CVE-2024-26989

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> arm64: hibernate: Fix level3 translation fault in swsusp_save()<br /> <br /> On arm64 machines, swsusp_save() faults if it attempts to access<br /> MEMBLOCK_NOMAP memory ranges. This can be reproduced in QEMU using UEFI<br /> when booting with rodata=off debug_pagealloc=off and CONFIG_KFENCE=n:<br /> <br /> Unable to handle kernel paging request at virtual address ffffff8000000000<br /> Mem abort info:<br /> ESR = 0x0000000096000007<br /> EC = 0x25: DABT (current EL), IL = 32 bits<br /> SET = 0, FnV = 0<br /> EA = 0, S1PTW = 0<br /> FSC = 0x07: level 3 translation fault<br /> Data abort info:<br /> ISV = 0, ISS = 0x00000007, ISS2 = 0x00000000<br /> CM = 0, WnR = 0, TnD = 0, TagAccess = 0<br /> GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0<br /> swapper pgtable: 4k pages, 39-bit VAs, pgdp=00000000eeb0b000<br /> [ffffff8000000000] pgd=180000217fff9803, p4d=180000217fff9803, pud=180000217fff9803, pmd=180000217fff8803, pte=0000000000000000<br /> Internal error: Oops: 0000000096000007 [#1] SMP<br /> Internal error: Oops: 0000000096000007 [#1] SMP<br /> Modules linked in: xt_multiport ipt_REJECT nf_reject_ipv4 xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 libcrc32c iptable_filter bpfilter rfkill at803x snd_hda_codec_hdmi snd_hda_intel snd_intel_dspcfg dwmac_generic stmmac_platform snd_hda_codec stmmac joydev pcs_xpcs snd_hda_core phylink ppdev lp parport ramoops reed_solomon ip_tables x_tables nls_iso8859_1 vfat multipath linear amdgpu amdxcp drm_exec gpu_sched drm_buddy hid_generic usbhid hid radeon video drm_suballoc_helper drm_ttm_helper ttm i2c_algo_bit drm_display_helper cec drm_kms_helper drm<br /> CPU: 0 PID: 3663 Comm: systemd-sleep Not tainted 6.6.2+ #76<br /> Source Version: 4e22ed63a0a48e7a7cff9b98b7806d8d4add7dc0<br /> Hardware name: Greatwall GW-XXXXXX-XXX/GW-XXXXXX-XXX, BIOS KunLun BIOS V4.0 01/19/2021<br /> pstate: 600003c5 (nZCv DAIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)<br /> pc : swsusp_save+0x280/0x538<br /> lr : swsusp_save+0x280/0x538<br /> sp : ffffffa034a3fa40<br /> x29: ffffffa034a3fa40 x28: ffffff8000001000 x27: 0000000000000000<br /> x26: ffffff8001400000 x25: ffffffc08113e248 x24: 0000000000000000<br /> x23: 0000000000080000 x22: ffffffc08113e280 x21: 00000000000c69f2<br /> x20: ffffff8000000000 x19: ffffffc081ae2500 x18: 0000000000000000<br /> x17: 6666662074736420 x16: 3030303030303030 x15: 3038666666666666<br /> x14: 0000000000000b69 x13: ffffff9f89088530 x12: 00000000ffffffea<br /> x11: 00000000ffff7fff x10: 00000000ffff7fff x9 : ffffffc08193f0d0<br /> x8 : 00000000000bffe8 x7 : c0000000ffff7fff x6 : 0000000000000001<br /> x5 : ffffffa0fff09dc8 x4 : 0000000000000000 x3 : 0000000000000027<br /> x2 : 0000000000000000 x1 : 0000000000000000 x0 : 000000000000004e<br /> Call trace:<br /> swsusp_save+0x280/0x538<br /> swsusp_arch_suspend+0x148/0x190<br /> hibernation_snapshot+0x240/0x39c<br /> hibernate+0xc4/0x378<br /> state_store+0xf0/0x10c<br /> kobj_attr_store+0x14/0x24<br /> <br /> The reason is swsusp_save() -&gt; copy_data_pages() -&gt; page_is_saveable()<br /> -&gt; kernel_page_present() assuming that a page is always present when<br /> can_set_direct_map() is false (all of rodata_full,<br /> debug_pagealloc_enabled() and arm64_kfence_can_set_direct_map() false),<br /> irrespective of the MEMBLOCK_NOMAP ranges. Such MEMBLOCK_NOMAP regions<br /> should not be saved during hibernation.<br /> <br /> This problem was introduced by changes to the pfn_valid() logic in<br /> commit a7d9f306ba70 ("arm64: drop pfn_valid_within() and simplify<br /> pfn_valid()").<br /> <br /> Similar to other architectures, drop the !can_set_direct_map() check in<br /> kernel_page_present() so that page_is_savable() skips such pages.<br /> <br /> [catalin.marinas@arm.com: rework commit message]