gsd-2024-26670
Vulnerability from gsd
Modified
2024-02-20 06:02
Details
In the Linux kernel, the following vulnerability has been resolved:
arm64: entry: fix ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD
Currently the ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD workaround isn't
quite right, as it is supposed to be applied after the last explicit
memory access, but is immediately followed by an LDR.
The ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD workaround is used to
handle Cortex-A520 erratum 2966298 and Cortex-A510 erratum 3117295,
which are described in:
* https://developer.arm.com/documentation/SDEN2444153/0600/?lang=en
* https://developer.arm.com/documentation/SDEN1873361/1600/?lang=en
In both cases the workaround is described as:
| If pagetable isolation is disabled, the context switch logic in the
| kernel can be updated to execute the following sequence on affected
| cores before exiting to EL0, and after all explicit memory accesses:
|
| 1. A non-shareable TLBI to any context and/or address, including
| unused contexts or addresses, such as a `TLBI VALE1 Xzr`.
|
| 2. A DSB NSH to guarantee completion of the TLBI.
The important part being that the TLBI+DSB must be placed "after all
explicit memory accesses".
Unfortunately, as-implemented, the TLBI+DSB is immediately followed by
an LDR, as we have:
| alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD
| tlbi vale1, xzr
| dsb nsh
| alternative_else_nop_endif
| alternative_if_not ARM64_UNMAP_KERNEL_AT_EL0
| ldr lr, [sp, #S_LR]
| add sp, sp, #PT_REGS_SIZE // restore sp
| eret
| alternative_else_nop_endif
|
| [ ... KPTI exception return path ... ]
This patch fixes this by reworking the logic to place the TLBI+DSB
immediately before the ERET, after all explicit memory accesses.
The ERET is currently in a separate alternative block, and alternatives
cannot be nested. To account for this, the alternative block for
ARM64_UNMAP_KERNEL_AT_EL0 is replaced with a single alternative branch
to skip the KPTI logic, with the new shape of the logic being:
| alternative_insn "b .L_skip_tramp_exit_\@", nop, ARM64_UNMAP_KERNEL_AT_EL0
| [ ... KPTI exception return path ... ]
| .L_skip_tramp_exit_\@:
|
| ldr lr, [sp, #S_LR]
| add sp, sp, #PT_REGS_SIZE // restore sp
|
| alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD
| tlbi vale1, xzr
| dsb nsh
| alternative_else_nop_endif
| eret
The new structure means that the workaround is only applied when KPTI is
not in use; this is fine as noted in the documented implications of the
erratum:
| Pagetable isolation between EL0 and higher level ELs prevents the
| issue from occurring.
... and as per the workaround description quoted above, the workaround
is only necessary "If pagetable isolation is disabled".
Aliases
{ "gsd": { "metadata": { "exploitCode": "unknown", "remediation": "unknown", "reportConfidence": "confirmed", "type": "vulnerability" }, "osvSchema": { "aliases": [ "CVE-2024-26670" ], "details": "In the Linux kernel, the following vulnerability has been resolved:\n\narm64: entry: fix ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n\nCurrently the ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD workaround isn\u0027t\nquite right, as it is supposed to be applied after the last explicit\nmemory access, but is immediately followed by an LDR.\n\nThe ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD workaround is used to\nhandle Cortex-A520 erratum 2966298 and Cortex-A510 erratum 3117295,\nwhich are described in:\n\n* https://developer.arm.com/documentation/SDEN2444153/0600/?lang=en\n* https://developer.arm.com/documentation/SDEN1873361/1600/?lang=en\n\nIn both cases the workaround is described as:\n\n| If pagetable isolation is disabled, the context switch logic in the\n| kernel can be updated to execute the following sequence on affected\n| cores before exiting to EL0, and after all explicit memory accesses:\n|\n| 1. A non-shareable TLBI to any context and/or address, including\n| unused contexts or addresses, such as a `TLBI VALE1 Xzr`.\n|\n| 2. A DSB NSH to guarantee completion of the TLBI.\n\nThe important part being that the TLBI+DSB must be placed \"after all\nexplicit memory accesses\".\n\nUnfortunately, as-implemented, the TLBI+DSB is immediately followed by\nan LDR, as we have:\n\n| alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n| \ttlbi\tvale1, xzr\n| \tdsb\tnsh\n| alternative_else_nop_endif\n| alternative_if_not ARM64_UNMAP_KERNEL_AT_EL0\n| \tldr\tlr, [sp, #S_LR]\n| \tadd\tsp, sp, #PT_REGS_SIZE\t\t// restore sp\n| \teret\n| alternative_else_nop_endif\n|\n| [ ... KPTI exception return path ... ]\n\nThis patch fixes this by reworking the logic to place the TLBI+DSB\nimmediately before the ERET, after all explicit memory accesses.\n\nThe ERET is currently in a separate alternative block, and alternatives\ncannot be nested. To account for this, the alternative block for\nARM64_UNMAP_KERNEL_AT_EL0 is replaced with a single alternative branch\nto skip the KPTI logic, with the new shape of the logic being:\n\n| alternative_insn \"b .L_skip_tramp_exit_\\@\", nop, ARM64_UNMAP_KERNEL_AT_EL0\n| \t[ ... KPTI exception return path ... ]\n| .L_skip_tramp_exit_\\@:\n|\n| \tldr\tlr, [sp, #S_LR]\n| \tadd\tsp, sp, #PT_REGS_SIZE\t\t// restore sp\n|\n| alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n| \ttlbi\tvale1, xzr\n| \tdsb\tnsh\n| alternative_else_nop_endif\n| \teret\n\nThe new structure means that the workaround is only applied when KPTI is\nnot in use; this is fine as noted in the documented implications of the\nerratum:\n\n| Pagetable isolation between EL0 and higher level ELs prevents the\n| issue from occurring.\n\n... and as per the workaround description quoted above, the workaround\nis only necessary \"If pagetable isolation is disabled\".", "id": "GSD-2024-26670", "modified": "2024-02-20T06:02:29.146023Z", "schema_version": "1.4.0" } }, "namespaces": { "cve.org": { "CVE_data_meta": { "ASSIGNER": "cve@kernel.org", "ID": "CVE-2024-26670", "STATE": "PUBLIC" }, "affects": { "vendor": { "vendor_data": [ { "product": { "product_data": [ { "product_name": "Linux", "version": { "version_data": [ { "version_affected": "\u003c", "version_name": "471470bc7052", "version_value": "58eb5c07f417" }, { "version_value": "not down converted", "x_cve_json_5_version_data": { "defaultStatus": "affected", "versions": [ { "status": "affected", "version": "6.6" }, { "lessThan": "6.6", "status": "unaffected", "version": "0", "versionType": "custom" }, { "lessThanOrEqual": "6.6.*", "status": "unaffected", "version": "6.6.15", "versionType": "custom" }, { "lessThanOrEqual": "6.7.*", "status": "unaffected", "version": "6.7.3", "versionType": "custom" }, { "lessThanOrEqual": "*", "status": "unaffected", "version": "6.8", "versionType": "original_commit_for_fix" } ] } } ] } } ] }, "vendor_name": "Linux" } ] } }, "data_format": "MITRE", "data_type": "CVE", "data_version": "4.0", "description": { "description_data": [ { "lang": "eng", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\narm64: entry: fix ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n\nCurrently the ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD workaround isn\u0027t\nquite right, as it is supposed to be applied after the last explicit\nmemory access, but is immediately followed by an LDR.\n\nThe ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD workaround is used to\nhandle Cortex-A520 erratum 2966298 and Cortex-A510 erratum 3117295,\nwhich are described in:\n\n* https://developer.arm.com/documentation/SDEN2444153/0600/?lang=en\n* https://developer.arm.com/documentation/SDEN1873361/1600/?lang=en\n\nIn both cases the workaround is described as:\n\n| If pagetable isolation is disabled, the context switch logic in the\n| kernel can be updated to execute the following sequence on affected\n| cores before exiting to EL0, and after all explicit memory accesses:\n|\n| 1. A non-shareable TLBI to any context and/or address, including\n| unused contexts or addresses, such as a `TLBI VALE1 Xzr`.\n|\n| 2. A DSB NSH to guarantee completion of the TLBI.\n\nThe important part being that the TLBI+DSB must be placed \"after all\nexplicit memory accesses\".\n\nUnfortunately, as-implemented, the TLBI+DSB is immediately followed by\nan LDR, as we have:\n\n| alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n| \ttlbi\tvale1, xzr\n| \tdsb\tnsh\n| alternative_else_nop_endif\n| alternative_if_not ARM64_UNMAP_KERNEL_AT_EL0\n| \tldr\tlr, [sp, #S_LR]\n| \tadd\tsp, sp, #PT_REGS_SIZE\t\t// restore sp\n| \teret\n| alternative_else_nop_endif\n|\n| [ ... KPTI exception return path ... ]\n\nThis patch fixes this by reworking the logic to place the TLBI+DSB\nimmediately before the ERET, after all explicit memory accesses.\n\nThe ERET is currently in a separate alternative block, and alternatives\ncannot be nested. To account for this, the alternative block for\nARM64_UNMAP_KERNEL_AT_EL0 is replaced with a single alternative branch\nto skip the KPTI logic, with the new shape of the logic being:\n\n| alternative_insn \"b .L_skip_tramp_exit_\\@\", nop, ARM64_UNMAP_KERNEL_AT_EL0\n| \t[ ... KPTI exception return path ... ]\n| .L_skip_tramp_exit_\\@:\n|\n| \tldr\tlr, [sp, #S_LR]\n| \tadd\tsp, sp, #PT_REGS_SIZE\t\t// restore sp\n|\n| alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n| \ttlbi\tvale1, xzr\n| \tdsb\tnsh\n| alternative_else_nop_endif\n| \teret\n\nThe new structure means that the workaround is only applied when KPTI is\nnot in use; this is fine as noted in the documented implications of the\nerratum:\n\n| Pagetable isolation between EL0 and higher level ELs prevents the\n| issue from occurring.\n\n... and as per the workaround description quoted above, the workaround\nis only necessary \"If pagetable isolation is disabled\"." } ] }, "generator": { "engine": "bippy-b4257b672505" }, "problemtype": { "problemtype_data": [ { "description": [ { "lang": "eng", "value": "n/a" } ] } ] }, "references": { "reference_data": [ { "name": "https://git.kernel.org/stable/c/58eb5c07f41704464b9acc09ab0707b6769db6c0", "refsource": "MISC", "url": "https://git.kernel.org/stable/c/58eb5c07f41704464b9acc09ab0707b6769db6c0" }, { "name": "https://git.kernel.org/stable/c/baa0aaac16432019651e0d60c41cd34a0c3c3477", "refsource": "MISC", "url": "https://git.kernel.org/stable/c/baa0aaac16432019651e0d60c41cd34a0c3c3477" }, { "name": "https://git.kernel.org/stable/c/832dd634bd1b4e3bbe9f10b9c9ba5db6f6f2b97f", "refsource": "MISC", "url": "https://git.kernel.org/stable/c/832dd634bd1b4e3bbe9f10b9c9ba5db6f6f2b97f" } ] } }, "nvd.nist.gov": { "cve": { "descriptions": [ { "lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\narm64: entry: fix ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n\nCurrently the ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD workaround isn\u0027t\nquite right, as it is supposed to be applied after the last explicit\nmemory access, but is immediately followed by an LDR.\n\nThe ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD workaround is used to\nhandle Cortex-A520 erratum 2966298 and Cortex-A510 erratum 3117295,\nwhich are described in:\n\n* https://developer.arm.com/documentation/SDEN2444153/0600/?lang=en\n* https://developer.arm.com/documentation/SDEN1873361/1600/?lang=en\n\nIn both cases the workaround is described as:\n\n| If pagetable isolation is disabled, the context switch logic in the\n| kernel can be updated to execute the following sequence on affected\n| cores before exiting to EL0, and after all explicit memory accesses:\n|\n| 1. A non-shareable TLBI to any context and/or address, including\n| unused contexts or addresses, such as a `TLBI VALE1 Xzr`.\n|\n| 2. A DSB NSH to guarantee completion of the TLBI.\n\nThe important part being that the TLBI+DSB must be placed \"after all\nexplicit memory accesses\".\n\nUnfortunately, as-implemented, the TLBI+DSB is immediately followed by\nan LDR, as we have:\n\n| alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n| \ttlbi\tvale1, xzr\n| \tdsb\tnsh\n| alternative_else_nop_endif\n| alternative_if_not ARM64_UNMAP_KERNEL_AT_EL0\n| \tldr\tlr, [sp, #S_LR]\n| \tadd\tsp, sp, #PT_REGS_SIZE\t\t// restore sp\n| \teret\n| alternative_else_nop_endif\n|\n| [ ... KPTI exception return path ... ]\n\nThis patch fixes this by reworking the logic to place the TLBI+DSB\nimmediately before the ERET, after all explicit memory accesses.\n\nThe ERET is currently in a separate alternative block, and alternatives\ncannot be nested. To account for this, the alternative block for\nARM64_UNMAP_KERNEL_AT_EL0 is replaced with a single alternative branch\nto skip the KPTI logic, with the new shape of the logic being:\n\n| alternative_insn \"b .L_skip_tramp_exit_\\@\", nop, ARM64_UNMAP_KERNEL_AT_EL0\n| \t[ ... KPTI exception return path ... ]\n| .L_skip_tramp_exit_\\@:\n|\n| \tldr\tlr, [sp, #S_LR]\n| \tadd\tsp, sp, #PT_REGS_SIZE\t\t// restore sp\n|\n| alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD\n| \ttlbi\tvale1, xzr\n| \tdsb\tnsh\n| alternative_else_nop_endif\n| \teret\n\nThe new structure means that the workaround is only applied when KPTI is\nnot in use; this is fine as noted in the documented implications of the\nerratum:\n\n| Pagetable isolation between EL0 and higher level ELs prevents the\n| issue from occurring.\n\n... and as per the workaround description quoted above, the workaround\nis only necessary \"If pagetable isolation is disabled\"." }, { "lang": "es", "value": "En el kernel de Linux, se ha resuelto la siguiente vulnerabilidad: arm64: entrada: arreglar ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD Actualmente, la soluci\u00f3n alternativa ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD no es del todo correcta, ya que se supone que debe aplicarse despu\u00e9s del \u00faltimo acceso expl\u00edcito a la memoria, pero va seguida inmediatamente por un LDR. La soluci\u00f3n alternativa ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD se utiliza para manejar las erratas Cortex-A520 2966298 y Cortex-A510 erratas 3117295, que se describen en: * https://developer.arm.com/documentation/SDEN2444153/0600/?lang=en * https:// desarrollador.arm.com/documentation/SDEN1873361/1600/?lang=en En ambos casos, la soluci\u00f3n se describe como: | Si el aislamiento de la tabla de p\u00e1ginas est\u00e1 deshabilitado, la l\u00f3gica de cambio de contexto en | El kernel se puede actualizar para ejecutar la siguiente secuencia en los afectados | n\u00facleos antes de salir a EL0 y despu\u00e9s de todos los accesos expl\u00edcitos a la memoria: | | 1. Un TLBI que no se puede compartir en cualquier contexto y/o direcci\u00f3n, incluido | contextos o direcciones no utilizados, como `TLBI VALE1 Xzr`. | | 2. Un OSD NSH para garantizar la finalizaci\u00f3n del TLBI. La parte importante es que el TLBI+DSB debe colocarse \"despu\u00e9s de todos los accesos expl\u00edcitos a la memoria\". Desafortunadamente, tal como se implement\u00f3, el TLBI+DSB es seguido inmediatamente por un LDR, como tenemos: | Alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD | tlbi vale1, xzr | dsb nsh | alternativa_else_nop_endif | alternativa_si_no ARM64_UNMAP_KERNEL_AT_EL0 | ldr lr, [sp, #S_LR] | agregar sp, sp, #PT_REGS_SIZE // restaurar sp | eremita | alternativa_else_nop_endif | | [... Ruta de retorno de excepci\u00f3n KPTI...] Este parche soluciona este problema reelaborando la l\u00f3gica para colocar TLBI+DSB inmediatamente antes de ERET, despu\u00e9s de todos los accesos expl\u00edcitos a la memoria. El ERET se encuentra actualmente en un bloque alternativo separado y las alternativas no se pueden anidar. Para tener en cuenta esto, el bloque alternativo para ARM64_UNMAP_KERNEL_AT_EL0 se reemplaza con una \u00fanica rama alternativa para omitir la l\u00f3gica KPTI, siendo la nueva forma de la l\u00f3gica: | Alternative_insn \"b .L_skip_tramp_exit_\\@\", nop, ARM64_UNMAP_KERNEL_AT_EL0 | [... Ruta de retorno de excepci\u00f3n KPTI...] | .L_skip_tramp_exit_\\@: | | ldr lr, [sp, #S_LR] | agregar sp, sp, #PT_REGS_SIZE // restaurar sp | | Alternative_if ARM64_WORKAROUND_SPECULATIVE_UNPRIV_LOAD | tlbi vale1, xzr | dsb nsh | alternativa_else_nop_endif | eret La nueva estructura significa que la soluci\u00f3n alternativa s\u00f3lo se aplica cuando KPTI no est\u00e1 en uso; esto est\u00e1 bien, como se indica en las implicaciones documentadas de la fe de erratas: | El aislamiento de la tabla de paginaci\u00f3n entre EL0 y los EL de nivel superior evita que | que ocurra el problema. ... y seg\u00fan la descripci\u00f3n de la soluci\u00f3n citada anteriormente, la soluci\u00f3n s\u00f3lo es necesaria \"si el aislamiento de la tabla de p\u00e1ginas est\u00e1 deshabilitado\"." } ], "id": "CVE-2024-26670", "lastModified": "2024-04-02T12:50:42.233", "metrics": {}, "published": "2024-04-02T07:15:43.760", "references": [ { "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "url": "https://git.kernel.org/stable/c/58eb5c07f41704464b9acc09ab0707b6769db6c0" }, { "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "url": "https://git.kernel.org/stable/c/832dd634bd1b4e3bbe9f10b9c9ba5db6f6f2b97f" }, { "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "url": "https://git.kernel.org/stable/c/baa0aaac16432019651e0d60c41cd34a0c3c3477" } ], "sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "vulnStatus": "Awaiting Analysis" } } } }
Loading...