CVE-2021-47531

En el kernel de Linux, se ha resuelto la siguiente vulnerabilidad: drm/msm: corrige mmap para incluir VM_IO y VM_DONTDUMP. En el commit 510410bfc034 ("drm/msm: implementa mmap como función de objeto GEM") Nosotros cambiamos a un método nuevo / más limpio de hacer las cosas. Eso es bueno, pero nos perdimos un poquito. Antes de esa confirmación, solíamos ejecutar _primero_ el caso drm_gem_mmap_obj() donde `obj->funcs->mmap()` era NULL. Eso significó que ejecutamos: vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP; vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags)); vma->vm_page_prot = pgprot_decrypted(vma->vm_page_prot); ...y _luego_ modificamos esas asignaciones con las nuestras. Ahora que `obj->funcs->mmap()` ya no es NULL, no ejecutamos el código predeterminado. Parece que el hecho de que vm_flags obtuviera VM_IO/VM_DONTDUMP fue importante porque ahora estamos teniendo fallas en Chromebooks que usan ARC++ al cerrar sesión. Específicamente, un bloqueo que se ve así (esto ocurre en un kernel 5.10 con backports relevantes, pero también se ve en un kernel 5.15): No se puede manejar la solicitud de paginación del kernel en la dirección virtual ffffffc008000000 Información de cancelación de memoria: ESR = 0x96000006 EC = 0x25: DABT ( EL actual), IL = 32 bits SET = 0, FnV = 0 EA = 0, S1PTW = 0 Información de cancelación de datos: ISV = 0, ISS = 0x00000006 CM = 0, WnR = 0 tabla de intercambio: 4k páginas, VA de 39 bits , pgdp=000000008293d000 [ffffffc008000000] pgd=00000001002b3003, p4d=00000001002b3003, pud=00000001002b3003, pmd=0000000000000000 Error interno: operaciones: 96000006 [#1] PREEMPT SMP [...] CPU: 7 PID: 15734 Comunicaciones: crash_dump64 Contaminado: GW 5.10.67 #1 [...] Nombre del hardware: Qualcomm Technologies, Inc. sc7280 IDP SKU2 plataforma (DT) pstate: 80400009 (Nzcv daif +PAN -UAO -TCO BTYPE=--) pc: __arch_copy_to_user+0xc0/0x30c lr: copyout+0xac/0x14c [...] Rastreo de llamadas: __arch_copy_to_user+0xc0/0x30c copy_page_to_iter+0x1a0/0x294 Process_vm_rw_core+0x240/0x408 Process_vm_rw+0x110/0x16c __arm64_sys_process_vm_readv+0x30/0 x3c el0_svc_common+0xf8/0x250 do_el0_svc+0x30/0x80 el0_svc+0x10/0x1c el0_sync_handler+0x78/0x108 el0_sync+0x184/0x1c0 Código: f8408423 f80008c3 910020c6 36100082 (b8404423) Agreguemos las dos banderas nuevamente. Mientras estamos en esto, el hecho de que no estemos ejecutando medios predeterminados que _no_ necesitamos borrar VM_PFNMAP, así que elimínelo y guarde una instrucción. NOTA: se confirmó que VM_IO era el indicador importante para solucionar el problema que estaba viendo, pero volver a agregar VM_DONTDUMP parece algo sensato, así que lo estoy haciendo también.

In the Linux kernel, the following vulnerability has been resolved: drm/msm: Fix mmap to include VM_IO and VM_DONTDUMP In commit 510410bfc034 ("drm/msm: Implement mmap as GEM object function") we switched to a new/cleaner method of doing things. That's good, but we missed a little bit. Before that commit, we used to _first_ run through the drm_gem_mmap_obj() case where `obj->funcs->mmap()` was NULL. That meant that we ran: vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP; vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags)); vma->vm_page_prot = pgprot_decrypted(vma->vm_page_prot); ...and _then_ we modified those mappings with our own. Now that `obj->funcs->mmap()` is no longer NULL we don't run the default code. It looks like the fact that the vm_flags got VM_IO / VM_DONTDUMP was important because we're now getting crashes on Chromebooks that use ARC++ while logging out. Specifically a crash that looks like this (this is on a 5.10 kernel w/ relevant backports but also seen on a 5.15 kernel): Unable to handle kernel paging request at virtual address ffffffc008000000 Mem abort info: ESR = 0x96000006 EC = 0x25: DABT (current EL), IL = 32 bits SET = 0, FnV = 0 EA = 0, S1PTW = 0 Data abort info: ISV = 0, ISS = 0x00000006 CM = 0, WnR = 0 swapper pgtable: 4k pages, 39-bit VAs, pgdp=000000008293d000 [ffffffc008000000] pgd=00000001002b3003, p4d=00000001002b3003, pud=00000001002b3003, pmd=0000000000000000 Internal error: Oops: 96000006 [#1] PREEMPT SMP [...] CPU: 7 PID: 15734 Comm: crash_dump64 Tainted: G W 5.10.67 #1 [...] Hardware name: Qualcomm Technologies, Inc. sc7280 IDP SKU2 platform (DT) pstate: 80400009 (Nzcv daif +PAN -UAO -TCO BTYPE=--) pc : __arch_copy_to_user+0xc0/0x30c lr : copyout+0xac/0x14c [...] Call trace: __arch_copy_to_user+0xc0/0x30c copy_page_to_iter+0x1a0/0x294 process_vm_rw_core+0x240/0x408 process_vm_rw+0x110/0x16c __arm64_sys_process_vm_readv+0x30/0x3c el0_svc_common+0xf8/0x250 do_el0_svc+0x30/0x80 el0_svc+0x10/0x1c el0_sync_handler+0x78/0x108 el0_sync+0x184/0x1c0 Code: f8408423 f80008c3 910020c6 36100082 (b8404423) Let's add the two flags back in. While we're at it, the fact that we aren't running the default means that we _don't_ need to clear out VM_PFNMAP, so remove that and save an instruction. NOTE: it was confirmed that VM_IO was the important flag to fix the problem I was seeing, but adding back VM_DONTDUMP seems like a sane thing to do so I'm doing that too.