arm64: Rework checks for broken Cavium HW in the PI code
Marc Zyngier
maintainerPATCH v2 1/1 arm64: Rework checks for broken Cavium HW in the PI code
Calling into the MIDR checking framework from the PI code has recently become much harder, due to the new fancy "multi-MIDR" support that relies on tables being populated at boot time, but not that early that they are available to the PI code. There are additional issues with this framework, as the code really isn't position independend *at all*.
This leads to some ugly breakages, as reported by Ada.
It so appears that the only reason for the PI code to call into the MIDR checking code is to cope with The Most Broken ARM64 System Ever, aka Cavium ThunderX, which cannot deal with nG attributes that result of the combination of KASLR and KPTI as a consequence of Erratum 27456.
Duplicate the check for the erratum in the PI code, removing the dependency on the bulk of the MIDR checking framework. This allows dropping that same check from kaslr_requires_kpti(), as the KPTI code already relies on the ARM64_WORKAROUND_CAVIUM_27456 cap.
---
Notes:
* From v1 [1]:
- Preserved KASLR functionnality by duplicating the MIDR checks in the PI code. While this is a bit ugly, it keeps everything working for another day, and removes a duplicate check in the KPTI code.
- Tested in a VM to check that KASLR was still up and running. Yay!
[1] https://lore.kernel.org/r/[email protected]
arch/arm64/include/asm/mmu.h | 11 ----------- arch/arm64/kernel/cpu_errata.c | 2 +- arch/arm64/kernel/image-vars.h | 4 ---- arch/arm64/kernel/pi/map_kernel.c | 25 ++++++++++++++++++++++++- 4 files changed, 25 insertions(+), 17 deletions(-)
arch/arm64/include/asm/mmu.h +0 -11
@@ -94,17 +94,6 @@ static inline bool kaslr_requires_kpti(void)return false; } -/* -* Systems affected by Cavium erratum 24756 are incompatible -* with KPTI. -*/ -if (IS_ENABLED (CONFIG_CAVIUM_ERRATUM_27456 )) { -extern const struct midr_range cavium_erratum_27456_cpus[]; - -if (is_midr_in_range_list (cavium_erratum_27456_cpus)) -return false; - } -return true; }
arch/arm64/kernel/cpu_errata.c +1 -1
@@ -335,7 +335,7 @@ static const struct midr_range cavium_erratum_23154_cpus[] = {#endif #ifdef CONFIG_CAVIUM_ERRATUM_27456 -const struct midr_range cavium_erratum_27456_cpus[] = { +static const struct midr_range cavium_erratum_27456_cpus[] = {/* Cavium ThunderX, T88 pass 1.x - 2.1 */ MIDR_RANGE (MIDR_THUNDERX ,0 ,0 ,1 ,1 ), /* Cavium ThunderX, T81 pass 1.0 */
arch/arm64/kernel/image-vars.h +0 -4
@@ -47,10 +47,6 @@ PROVIDE(__pi_id_aa64smfr0_override = id_aa64smfr0_override);PROVIDE (__pi_id_aa64zfr0_override = id_aa64zfr0_override);PROVIDE (__pi_arm64_sw_feature_override = arm64_sw_feature_override);PROVIDE (__pi_arm64_use_ng_mappings = arm64_use_ng_mappings); -#ifdef CONFIG_CAVIUM_ERRATUM_27456 -PROVIDE (__pi_cavium_erratum_27456_cpus = cavium_erratum_27456_cpus); -PROVIDE (__pi_is_midr_in_range_list = is_midr_in_range_list); -#endif PROVIDE (__pi__ctype = _ctype);PROVIDE (__pi_memstart_offset_seed = memstart_offset_seed);
arch/arm64/kernel/pi/map_kernel.c +24 -1
@@ -207,6 +207,29 @@ static void __init map_fdt(u64 fdt)dsb (ishst); } +/* +* PI version of the Cavium Eratum 27456 detection, which makes it +* impossible to use non-global mappings. +*/ +static bool __init ng_mappings_allowed (void ) +{ +static const struct midr_range cavium_erratum_27456_cpus[]__initconst = { +/* Cavium ThunderX, T88 pass 1.x - 2.1 */ +MIDR_RANGE (MIDR_THUNDERX ,0 ,0 ,1 ,1 ), +/* Cavium ThunderX, T81 pass 1.0 */ +MIDR_REV (MIDR_THUNDERX_81XX ,0 ,0 ), + {}, + }; + +for (const struct midr_range *r = cavium_erratum_27456_cpus; r->model; r++) { +if (midr_is_cpu_model_range (read_cpuid_id (), r->model, + r->rv_min, r->rv_max)) +return false; + } + +return true; +} + asmlinkage void __init early_map_kernel(u64 boot_status, void *fdt) {static char const chosen_str[]__initconst ="/chosen" ;
@@ -246,7 +269,7 @@ asmlinkage void __init early_map_kernel(u64 boot_status, void *fdt)u64 kaslr_seed =kaslr_early_init (fdt, chosen);if (kaslr_seed &&kaslr_requires_kpti ()) - arm64_use_ng_mappings = true; + arm64_use_ng_mappings =ng_mappings_allowed (); kaslr_offset |= kaslr_seed & ~(MIN_KIMG_ALIGN -1 ); }
Signature · 1 line
2.39.2
Quoted 18 lines · Marc Zyngier
Calling into the MIDR checking framework from the PI code has recently become much harder, due to the new fancy "multi-MIDR" support that relies on tables being populated at boot time, but not that early that they are available to the PI code. There are additional issues with this framework, as the code really isn't position independend *at all*.
This leads to some ugly breakages, as reported by Ada.
It so appears that the only reason for the PI code to call into the MIDR checking code is to cope with The Most Broken ARM64 System Ever, aka Cavium ThunderX, which cannot deal with nG attributes that result of the combination of KASLR and KPTI as a consequence of Erratum 27456.
Duplicate the check for the erratum in the PI code, removing the dependency on the bulk of the MIDR checking framework. This allows dropping that same check from
kaslr_requires_kpti(), as the KPTI code already relies on theARM64_WORKAROUND_CAVIUM_27456cap.
Quoted 8 lines
Fixes c8c2647e69bed ("arm64: Make _midr_in_range_list() an exported function")Reported-by Ada Couprie Diaz <[email protected]>Signed-off-by Marc Zyngier <[email protected]>Cc Catalin Marinas <[email protected]>Cc Will Deacon <[email protected]>Cc Shameer Kolothum <[email protected]>Cc Oliver Upton <[email protected]>
I think the fastest path to Linus for this patch is through the arm64 tree. Catalin, in the interest of getting this fixed ASAP, could you pick this up?
I'll gladly take it otherwise.
Quoted 116 lines
---
Notes:
* From v1 [1]:- Preserved KASLR functionnality by duplicating the MIDR checks in the PI code. While this is a bit ugly, it keeps everything working for another day, and removes a duplicate check in the KPTI code.
- Tested in a VM to check that KASLR was still up and running. Yay!
[1] https://lore.kernel.org/r/[email protected]
arch/arm64/include/asm/mmu.h | 11 ----------- arch/arm64/kernel/cpu_errata.c | 2 +- arch/arm64/kernel/image-vars.h | 4 ---- arch/arm64/kernel/pi/map_kernel.c | 25 ++++++++++++++++++++++++- 4 files changed, 25 insertions(+), 17 deletions(-)arch/arm64/include/asm/mmu.h +0
-11@@ -94,17 +94,6 @@ static inline bool kaslr_requires_kpti(void)return false; } -/* -* Systems affected by Cavium erratum 24756 are incompatible -* with KPTI. -*/ -if (IS_ENABLED (CONFIG_CAVIUM_ERRATUM_27456 )) { -extern const struct midr_range cavium_erratum_27456_cpus[]; - -if (is_midr_in_range_list (cavium_erratum_27456_cpus)) -return false; - } -return true; }arch/arm64/kernel/cpu_errata.c +1
-1@@ -335,7 +335,7 @@ static const struct midr_range cavium_erratum_23154_cpus[] = {#endif #ifdef CONFIG_CAVIUM_ERRATUM_27456 -const struct midr_range cavium_erratum_27456_cpus[] = { +static const struct midr_range cavium_erratum_27456_cpus[] = {/* Cavium ThunderX, T88 pass 1.x - 2.1 */ MIDR_RANGE (MIDR_THUNDERX ,0 ,0 ,1 ,1 ), /* Cavium ThunderX, T81 pass 1.0 */arch/arm64/kernel/image-vars.h +0
-4@@ -47,10 +47,6 @@ PROVIDE(__pi_id_aa64smfr0_override = id_aa64smfr0_override);PROVIDE (__pi_id_aa64zfr0_override = id_aa64zfr0_override);PROVIDE (__pi_arm64_sw_feature_override = arm64_sw_feature_override);PROVIDE (__pi_arm64_use_ng_mappings = arm64_use_ng_mappings); -#ifdef CONFIG_CAVIUM_ERRATUM_27456 -PROVIDE (__pi_cavium_erratum_27456_cpus = cavium_erratum_27456_cpus); -PROVIDE (__pi_is_midr_in_range_list = is_midr_in_range_list); -#endif PROVIDE (__pi__ctype = _ctype);PROVIDE (__pi_memstart_offset_seed = memstart_offset_seed);arch/arm64/kernel/pi/map_kernel.c +24
-1@@ -207,6 +207,29 @@ static void __init map_fdt(u64 fdt)dsb (ishst); } +/* +* PI version of the Cavium Eratum 27456 detection, which makes it +* impossible to use non-global mappings. +*/ +static bool __init ng_mappings_allowed (void ) +{ +static const struct midr_range cavium_erratum_27456_cpus[]__initconst = { +/* Cavium ThunderX, T88 pass 1.x - 2.1 */ +MIDR_RANGE (MIDR_THUNDERX ,0 ,0 ,1 ,1 ), +/* Cavium ThunderX, T81 pass 1.0 */ +MIDR_REV (MIDR_THUNDERX_81XX ,0 ,0 ), + {}, + }; + +for (const struct midr_range *r = cavium_erratum_27456_cpus; r->model; r++) { +if (midr_is_cpu_model_range (read_cpuid_id (), r->model, + r->rv_min, r->rv_max)) +return false; + } + +return true; +} + asmlinkage void __init early_map_kernel(u64 boot_status, void *fdt) {static char const chosen_str[]__initconst ="/chosen" ;@@ -246,7 +269,7 @@ asmlinkage void __init early_map_kernel(u64 boot_status, void *fdt)u64 kaslr_seed =kaslr_early_init (fdt, chosen);if (kaslr_seed &&kaslr_requires_kpti ()) - arm64_use_ng_mappings = true; + arm64_use_ng_mappings =ng_mappings_allowed (); kaslr_offset |= kaslr_seed & ~(MIN_KIMG_ALIGN -1 ); }Signature · 1 line
2.39.2
Quoted 18 lines · Marc Zyngier
Calling into the MIDR checking framework from the PI code has recently become much harder, due to the new fancy "multi-MIDR" support that relies on tables being populated at boot time, but not that early that they are available to the PI code. There are additional issues with this framework, as the code really isn't position independend *at all*.
This leads to some ugly breakages, as reported by Ada.
It so appears that the only reason for the PI code to call into the MIDR checking code is to cope with The Most Broken ARM64 System Ever, aka Cavium ThunderX, which cannot deal with nG attributes that result of the combination of KASLR and KPTI as a consequence of Erratum 27456.
Duplicate the check for the erratum in the PI code, removing the dependency on the bulk of the MIDR checking framework. This allows dropping that same check from
kaslr_requires_kpti(), as the KPTI code already relies on theARM64_WORKAROUND_CAVIUM_27456cap.
Quoted 8 lines
Fixes c8c2647e69bed ("arm64: Make _midr_in_range_list() an exported function")Reported-by Ada Couprie Diaz <[email protected]>Signed-off-by Marc Zyngier <[email protected]>Cc Catalin Marinas <[email protected]>Cc Will Deacon <[email protected]>Cc Shameer Kolothum <[email protected]>Cc Oliver Upton <[email protected]>
Oliver, if you are in a timezone where you are still working, please pick it up. I might not have time until Monday otherwise.
Thanks.
Signature · 1 line
Catalin
Oliver Upton
AppliedQuoted 9 lines · Marc Zyngier
Calling into the MIDR checking framework from the PI code has recently become much harder, due to the new fancy "multi-MIDR" support that relies on tables being populated at boot time, but not that early that they are available to the PI code. There are additional issues with this framework, as the code really isn't position independend *at all*.
This leads to some ugly breakages, as reported by Ada.
[...]
Applied to fixes (and off to Paolo), thanks!
[1/1] arm64: Rework checks for broken Cavium HW in the PI code
https://git.kernel.org/kvmarm/kvmarm/c/117c3b21d3c7
--
Best,
Oliver