nvd-json-data-feeds/CVE-2021/CVE-2021-474xx/CVE-2021-47414.json

{
  "id": "CVE-2021-47414",
  "sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
  "published": "2024-05-21T15:15:26.887",
  "lastModified": "2024-11-21T06:36:05.703",
  "vulnStatus": "Awaiting Analysis",
  "cveTags": [],
  "descriptions": [
    {
      "lang": "en",
      "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nriscv: Flush current cpu icache before other cpus\n\nOn SiFive Unmatched, I recently fell onto the following BUG when booting:\n\n[    0.000000] ftrace: allocating 36610 entries in 144 pages\n[    0.000000] Oops - illegal instruction [#1]\n[    0.000000] Modules linked in:\n[    0.000000] CPU: 0 PID: 0 Comm: swapper Not tainted 5.13.1+ #5\n[    0.000000] Hardware name: SiFive HiFive Unmatched A00 (DT)\n[    0.000000] epc : riscv_cpuid_to_hartid_mask+0x6/0xae\n[    0.000000]  ra : __sbi_rfence_v02+0xc8/0x10a\n[    0.000000] epc : ffffffff80007240 ra : ffffffff80009964 sp : ffffffff81803e10\n[    0.000000]  gp : ffffffff81a1ea70 tp : ffffffff8180f500 t0 : ffffffe07fe30000\n[    0.000000]  t1 : 0000000000000004 t2 : 0000000000000000 s0 : ffffffff81803e60\n[    0.000000]  s1 : 0000000000000000 a0 : ffffffff81a22238 a1 : ffffffff81803e10\n[    0.000000]  a2 : 0000000000000000 a3 : 0000000000000000 a4 : 0000000000000000\n[    0.000000]  a5 : 0000000000000000 a6 : ffffffff8000989c a7 : 0000000052464e43\n[    0.000000]  s2 : ffffffff81a220c8 s3 : 0000000000000000 s4 : 0000000000000000\n[    0.000000]  s5 : 0000000000000000 s6 : 0000000200000100 s7 : 0000000000000001\n[    0.000000]  s8 : ffffffe07fe04040 s9 : ffffffff81a22c80 s10: 0000000000001000\n[    0.000000]  s11: 0000000000000004 t3 : 0000000000000001 t4 : 0000000000000008\n[    0.000000]  t5 : ffffffcf04000808 t6 : ffffffe3ffddf188\n[    0.000000] status: 0000000200000100 badaddr: 0000000000000000 cause: 0000000000000002\n[    0.000000] [<ffffffff80007240>] riscv_cpuid_to_hartid_mask+0x6/0xae\n[    0.000000] [<ffffffff80009474>] sbi_remote_fence_i+0x1e/0x26\n[    0.000000] [<ffffffff8000b8f4>] flush_icache_all+0x12/0x1a\n[    0.000000] [<ffffffff8000666c>] patch_text_nosync+0x26/0x32\n[    0.000000] [<ffffffff8000884e>] ftrace_init_nop+0x52/0x8c\n[    0.000000] [<ffffffff800f051e>] ftrace_process_locs.isra.0+0x29c/0x360\n[    0.000000] [<ffffffff80a0e3c6>] ftrace_init+0x80/0x130\n[    0.000000] [<ffffffff80a00f8c>] start_kernel+0x5c4/0x8f6\n[    0.000000] ---[ end trace f67eb9af4d8d492b ]---\n[    0.000000] Kernel panic - not syncing: Attempted to kill the idle task!\n[    0.000000] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]---\n\nWhile ftrace is looping over a list of addresses to patch, it always failed\nwhen patching the same function: riscv_cpuid_to_hartid_mask. Looking at the\nbacktrace, the illegal instruction is encountered in this same function.\nHowever, patch_text_nosync, after patching the instructions, calls\nflush_icache_range. But looking at what happens in this function:\n\nflush_icache_range -> flush_icache_all\n                   -> sbi_remote_fence_i\n                   -> __sbi_rfence_v02\n                   -> riscv_cpuid_to_hartid_mask\n\nThe icache and dcache of the current cpu are never synchronized between the\npatching of riscv_cpuid_to_hartid_mask and calling this same function.\n\nSo fix this by flushing the current cpu's icache before asking for the other\ncpus to do the same."
    },
    {
      "lang": "es",
      "value": "En el kernel de Linux, se resolvi\u00f3 la siguiente vulnerabilidad: riscv: vaciar el icache de la CPU actual antes que otras CPU. En SiFive Unmatched, recientemente encontr\u00e9 el siguiente BUGal arrancar: [0.000000] ftrace: asignar 36610 entradas en 144 p\u00e1ginas [0.000000] Ups - instrucci\u00f3n ilegal [#1] [ 0.000000] M\u00f3dulos vinculados en: [ 0.000000] CPU: 0 PID: 0 Comm: swapper Not tainted 5.13.1+ #5 [ 0.000000] Nombre del hardware: SiFive HiFive Unmatched A00 (DT) [ 0.000000] epc: riscv_cpuid_to_hartid_mask+0x6/0xae [ 0.000000] ra : __sbi_rfence_v02+0xc8/0x10a [ 0.000000] epc : ffffffff80007240 ra : ffffffff80009964 sp : ffffffff81803e10 [ 0 .000000] gp : ffffffff81a1ea70 tp : ffffffff8180f500 t0 : ffffffe07fe30000 [ 0.000000] t1 : 00000000000000004 t2 : 00000000000000000 s0: ffffffff81803e60 [0.000000] s1: 0000000000000000 a0: ffffffff81a22238 a1: ffffffff81803e10 [0.000000] a2: 000000000000000 a3: 0000000 000000000 a4: 0000000000000000 [0.000000] a5: 0000000000000000 a6: ffffffff8000989c a7: 0000000052464e43 [0.000000] s2: ffffffff81a220c8 s3: 0000000000000000 s4: 0000000000000000 [ 0.000000] s5 : 0000000000000000 s6 : 0000000200000100 s7 : 0000000000000001 [ 0.000000] s8 : ffffffe07fe04040 s9 : ffffffff81a22c80 s10: 0000000000001000 [0.000000] s11: 0000000000000004 t3: 0000000000000001 t4: 0000000000000008 [0.000000] t5: ffcf04000808 t6: ffffffe3ffddf188 [0.000000] estado : 0000000200000100 badaddr: 0000000000000000 causa: 0000000000000002 [ 0.000000] [] riscv_cpuid_to_hartid_mask+0x6/0xae [ 0.000000] ffffffff80009474&gt;] sbi_remote_fence_i+0x1e/0x26 [ 0.000000] [] Flush_icache_all+0x12/0x1a [ 0.000000] [] patch_text_nosync+0x26/0x32 [ 0.000000] [] ftrace_init_nop+0x52/0x8c [ 0.000000] [] ftrace_process_locs.isra.0+0x29c/ 0x360 [ 0.000000] [] ftrace_init+ 0x80/0x130 [ 0.000000] [] start_kernel+0x5c4/0x8f6 [ 0.000000] ---[ end trace f67eb9af4d8d492b ]--- [ 0.000000] P\u00e1nico en el kernel - no se sincroniza: \u00a1Intent\u00f3 finalizar la tarea inactiva! [0.000000] ---[ fin del p\u00e1nico del kernel: no se sincroniza: \u00a1se intent\u00f3 finalizar la tarea inactiva! ]--- Mientras ftrace recorre una lista de direcciones para parchear, siempre fallaba al parchear la misma funci\u00f3n: riscv_cpuid_to_hartid_mask. Al observar el seguimiento, la instrucci\u00f3n ilegal se encuentra en esta misma funci\u00f3n. Sin embargo, patch_text_nosync, despu\u00e9s de parchear las instrucciones, llama a flush_icache_range. Pero observando lo que sucede en esta funci\u00f3n: Flush_icache_range -&gt; Flush_icache_all -&gt; sbi_remote_fence_i -&gt; __sbi_rfence_v02 -&gt; riscv_cpuid_to_hartid_mask El icache y el dcache de la CPU actual nunca se sincronizan entre el parche de riscv_cpuid_to_hartid_mask y la llamada a esta misma funci\u00f3n. As\u00ed que solucione este problema limpiando el icache de la CPU actual antes de pedirle a las otras CPU que hagan lo mismo."
    }
  ],
  "metrics": {},
  "references": [
    {
      "url": "https://git.kernel.org/stable/c/427faa29e06f0709476ea1bd59758f997ec8b64e",
      "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
    },
    {
      "url": "https://git.kernel.org/stable/c/bb8958d5dc79acbd071397abb57b8756375fe1ce",
      "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
    },
    {
      "url": "https://git.kernel.org/stable/c/f1c7aa87c423e765e3862349c2f095fdfccdd9b3",
      "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
    },
    {
      "url": "https://git.kernel.org/stable/c/427faa29e06f0709476ea1bd59758f997ec8b64e",
      "source": "af854a3a-2127-422b-91ae-364da2661108"
    },
    {
      "url": "https://git.kernel.org/stable/c/bb8958d5dc79acbd071397abb57b8756375fe1ce",
      "source": "af854a3a-2127-422b-91ae-364da2661108"
    },
    {
      "url": "https://git.kernel.org/stable/c/f1c7aa87c423e765e3862349c2f095fdfccdd9b3",
      "source": "af854a3a-2127-422b-91ae-364da2661108"
    }
  ]
}