Re: [PATCH v3 31/34] xen/riscv: add minimal stuff to mm.h to build full Xen

[Jan Beulich <jbeulich@suse.com>]

On 22.12.2023 16:13, Oleksii Kurochko wrote:
> Signed-off-by: Oleksii Kurochko <oleksii.kurochko@gmail.com>
> ---
> Changes in V3:
>  - update the commit message

??? (yet again)

> --- a/xen/arch/riscv/include/asm/mm.h
> +++ b/xen/arch/riscv/include/asm/mm.h
> @@ -3,8 +3,251 @@
>  #ifndef _ASM_RISCV_MM_H
>  #define _ASM_RISCV_MM_H
>  
> +#include <public/xen.h>
> +#include <xen/pdx.h>
> +#include <xen/types.h>
> +
> +#include <asm/page.h>
>  #include <asm/page-bits.h>
>  
> +#define paddr_to_pdx(pa)    mfn_to_pdx(maddr_to_mfn(pa))
> +#define gfn_to_gaddr(gfn)   pfn_to_paddr(gfn_x(gfn))
> +#define gaddr_to_gfn(ga)    _gfn(paddr_to_pfn(ga))
> +#define mfn_to_maddr(mfn)   pfn_to_paddr(mfn_x(mfn))
> +#define maddr_to_mfn(ma)    _mfn(paddr_to_pfn(ma))
> +#define vmap_to_mfn(va)     maddr_to_mfn(virt_to_maddr((vaddr_t)va))
> +#define vmap_to_page(va)    mfn_to_page(vmap_to_mfn(va))

Everything you have above ...

> +#define paddr_to_pdx(pa)    mfn_to_pdx(maddr_to_mfn(pa))
> +#define gfn_to_gaddr(gfn)   pfn_to_paddr(gfn_x(gfn))
> +#define gaddr_to_gfn(ga)    _gfn(paddr_to_pfn(ga))
> +#define mfn_to_maddr(mfn)   pfn_to_paddr(mfn_x(mfn))
> +#define maddr_to_mfn(ma)    _mfn(paddr_to_pfn(ma))
> +#define vmap_to_mfn(va)     maddr_to_mfn(virt_to_maddr((vaddr_t)va))
> +#define vmap_to_page(va)    mfn_to_page(vmap_to_mfn(va))

... appears a 2nd time right afterwards.

> +#define virt_to_maddr(va) ((paddr_t)((vaddr_t)(va) & PADDR_MASK))
> +#define maddr_to_virt(pa) ((void *)((paddr_t)(pa) | DIRECTMAP_VIRT_START))
> +
> +/* Convert between Xen-heap virtual addresses and machine frame numbers. */
> +#define __virt_to_mfn(va) (virt_to_maddr(va) >> PAGE_SHIFT)
> +#define __mfn_to_virt(mfn) maddr_to_virt((paddr_t)(mfn) << PAGE_SHIFT)

These would imo better use maddr_to_mfn() and mfn_to_maddr(), rather than
kind of open-coding them. The former could also use PFN_DOWN() as an
alternative.

> +/* Convert between Xen-heap virtual addresses and page-info structures. */
> +static inline struct page_info *virt_to_page(const void *v)
> +{
> +    BUG();
> +    return NULL;
> +}
> +
> +/*
> + * We define non-underscored wrappers for above conversion functions.
> + * These are overriden in various source files while underscored version
> + * remain intact.
> + */
> +#define virt_to_mfn(va)     __virt_to_mfn(va)
> +#define mfn_to_virt(mfn)    __mfn_to_virt(mfn)

Is this really still needed? Would be pretty nice if in a new port we
could get to start cleanly right away (i.e. by not needing per-file
overrides, but using type-safe expansions here right away).

> +struct page_info
> +{
> +    /* Each frame can be threaded onto a doubly-linked list. */
> +    struct page_list_entry list;
> +
> +    /* Reference count and various PGC_xxx flags and fields. */
> +    unsigned long count_info;
> +
> +    /* Context-dependent fields follow... */
> +    union {
> +        /* Page is in use: ((count_info & PGC_count_mask) != 0). */
> +        struct {
> +            /* Type reference count and various PGT_xxx flags and fields. */
> +            unsigned long type_info;
> +        } inuse;
> +        /* Page is on a free list: ((count_info & PGC_count_mask) == 0). */
> +        union {
> +            struct {
> +                /*
> +                 * Index of the first *possibly* unscrubbed page in the buddy.
> +                 * One more bit than maximum possible order to accommodate
> +                 * INVALID_DIRTY_IDX.
> +                 */
> +#define INVALID_DIRTY_IDX ((1UL << (MAX_ORDER + 1)) - 1)
> +                unsigned long first_dirty:MAX_ORDER + 1;
> +
> +                /* Do TLBs need flushing for safety before next page use? */
> +                bool need_tlbflush:1;
> +
> +#define BUDDY_NOT_SCRUBBING    0
> +#define BUDDY_SCRUBBING        1
> +#define BUDDY_SCRUB_ABORT      2
> +                unsigned long scrub_state:2;
> +            };
> +
> +                unsigned long val;
> +            } free;

Indentation is wrong (and thus misleading) for these two lines.

> +
> +    } u;

Nit: I don't see the value of the trailing blank line inside the
union.

> +    union {
> +        /* Page is in use, but not as a shadow. */

I question the appicability of "shadow" here.

> +        struct {
> +            /* Owner of this page (zero if page is anonymous). */
> +            struct domain *domain;
> +        } inuse;
> +
> +        /* Page is on a free list. */
> +        struct {
> +            /* Order-size of the free chunk this page is the head of. */
> +            unsigned int order;
> +        } free;
> +
> +    } v;
> +
> +    union {
> +        /*
> +         * Timestamp from 'TLB clock', used to avoid extra safety flushes.
> +         * Only valid for: a) free pages, and b) pages with zero type count
> +         */
> +        uint32_t tlbflush_timestamp;
> +    };
> +    uint64_t pad;
> +};
> +
> +#define frame_table ((struct page_info *)FRAMETABLE_VIRT_START)
> +
> +/* PDX of the first page in the frame table. */
> +extern unsigned long frametable_base_pdx;

From this I conclude memory on RISC-V systems may not start at (or near) 0?

> +/* Convert between machine frame numbers and page-info structures. */
> +#define mfn_to_page(mfn)                                            \
> +    (frame_table + (mfn_to_pdx(mfn) - frametable_base_pdx))
> +#define page_to_mfn(pg)                                             \
> +    pdx_to_mfn((unsigned long)((pg) - frame_table) + frametable_base_pdx)
> +
> +static inline void *page_to_virt(const struct page_info *pg)
> +{
> +    return mfn_to_virt(mfn_x(page_to_mfn(pg)));
> +}
> +
> +/*
> + * Common code requires get_page_type and put_page_type.
> + * We don't care about typecounts so we just do the minimum to make it
> + * happy.
> + */
> +static inline int get_page_type(struct page_info *page, unsigned long type)
> +{
> +    return 1;
> +}
> +
> +static inline void put_page_type(struct page_info *page)
> +{
> +}
> +
> +static inline void put_page_and_type(struct page_info *page)
> +{
> +    put_page_type(page);
> +    put_page(page);
> +}
> +
> +/*
> + * RISC-V does not have an M2P, but common code expects a handful of
> + * M2P-related defines and functions. Provide dummy versions of these.
> + */
> +#define INVALID_M2P_ENTRY        (~0UL)
> +#define SHARED_M2P_ENTRY         (~0UL - 1UL)
> +#define SHARED_M2P(_e)           ((_e) == SHARED_M2P_ENTRY)
> +
> +/* Xen always owns P2M on RISC-V */
> +#define set_gpfn_from_mfn(mfn, pfn) do { (void) (mfn), (void)(pfn); } while (0)

Nit: Stray blank again after cast.

> +#define mfn_to_gfn(d, mfn) ((void)(d), _gfn(mfn_x(mfn)))

What's the relation of the comment with these two #define-s?

> +#define PDX_GROUP_SHIFT (16 + 5)
> +
> +static inline unsigned long domain_get_maximum_gpfn(struct domain *d)
> +{
> +    BUG();
> +    return 0;
> +}
> +
> +static inline long arch_memory_op(int op, XEN_GUEST_HANDLE_PARAM(void) arg)
> +{
> +    BUG();
> +    return 0;
> +}
> +
> +/*
> + * On RISCV, all the RAM is currently direct mapped in Xen.
> + * Hence return always true.
> + */
> +static inline bool arch_mfns_in_directmap(unsigned long mfn, unsigned long nr)
> +{
> +    return true;
> +}
> +
> +#define PG_shift(idx)   (BITS_PER_LONG - (idx))
> +#define PG_mask(x, idx) (x ## UL << PG_shift(idx))
> +
> +#define PGT_none          PG_mask(0, 1)  /* no special uses of this page   */
> +#define PGT_writable_page PG_mask(1, 1)  /* has writable mappings?         */
> +#define PGT_type_mask     PG_mask(1, 1)  /* Bits 31 or 63.                 */
> +
> + /* Count of uses of this frame as its current type. */
> +#define PGT_count_width   PG_shift(2)
> +#define PGT_count_mask    ((1UL<<PGT_count_width)-1)

Nit: Style (missing blanks around binary operators). Also a few more
times further down.

> +/*
> + * Page needs to be scrubbed. Since this bit can only be set on a page that is
> + * free (i.e. in PGC_state_free) we can reuse PGC_allocated bit.
> + */
> +#define _PGC_need_scrub   _PGC_allocated
> +#define PGC_need_scrub    PGC_allocated
> +
> +//  /* Cleared when the owning guest 'frees' this page. */

Why a commented out comment?

> +#define _PGC_allocated    PG_shift(1)
> +#define PGC_allocated     PG_mask(1, 1)
> +  /* Page is Xen heap? */
> +#define _PGC_xen_heap     PG_shift(2)
> +#define PGC_xen_heap      PG_mask(1, 2)
> +/* Page is broken? */
> +#define _PGC_broken       PG_shift(7)
> +#define PGC_broken        PG_mask(1, 7)
> + /* Mutually-exclusive page states: { inuse, offlining, offlined, free }. */

Can similar comments in this block please all be similarly indented
(or not)?

> +#define PGC_state         PG_mask(3, 9)
> +#define PGC_state_inuse   PG_mask(0, 9)
> +#define PGC_state_offlining PG_mask(1, 9)
> +#define PGC_state_offlined PG_mask(2, 9)
> +#define PGC_state_free    PG_mask(3, 9)
> +// #define page_state_is(pg, st) (((pg)->count_info&PGC_state) == PGC_state_##st)

???

> +/* Count of references to this frame. */
> +#define PGC_count_width   PG_shift(9)
> +#define PGC_count_mask    ((1UL<<PGC_count_width)-1)
> +
> +#define page_state_is(pg, st) (((pg)->count_info&PGC_state) == PGC_state_##st)

And here it then "properly" appears?

> +#define _PGC_extra        PG_shift(10)
> +#define PGC_extra         PG_mask(1, 10)
> +
> +#define is_xen_heap_page(page) ((page)->count_info & PGC_xen_heap)
> +#define is_xen_heap_mfn(mfn) \
> +    (mfn_valid(mfn) && is_xen_heap_page(mfn_to_page(mfn)))
> +
> +#define is_xen_fixed_mfn(mfn)                                   \
> +    ((mfn_to_maddr(mfn) >= virt_to_maddr(&_start)) &&           \
> +     (mfn_to_maddr(mfn) <= virt_to_maddr((vaddr_t)_end - 1)))

Why does _start need prefixing wuth & and _end prefixing with a cast?
First and foremost both want to be consistent. And then preferably
with as little extra clutter as possible.

> +#define page_get_owner(_p)    (_p)->v.inuse.domain
> +#define page_set_owner(_p,_d) ((_p)->v.inuse.domain = (_d))
> +
> +/* TODO: implement */
> +#define mfn_valid(mfn) ({ (void) (mfn); 0; })
> +
> +#define mfn_to_gfn(d, mfn) ((void)(d), _gfn(mfn_x(mfn)))

This appeared further up already.

> +#define domain_set_alloc_bitsize(d) ((void)0)

Better ((void)(d)) ? And then ...

> +#define domain_clamp_alloc_bitsize(d, b) (b)

... ((void)(d), (b)) here?

Jan