| Message ID | 20210607024318.3988467-6-penny.zheng@arm.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Domain on Static Allocation | expand |
On 07.06.2021 04:43, Penny Zheng wrote: > Introduce new interface assign_pages_nr to deal with when page number is > not in a power-of-two, which will save the trouble each time user needs > to split the size in a power of 2 to use assign_pages. First of all I still don't see why in this one special case it is a meaningful burden to do the count-to-order conversion in the caller you mean to add, and hence why we really need this new function (to keep it simple, you could even have the caller not break down to arbitrary power-of-2 chunks, but simply iterate over all individual [order-0] pages). The more that I'm not happy with the chosen name, despite it having been suggested during v1 review. _If_ we needed two functions, imo they ought to be named assign_page() (dealing with a single page of the given order) and assign_pages(). Backporting confusion could be helped by altering the order of parameters, such that the compiler would point out that adjustments at call sites are needed. Irrespective of this a few remarks on the code change itself: > --- a/xen/common/page_alloc.c > +++ b/xen/common/page_alloc.c > @@ -2301,14 +2301,14 @@ void init_domheap_pages(paddr_t ps, paddr_t pe) > } > > > -int assign_pages( > +int assign_pages_nr( > struct domain *d, > struct page_info *pg, > - unsigned int order, > + unsigned int nr_pfns, Even leaving the naming aspect of "pfns" aside, I can't see why this can't be simply "nr" (of appropriate type, see next remark). > unsigned int memflags) > { > int rc = 0; > - unsigned long i; > + unsigned int i; This is not an acceptable type change, at least not as long as it's not justified at all in the description. While both Arm and x86 will be fine this way, the code here is supposed to be generic, and hence would better remain generally correct. > --- a/xen/include/xen/mm.h > +++ b/xen/include/xen/mm.h > @@ -131,12 +131,18 @@ int query_page_offline(mfn_t mfn, uint32_t *status); > > void heap_init_late(void); > > -int assign_pages( > +int assign_pages_nr( > struct domain *d, > struct page_info *pg, > - unsigned int order, > + unsigned int nr_pfns, > unsigned int memflags); > > + int assign_pages( > + struct domain *d, > + struct page_info *pg, > + unsigned int order, > + unsigned int memflags); Bogus extra leading space on all lines of this addition. Jan
Hi, On 10/06/2021 10:49, Jan Beulich wrote: > On 07.06.2021 04:43, Penny Zheng wrote: >> Introduce new interface assign_pages_nr to deal with when page number is >> not in a power-of-two, which will save the trouble each time user needs >> to split the size in a power of 2 to use assign_pages. > > First of all I still don't see why in this one special case it is a > meaningful burden to do the count-to-order conversion in the caller you > mean to add, This sort of works for one caller. However, I would expect some more user in the future (we use it for Live-Update). > and hence why we really need this new function (to keep it > simple, you could even have the caller not break down to arbitrary > power-of-2 chunks, but simply iterate over all individual [order-0] > pages). The function assign_pages() will always use 1U << order (and sadly 1 << order). So we would end up to convert the count in multiple order for then directly converting back to a number. To me, this sounds rather pointless... There are also a slight benefits to call assign_pages() a single time during boot because it will reduce the number of time we need to lock/unlock d->page_alloc_lock. > The more that I'm not happy with the chosen name, despite it > having been suggested during v1 review. _If_ we needed two functions, > imo they ought to be named assign_page() (dealing with a single page of > the given order) and assign_pages(). Backporting confusion could be > helped by altering the order of parameters, such that the compiler > would point out that adjustments at call sites are needed. > > Irrespective of this a few remarks on the code change itself: > >> --- a/xen/common/page_alloc.c >> +++ b/xen/common/page_alloc.c >> @@ -2301,14 +2301,14 @@ void init_domheap_pages(paddr_t ps, paddr_t pe) >> } >> >> >> -int assign_pages( >> +int assign_pages_nr( >> struct domain *d, >> struct page_info *pg, >> - unsigned int order, >> + unsigned int nr_pfns, > > Even leaving the naming aspect of "pfns" aside, I can't see why this > can't be simply "nr" (of appropriate type, see next remark). > >> unsigned int memflags) >> { >> int rc = 0; >> - unsigned long i; >> + unsigned int i; > > This is not an acceptable type change, at least not as long as it's not > justified at all in the description. While both Arm and x86 will be > fine this way, the code here is supposed to be generic, and hence would > better remain generally correct. I would like to point out the code is already not correct as we are using 1U << order or worse 1 << order :). Cheers,
On 30.06.2021 20:29, Julien Grall wrote: > On 10/06/2021 10:49, Jan Beulich wrote: >> On 07.06.2021 04:43, Penny Zheng wrote: >>> Introduce new interface assign_pages_nr to deal with when page number is >>> not in a power-of-two, which will save the trouble each time user needs >>> to split the size in a power of 2 to use assign_pages. >> >> First of all I still don't see why in this one special case it is a >> meaningful burden to do the count-to-order conversion in the caller you >> mean to add, > > This sort of works for one caller. However, I would expect some more > user in the future (we use it for Live-Update). > >> and hence why we really need this new function (to keep it >> simple, you could even have the caller not break down to arbitrary >> power-of-2 chunks, but simply iterate over all individual [order-0] >> pages). > > The function assign_pages() will always use 1U << order (and sadly 1 << > order). So we would end up to convert the count in multiple order for > then directly converting back to a number. To me, this sounds rather > pointless... > > There are also a slight benefits to call assign_pages() a single time > during boot because it will reduce the number of time we need to > lock/unlock d->page_alloc_lock. Well, all of this is why I did add ... >> The more that I'm not happy with the chosen name, despite it >> having been suggested during v1 review. _If_ we needed two functions, >> imo they ought to be named assign_page() (dealing with a single page of >> the given order) and assign_pages(). Backporting confusion could be >> helped by altering the order of parameters, such that the compiler >> would point out that adjustments at call sites are needed. ... this. Not sure whether you not commenting on it means you agree with the proposal. >>> --- a/xen/common/page_alloc.c >>> +++ b/xen/common/page_alloc.c >>> @@ -2301,14 +2301,14 @@ void init_domheap_pages(paddr_t ps, paddr_t pe) >>> } >>> >>> >>> -int assign_pages( >>> +int assign_pages_nr( >>> struct domain *d, >>> struct page_info *pg, >>> - unsigned int order, >>> + unsigned int nr_pfns, >> >> Even leaving the naming aspect of "pfns" aside, I can't see why this >> can't be simply "nr" (of appropriate type, see next remark). >> >>> unsigned int memflags) >>> { >>> int rc = 0; >>> - unsigned long i; >>> + unsigned int i; >> >> This is not an acceptable type change, at least not as long as it's not >> justified at all in the description. While both Arm and x86 will be >> fine this way, the code here is supposed to be generic, and hence would >> better remain generally correct. > > I would like to point out the code is already not correct as we are > using 1U << order or worse 1 << order :). Indeed there are improvements (towards being consistent) to be made. But this is not an excuse to make things worse here. At least one of the two loops already properly uses 1ul; sadly that's only debugging code. And of course something like domain_tot_pages() (and the underlying field) dealing with "unsigned int" doesn't help consistency either. As it stands we're limiting ourselves to 8Tb VMs, as it seems, and for no good reason. Jan
Hi Jan, On 01/07/2021 09:26, Jan Beulich wrote: > On 30.06.2021 20:29, Julien Grall wrote: >> On 10/06/2021 10:49, Jan Beulich wrote: >>> On 07.06.2021 04:43, Penny Zheng wrote: >>>> Introduce new interface assign_pages_nr to deal with when page number is >>>> not in a power-of-two, which will save the trouble each time user needs >>>> to split the size in a power of 2 to use assign_pages. >>> >>> First of all I still don't see why in this one special case it is a >>> meaningful burden to do the count-to-order conversion in the caller you >>> mean to add, >> >> This sort of works for one caller. However, I would expect some more >> user in the future (we use it for Live-Update). >> >>> and hence why we really need this new function (to keep it >>> simple, you could even have the caller not break down to arbitrary >>> power-of-2 chunks, but simply iterate over all individual [order-0] >>> pages). >> >> The function assign_pages() will always use 1U << order (and sadly 1 << >> order). So we would end up to convert the count in multiple order for >> then directly converting back to a number. To me, this sounds rather >> pointless... >> >> There are also a slight benefits to call assign_pages() a single time >> during boot because it will reduce the number of time we need to >> lock/unlock d->page_alloc_lock. > > Well, all of this is why I did add ... > >>> The more that I'm not happy with the chosen name, despite it >>> having been suggested during v1 review. _If_ we needed two functions, >>> imo they ought to be named assign_page() (dealing with a single page of >>> the given order) and assign_pages(). Backporting confusion could be >>> helped by altering the order of parameters, such that the compiler >>> would point out that adjustments at call sites are needed. > > ... this. Oh, it wasn't entirely clear whether you were objecting of offering the possibility to pass a number of pages rather than an order. > Not sure whether you not commenting on it means you agree > with the proposal. Yes I am happy with your proposal. > >>>> --- a/xen/common/page_alloc.c >>>> +++ b/xen/common/page_alloc.c >>>> @@ -2301,14 +2301,14 @@ void init_domheap_pages(paddr_t ps, paddr_t pe) >>>> } >>>> >>>> >>>> -int assign_pages( >>>> +int assign_pages_nr( >>>> struct domain *d, >>>> struct page_info *pg, >>>> - unsigned int order, >>>> + unsigned int nr_pfns, >>> >>> Even leaving the naming aspect of "pfns" aside, I can't see why this >>> can't be simply "nr" (of appropriate type, see next remark). >>> >>>> unsigned int memflags) >>>> { >>>> int rc = 0; >>>> - unsigned long i; >>>> + unsigned int i; >>> >>> This is not an acceptable type change, at least not as long as it's not >>> justified at all in the description. While both Arm and x86 will be >>> fine this way, the code here is supposed to be generic, and hence would >>> better remain generally correct. >> >> I would like to point out the code is already not correct as we are >> using 1U << order or worse 1 << order :). > > Indeed there are improvements (towards being consistent) to be made. But > this is not an excuse to make things worse here. At least one of the two > loops already properly uses 1ul; sadly that's only debugging code. And > of course something like domain_tot_pages() (and the underlying field) > dealing with "unsigned int" doesn't help consistency either. As it stands > we're limiting ourselves to 8Tb VMs, as it seems, and for no good reason. I actually have a patch in my queue that hardens assign_pages(). I will aim to send it later today. Cheers,
On 01.07.2021 11:24, Julien Grall wrote: > On 01/07/2021 09:26, Jan Beulich wrote: >> On 30.06.2021 20:29, Julien Grall wrote: >>> On 10/06/2021 10:49, Jan Beulich wrote: >>>> On 07.06.2021 04:43, Penny Zheng wrote: >>>>> Introduce new interface assign_pages_nr to deal with when page number is >>>>> not in a power-of-two, which will save the trouble each time user needs >>>>> to split the size in a power of 2 to use assign_pages. >>>> >>>> First of all I still don't see why in this one special case it is a >>>> meaningful burden to do the count-to-order conversion in the caller you >>>> mean to add, >>> >>> This sort of works for one caller. However, I would expect some more >>> user in the future (we use it for Live-Update). >>> >>>> and hence why we really need this new function (to keep it >>>> simple, you could even have the caller not break down to arbitrary >>>> power-of-2 chunks, but simply iterate over all individual [order-0] >>>> pages). >>> >>> The function assign_pages() will always use 1U << order (and sadly 1 << >>> order). So we would end up to convert the count in multiple order for >>> then directly converting back to a number. To me, this sounds rather >>> pointless... >>> >>> There are also a slight benefits to call assign_pages() a single time >>> during boot because it will reduce the number of time we need to >>> lock/unlock d->page_alloc_lock. >> >> Well, all of this is why I did add ... >> >>>> The more that I'm not happy with the chosen name, despite it >>>> having been suggested during v1 review. _If_ we needed two functions, >>>> imo they ought to be named assign_page() (dealing with a single page of >>>> the given order) and assign_pages(). Backporting confusion could be >>>> helped by altering the order of parameters, such that the compiler >>>> would point out that adjustments at call sites are needed. >> >> ... this. > > Oh, it wasn't entirely clear whether you were objecting of offering the > possibility to pass a number of pages rather than an order. Easily understood: I indeed we trying to express my preference to stick to what we have, but trying to suggest a variant that I think I could live with. Jan
diff --git a/xen/common/page_alloc.c b/xen/common/page_alloc.c index 8c00262c04..e244d2e52e 100644 --- a/xen/common/page_alloc.c +++ b/xen/common/page_alloc.c @@ -2301,14 +2301,14 @@ void init_domheap_pages(paddr_t ps, paddr_t pe) } -int assign_pages( +int assign_pages_nr( struct domain *d, struct page_info *pg, - unsigned int order, + unsigned int nr_pfns, unsigned int memflags) { int rc = 0; - unsigned long i; + unsigned int i; spin_lock(&d->page_alloc_lock); @@ -2324,7 +2324,7 @@ int assign_pages( { unsigned int extra_pages = 0; - for ( i = 0; i < (1ul << order); i++ ) + for ( i = 0; i < nr_pfns; i++ ) { ASSERT(!(pg[i].count_info & ~PGC_extra)); if ( pg[i].count_info & PGC_extra ) @@ -2333,18 +2333,18 @@ int assign_pages( ASSERT(!extra_pages || ((memflags & MEMF_no_refcount) && - extra_pages == 1u << order)); + extra_pages == nr_pfns)); } #endif if ( pg[0].count_info & PGC_extra ) { - d->extra_pages += 1u << order; + d->extra_pages += nr_pfns; memflags &= ~MEMF_no_refcount; } else if ( !(memflags & MEMF_no_refcount) ) { - unsigned int tot_pages = domain_tot_pages(d) + (1 << order); + unsigned int tot_pages = domain_tot_pages(d) + nr_pfns; if ( unlikely(tot_pages > d->max_pages) ) { @@ -2356,10 +2356,10 @@ int assign_pages( } if ( !(memflags & MEMF_no_refcount) && - unlikely(domain_adjust_tot_pages(d, 1 << order) == (1 << order)) ) + unlikely(domain_adjust_tot_pages(d, nr_pfns) == nr_pfns) ) get_knownalive_domain(d); - for ( i = 0; i < (1 << order); i++ ) + for ( i = 0; i < nr_pfns; i++ ) { ASSERT(page_get_owner(&pg[i]) == NULL); page_set_owner(&pg[i], d); @@ -2374,6 +2374,14 @@ int assign_pages( return rc; } +int assign_pages( + struct domain *d, + struct page_info *pg, + unsigned int order, + unsigned int memflags) +{ + return assign_pages_nr(d, pg, (1U << order), memflags); +} struct page_info *alloc_domheap_pages( struct domain *d, unsigned int order, unsigned int memflags) diff --git a/xen/include/xen/mm.h b/xen/include/xen/mm.h index df25e55966..25d970e857 100644 --- a/xen/include/xen/mm.h +++ b/xen/include/xen/mm.h @@ -131,12 +131,18 @@ int query_page_offline(mfn_t mfn, uint32_t *status); void heap_init_late(void); -int assign_pages( +int assign_pages_nr( struct domain *d, struct page_info *pg, - unsigned int order, + unsigned int nr_pfns, unsigned int memflags); + int assign_pages( + struct domain *d, + struct page_info *pg, + unsigned int order, + unsigned int memflags); + /* Dump info to serial console */ void arch_dump_shared_mem_info(void);
Introduce new interface assign_pages_nr to deal with when page number is not in a power-of-two, which will save the trouble each time user needs to split the size in a power of 2 to use assign_pages. Signed-off-by: Penny Zheng <penny.zheng@arm.com> --- changes v2: - introduce new interface assign_pages_nr --- xen/common/page_alloc.c | 26 +++++++++++++++++--------- xen/include/xen/mm.h | 10 ++++++++-- 2 files changed, 25 insertions(+), 11 deletions(-)