Date: Sat, 22 Nov 2025 12:48:32 -0800 From: Mark Millard <marklmi@yahoo.com> To: Konstantin Belousov <kostikbel@gmail.com>, FreeBSD Current <freebsd-current@freebsd.org> Subject: Re: mmap( MAP_ANON) is broken on current. (was Still seeing Failed assertion: "p[i] == 0" on armv7 buildworld) [debug.vm_check_pg_zero=1 silent] Message-ID: <75CEABED-3CCB-4DB9-AC82-5980696C2A06@yahoo.com> References: <75CEABED-3CCB-4DB9-AC82-5980696C2A06.ref@yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Konstantin Belousov <kostikbel_at_gmail.com> wrote on Date: Sat, 22 Nov 2025 20:19:33 UTC : > On Sat, Nov 22, 2025 at 08:54:21PM +0100, Michal Meloun wrote: > > > > > > On 22.11.2025 19:45, Konstantin Belousov wrote: > > > On Sat, Nov 22, 2025 at 07:01:03PM +0100, Michal Meloun wrote: > > > > > Would you please gather the same ddebugging info, with this patch applied? > > > > Oups, sorry. > > > > In meantime, next round with he vm_map patch finished successfully. > > > > > > It was still the case of coalescing previous entry and the mapping. > > > > > > It is weird, the patch ensures that there is no pages in the object > > > backing the new region, and due to the ensured properties of the object, > > > there should be no way to create pages under us. > > > I am almost sure that the provided patch is correct, but it might be > > > some additional cases that I miss. > > > > > > Please apply the following debugging patch, it includes the vm_object' > > > part. Instead of allowing the corruption in userspace, kernel should > > > panic now. Can you confirm that? > > > > > > diff --git a/sys/vm/vm_map.c b/sys/vm/vm_map.c > > > . . . > > > > Unfortunately, KASSERT doesn't assert on failure. Don't hit me, please. :) > > > > . . . > > ... > Please in addition to the patch, enable debug.vm_check_pg_zero. # sysctl debug.vm_check_pg_zero=1 debug.vm_check_pg_zero: 0 -> 1 # sysctl debug.vm_check_pg_zero debug.vm_check_pg_zero: 1 # env WITH_META_MODE= make -j10 buildworld --- buildworld --- . . . Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaCodeComplete.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaConcept.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaConsumer.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaCoroutine.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaDecl.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaDeclAttr.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaDeclCXX.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaDeclObjC.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaExceptionSpec.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaExpr.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaExprCXX.pico Building /usr/obj/usr/src/i386.i386/lib/clang/libclang/Sema/SemaExprMember.pico <jemalloc>: /usr/src/contrib/jemalloc/include/jemalloc/internal/ehooks.h:170: Failed assertion: "p[i] == 0" Abort trap (core dumped) *** [Sema/SemaCodeComplete.pico] Error code 134 . . . === Mark Millard marklmi at yahoo.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?75CEABED-3CCB-4DB9-AC82-5980696C2A06>