lkml.org 
[lkml]   [2023]   [Feb]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    From
    SubjectRe: [PATCH v4 2/6] riscv: Rework kasan population functions
    Date
    Alexandre Ghiti <alexghiti@rivosinc.com> writes:

    > Our previous kasan population implementation used to have the final kasan
    > shadow region mapped with kasan_early_shadow_page, because we did not clean
    > the early mapping and then we had to populate the kasan region "in-place"
    > which made the code cumbersome.
    >
    > So now we clear the early mapping, establish a temporary mapping while we
    > populate the kasan shadow region with just the kernel regions that will
    > be used.
    >
    > This new version uses the "generic" way of going through a page table
    > that may be folded at runtime (avoid the XXX_next macros).
    >
    > It was tested with outline instrumentation on an Ubuntu kernel
    > configuration successfully.
    >
    > Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com>

    (One minor nit, that can be addressed later.)

    Reviewed-by: Björn Töpel <bjorn@rivosinc.com>

    > arch/riscv/mm/kasan_init.c | 361 +++++++++++++++++++------------------
    > 1 file changed, 183 insertions(+), 178 deletions(-)


    > @@ -482,7 +437,37 @@ static void __init kasan_shallow_populate(void *start, void *end)
    > unsigned long vend = PAGE_ALIGN((unsigned long)end);
    >
    > kasan_shallow_populate_pgd(vaddr, vend);
    > - local_flush_tlb_all();
    > +}
    > +
    > +static void create_tmp_mapping(void)
    > +{
    > + void *ptr;
    > + p4d_t *base_p4d;
    > +
    > + /*
    > + * We need to clean the early mapping: this is hard to achieve "in-place",
    > + * so install a temporary mapping like arm64 and x86 do.
    > + */
    > + memcpy(tmp_pg_dir, swapper_pg_dir, sizeof(pgd_t) * PTRS_PER_PGD);
    > +
    > + /* Copy the last p4d since it is shared with the kernel mapping. */
    > + if (pgtable_l5_enabled) {
    > + ptr = (p4d_t *)pgd_page_vaddr(*pgd_offset_k(KASAN_SHADOW_END));
    > + memcpy(tmp_p4d, ptr, sizeof(p4d_t) * PTRS_PER_P4D);
    > + set_pgd(&tmp_pg_dir[pgd_index(KASAN_SHADOW_END)],
    > + pfn_pgd(PFN_DOWN(__pa(tmp_p4d)), PAGE_TABLE));
    > + base_p4d = tmp_p4d;
    > + } else {
    > + base_p4d = (p4d_t *)tmp_pg_dir;
    > + }
    > +
    > + /* Copy the last pud since it is shared with the kernel mapping. */
    > + if (pgtable_l4_enabled) {
    > + ptr = (pud_t *)p4d_page_vaddr(*(base_p4d + p4d_index(KASAN_SHADOW_END)));
    > + memcpy(tmp_pud, ptr, sizeof(pud_t) * PTRS_PER_PUD);
    > + set_p4d(&base_p4d[p4d_index(KASAN_SHADOW_END)],
    > + pfn_p4d(PFN_DOWN(__pa(tmp_pud)), PAGE_TABLE));
    > + }
    > }
    >
    > void __init kasan_init(void)
    > @@ -490,10 +475,27 @@ void __init kasan_init(void)
    > phys_addr_t p_start, p_end;
    > u64 i;
    >
    > - if (IS_ENABLED(CONFIG_KASAN_VMALLOC))
    > + create_tmp_mapping();
    > + csr_write(CSR_SATP, PFN_DOWN(__pa(tmp_pg_dir)) | satp_mode);

    Nit: Maybe add a comment, why the sfence.vma is *not* required here. I
    tripped over it.


    Björn

    \
     
     \ /
      Last update: 2023-03-27 00:27    [W:2.378 / U:0.276 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site