Date: Thu, 23 Aug 2018 00:10:34 +0200 From: Michael Gmelin <freebsd@grem.de> To: Konstantin Belousov <kostikbel@gmail.com> Cc: John Baldwin <jhb@FreeBSD.org>, "freebsd-current@freebsd.org" <freebsd-current@freebsd.org>, Matthias Apitz <guru@unixarea.de> Subject: Re: Fatal trap 12: page fault on Acer Chromebook 720 (peppy) Message-ID: <1C7DACDC-36F2-4E65-8C75-7B7215BB6546@grem.de> In-Reply-To: <20180822211528.GB2340@kib.kiev.ua> References: <e82ed552-83b0-5331-3117-6750b8c205f7@FreeBSD.org> <07E28AC5-EBE6-4893-810A-6C03F07925C8@grem.de> <8726bc32-6023-bfe1-7600-5b2c706236f8@FreeBSD.org> <20180819165951.274d61b0@bsd64.grem.de> <20180819161642.GP2340@kib.kiev.ua> <20180820004512.5171fa75@bsd64.grem.de> <20180820150904.GS2340@kib.kiev.ua> <57B6DC4C-16EE-4B7B-B691-CB79D8C40289@grem.de> <20180822154603.GW2340@kib.kiev.ua> <C8BC1622-4F47-4E56-A428-46DAC5CE71B4@grem.de> <20180822211528.GB2340@kib.kiev.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
> On 22. Aug 2018, at 23:15, Konstantin Belousov <kostikbel@gmail.com> wrote= : >=20 >> On Wed, Aug 22, 2018 at 10:03:54PM +0200, Michael Gmelin wrote: >>=20 >>=20 >>>> On 22. Aug 2018, at 17:46, Konstantin Belousov <kostikbel@gmail.com> wr= ote: >>>>=20 >>>> On Tue, Aug 21, 2018 at 12:14:35AM +0200, Michael Gmelin wrote: >>>>=20 >>>>=20 >>>>>> On 20. Aug 2018, at 17:09, Konstantin Belousov <kostikbel@gmail.com> w= rote: >>>>>>=20 >>>>>> On Mon, Aug 20, 2018 at 12:45:12AM +0200, Michael Gmelin wrote: >>>>>>=20 >>>>>> See here for a screenshot (also including the output of "show pte >>>>>> 0xfffff80001000000"): >>>>>>=20 >>>>>> https://gist.github.com/grembo/78d0f2a100dd4f16775b85a118769658#file-= ddb1-png >>>>> It is too early for ddb routines to register. >>>>> Ok can you try the following debugging patch, to verify my guess ? >>>>>=20 >>>>> diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c >>>>> index 18777d23f09..cd05fdb763f 100644 >>>>> --- a/sys/amd64/amd64/pmap.c >>>>> +++ b/sys/amd64/amd64/pmap.c >>>>> @@ -1052,8 +1052,7 @@ create_pagetables(vm_paddr_t *firstaddr) >>>>> pd_p =3D (pd_entry_t *)DMPDkernphys; >>>>> for (i =3D 0; i < (NPDEPG * nkdmpde); i++) >>>>> pd_p[i] =3D (i << PDRSHIFT) | X86_PG_V | PG_PS | pg_g | >>>>> - X86_PG_M | X86_PG_A | pg_nx | >>>>> - bootaddr_rwx(i << PDRSHIFT); >>>>> + X86_PG_M | X86_PG_A | pg_nx | X86_PG_RW; >>>>> for (i =3D 0; i < nkdmpde; i++) >>>>> pdp_p[i] =3D (DMPDkernphys + ptoa(i)) | X86_PG_RW | >>>>> X86_PG_V; >>>>=20 >>>> With this change it boots okay (mptramp_pagetables is 0x1000000, as exp= ected). >>>=20 >>> Can you apply the following on top of the previous debugging patch and s= how >>> me the line printed ? >>>=20 >>> diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c >>> index 3d70532b7fd..613fa9f2165 100644 >>> --- a/sys/amd64/amd64/pmap.c >>> +++ b/sys/amd64/amd64/pmap.c >>> @@ -2662,6 +2662,7 @@ pmap_pinit0(pmap_t pmap) >>> pmap->pm_pcids[i].pm_gen =3D 1; >>> } >>> pmap_activate_boot(pmap); >>> +printf("bootaddr addr %#lx rwx %#lx btext %#lx _end %#lx brwsection %#l= x etext %#lx KERNBASE %#lx\n", 0x1000000UL, bootaddr_rwx(0x1000000UL), (uint= ptr_t)btext, (uintptr_t)_end, (uintptr_t)brwsection, (uintptr_t)etext, (uint= ptr_t)KERNBASE); >>> } >>>=20 >>> void >>=20 >> bootaddr addr 0x1000000 rwx 0 btext 0xffffffff80342000 _end 0xffffffff823= cf840 brwsection #ffffffff81a00000 etext 0xffffffff812041e4 KERNBASE 0xfffff= fff80000000 >>=20 >=20 > Try this, please. Revert all debugging pmap.c patches that I provided > before. >=20 > diff --git a/sys/amd64/amd64/mp_machdep.c b/sys/amd64/amd64/mp_machdep.c > index 4ca2e07e578..2ee8f862854 100644 > --- a/sys/amd64/amd64/mp_machdep.c > +++ b/sys/amd64/amd64/mp_machdep.c > @@ -87,6 +87,8 @@ __FBSDID("$FreeBSD$"); >=20 > #define GiB(v) (v ## ULL << 30) >=20 > +#define AP_BOOTPT_SZ (PAGE_SIZE * 3) > + > extern struct pcpu __pcpu[]; >=20 > /* Temporary variables for init_secondary() */ > @@ -101,45 +103,78 @@ char *dbg_stack; >=20 > static int start_ap(int apic_id); >=20 > +static bool > +is_kernel_paddr(vm_paddr_t pa) > +{ > + > + return (pa >=3D trunc_2mpage(btext - KERNBASE) && > + pa < round_page(_end - KERNBASE)); > +} > + > +static bool > +is_mpboot_good(vm_paddr_t start, vm_paddr_t end) > +{ > + > + return (start + AP_BOOTPT_SZ <=3D GiB(4) && > + end >=3D start + AP_BOOTPT_SZ && atop(end) < Maxmem); > +} > + > /* > * Calculate usable address in base memory for AP trampoline code. > */ > void > mp_bootaddress(vm_paddr_t *physmap, unsigned int *physmap_idx) > { > + vm_paddr_t start, end; > unsigned int i; > bool allocated; >=20 > alloc_ap_trampoline(physmap, physmap_idx); >=20 > + /* > + * Find a memory region big enough below the 4GB boundary to > + * store the initial page tables. Region must be mapped by > + * the direct map. > + * > + * Note that it needs to be aligned to a page boundary. > + */ > allocated =3D false; > for (i =3D *physmap_idx; i <=3D *physmap_idx; i -=3D 2) { > /* > - * Find a memory region big enough below the 4GB > - * boundary to store the initial page tables. Region > - * must be mapped by the direct map. > - * > - * Note that it needs to be aligned to a page > - * boundary. > + * First, try to chomp at the start of the physmap region. > + * Kernel binary might claim it already. > + */ > + start =3D round_page(physmap[i]); > + end =3D trunc_page(physmap[i + 1]); > + if (is_mpboot_good(start, end) && > + !is_kernel_paddr(start) && !is_kernel_paddr(end - 1)) { > + allocated =3D true; > + physmap[i] =3D start + AP_BOOTPT_SZ; > + break; > + } > + > + /* > + * Second, try to chomp at the end. Again, check > + * against kernel. > */ > - if (physmap[i] >=3D GiB(4) || physmap[i + 1] - > - round_page(physmap[i]) < PAGE_SIZE * 3 || > - atop(physmap[i + 1]) > Maxmem) > - continue; > - > - allocated =3D true; > - mptramp_pagetables =3D round_page(physmap[i]); > - physmap[i] =3D round_page(physmap[i]) + (PAGE_SIZE * 3); > + end =3D trunc_page(physmap[i + 1]); > + start =3D end - AP_BOOTPT_SZ; > + if (start >=3D physmap[i] && is_mpboot_good(start, end) && > + !is_kernel_paddr(start) && !is_kernel_paddr(end - 1)) { > + allocated =3D true; > + physmap[i + 1] =3D start; > + break; > + } > + } > + if (allocated) { > + mptramp_pagetables =3D start; > if (physmap[i] =3D=3D physmap[i + 1] && *physmap_idx !=3D 0) { > memmove(&physmap[i], &physmap[i + 2], > sizeof(*physmap) * (*physmap_idx - i + 2)); > *physmap_idx -=3D 2; > } > - break; > - } > - > - if (!allocated) { > - mptramp_pagetables =3D trunc_page(boot_address) - (PAGE_SIZE * 3)= ; > + } else { > + mptramp_pagetables =3D trunc_page(boot_address) - AP_BOOTPT_SZ; > if (bootverbose) > printf( > "Cannot find enough space for the initial AP page tables, placing them at %= #x", Reverted back to r337813 and applied the patch. Unfortunately it panics just= like before. Adding back physmap debugging like before shows that it=E2=80=99= s still using pages 0x1000-0x1003 (Stopped at native_start_all_aps+0x92: mov= q %rax,(%rsi)) Best, Michael
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1C7DACDC-36F2-4E65-8C75-7B7215BB6546>