-next/4.16 branch when > it is convenient. I have committed it to my for-next/4.16 branch. I will merge the branch after the tree has re-opened. Cheers,
diff --git a/xen/common/kernel.c b/xen/common/kernel.c index 7a345ae45e..d77756a81e 100644 --- a/xen/common/kernel.c +++ b/xen/common/kernel.c @@ -560,6 +560,10 @@ DO(xen_version)(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg) (1U << XENFEAT_hvm_callback_vector) | (has_pirq(d) ? (1U << XENFEAT_hvm_pirqs) : 0); #endif + if ( !paging_mode_translate(d) || is_domain_direct_mapped(d) ) + fi.submap |= (1U << XENFEAT_direct_mapped); + else + fi.submap |= (1U << XENFEAT_not_direct_mapped); break; default: return -EINVAL; diff --git a/xen/include/public/features.h b/xen/include/public/features.h index 1613b2aab8..9ee2f760ef 100644 --- a/xen/include/public/features.h +++ b/xen/include/public/features.h @@ -114,6 +114,20 @@ */ #define XENFEAT_linux_rsdp_unrestricted 15 +/* + * A direct-mapped (or 1:1 mapped) domain is a domain for which its + * local pages have gfn == mfn. If a domain is direct-mapped, + * XENFEAT_direct_mapped is set; otherwise XENFEAT_not_direct_mapped + * is set. + * + * If neither flag is set (e.g. older Xen releases) the assumptions are: + * - not auto_translated domains (x86 only) are always direct-mapped + * - on x86, auto_translated domains are not direct-mapped + * - on ARM, Dom0 is direct-mapped, DomUs are not + */ +#define XENFEAT_not_direct_mapped 16 +#define XENFEAT_direct_mapped 17 + #define XENFEAT_NR_SUBMAPS 1 #endif /* __XEN_PUBLIC_FEATURES_H__ */
Introduce two feature flags to tell the domain whether it is direct-mapped or not. It allows the guest kernel to make informed decisions on things such as swiotlb-xen enablement. The introduction of both flags (XENFEAT_direct_mapped and XENFEAT_not_direct_mapped) allows the guest kernel to avoid any guesswork if one of the two is present, or fallback to the current checks if neither of them is present. XENFEAT_direct_mapped is always set for not auto-translated guests. For auto-translated guests, only Dom0 on ARM is direct-mapped. Also, see is_domain_direct_mapped() which refers to auto-translated guests: xen/include/asm-arm/domain.h:is_domain_direct_mapped xen/include/asm-x86/domain.h:is_domain_direct_mapped Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com> CC: jbeulich@suse.com CC: andrew.cooper3@citrix.com CC: julien@xen.org --- Changes in v3: - swap 2 conditions of || - improve in-code comment to cover existing assumptions Changes in v2: - code style improvements - better comments - better commit message - not auto_translated domains are direct_mapped --- xen/common/kernel.c | 4 ++++ xen/include/public/features.h | 14 ++++++++++++++ 2 files changed, 18 insertions(+)