linux-kernel - Re: [PATCH] arm64: cpufeature: Enable Qualcomm erratas

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1d1a3dca-16ce-f541-5d78-e61ad24227e0@arm.com>
Date:   Tue, 29 Oct 2019 12:04:21 +0000
From:   Suzuki K Poulose <suzuki.poulose@....com>
To:     will@...nel.org, bjorn.andersson@...aro.org
Cc:     catalin.marinas@....com, linux-arm-kernel@...ts.infradead.org,
        linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org,
        stable@...r.kernel.org, broonie@...nel.org
Subject: Re: [PATCH] arm64: cpufeature: Enable Qualcomm erratas

On 10/29/2019 11:39 AM, Will Deacon wrote:
> On Mon, Oct 28, 2019 at 11:04:32PM -0700, Bjorn Andersson wrote:
>> With the introduction of 'cce360b54ce6 ("arm64: capabilities: Filter the
>> entries based on a given mask")' the Qualcomm erratas are no long
>> applied.
>>
>> The result of not applying errata 1003 is that MSM8996 runs into various
>> RCU stalls and fails to boot most of the times.
>>
>> Give both 1003 and 1009 a "type" to ensure they are not filtered out in
>> update_cpu_capabilities().
> 
> Oh nasty. Thanks for debugging and fixing this.
> 
>> Fixes: cce360b54ce6 ("arm64: capabilities: Filter the entries based on a given mask")
>> Cc: stable@...r.kernel.org
>> Reported-by: Mark Brown <broonie@...nel.org>
>> Suggested-by: Will Deacon <will@...nel.org>
>> Signed-off-by: Bjorn Andersson <bjorn.andersson@...aro.org>
>> ---
>>   arch/arm64/kernel/cpu_errata.c | 2 ++
>>   1 file changed, 2 insertions(+)
>>
>> diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
>> index df9465120e2f..cdd8df033536 100644
>> --- a/arch/arm64/kernel/cpu_errata.c
>> +++ b/arch/arm64/kernel/cpu_errata.c
>> @@ -780,6 +780,7 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
>>   	{
>>   		.desc = "Qualcomm Technologies Falkor/Kryo erratum 1003",
>>   		.capability = ARM64_WORKAROUND_QCOM_FALKOR_E1003,
>> +		.type = ARM64_CPUCAP_SCOPE_LOCAL_CPU,
>>   		.matches = cpucap_multi_entry_cap_matches,
> 
> This should probably be ARM64_CPUCAP_LOCAL_CPU_ERRATUM instead, but I'll
> want Suzuki's ack before I take the change.

Yes, it must be ARM64_CPUCAP_LOCAL_CPU_ERRATUM.

It may be a good idea to stick in a check to make sure that the scope is
set for all the capabilities in a separate patch. e.g,

diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index d260e3bdf07b..51a79b4a44eb 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -546,6 +546,8 @@ static void __init
  init_cpu_hwcaps_indirect_list_from_array(const struct 
arm64_cpu_capabilities *caps)
  {
  	for (; caps->matches; caps++) {
+		WARN(!cpucap_default_scope(caps),
+		     "Invalid scope for capability %d\n", caps->capability);
  		if (WARN(caps->capability >= ARM64_NCAPS,
  			"Invalid capability %d\n", caps->capability))
			continue;

Otherwise looks good to me.

>>   		.match_list = qcom_erratum_1003_list,
>>   	},
>> @@ -788,6 +789,7 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
>>   	{
>>   		.desc = "Qualcomm erratum 1009, ARM erratum 1286807",
>>   		.capability = ARM64_WORKAROUND_REPEAT_TLBI,
>> +		.type = ARM64_CPUCAP_SCOPE_LOCAL_CPU,
>>   		ERRATA_MIDR_RANGE_LIST(arm64_repeat_tlbi_cpus),
> 
> ERRATA_MIDR_RANGE_LIST sets the type already, so I think this is redundant.
> 
> Will
> 



Cheers
Suzuki