linux-kernel - Re: Re: [PATCH] x86/bugs: Rate-limit the SPECTRE

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>] [day] [month] [year] [list]
Message-Id: <20250918020128.40658-1-liuq131@chinatelecom.cn>
Date: Thu, 18 Sep 2025 10:01:28 +0800
From: Qiang Liu <liuq131@...natelecom.cn>
To: pawan.kumar.gupta@...ux.intel.com
Cc: tglx@...utronix.de,
	bp@...en8.de,
	peterz@...radead.org,
	jpoimboe@...nel.org,
	mingo@...hat.com,
	dave.hansen@...ux.intel.com,
	x86@...nel.org,
	hpa@...or.com,
	linux-kernel@...r.kernel.org,
	Qiang Liu <liuq131@...natelecom.cn>
Subject: Re: Re: [PATCH] x86/bugs: Rate-limit the SPECTRE_V2 messages

On Tue, 16 Sep 2025 22:11:54 -0700, Pawan Gupta wrote:
>> During stress-ng testing, excessive SPECTRE_V2 messages overwhelmed the
>> low-speed serial device, causing system soft lockup. This issue can be
>> resolved by implementing rate-limiting for the relevant log outputs.
>> 
>> The log as below:
>> [121017.083236] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.098606] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.102398] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.102421] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.102532] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.102550] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.102569] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.133670] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.148497] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.163674] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.185720] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.287675] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.300205] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.335075] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.346428] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.428517] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.450328] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.592131] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121017.592865] Spectre V2 : WARNING: Unprivileged eBPF is enabled with eIBRS on, data leaks possible via Spectre v2 BHB attacks!
>> [121037.642577] watchdog: BUG: soft lockup - CPU#68 stuck for 23s! [stress-ng-procf:2483051]
>> [121037.642578] Modules linked in: ...
>> [121037.642697] CPU: 68 PID: 2483051 Comm: stress-ng-procf Kdump: loaded Tainted: ...
>> [121037.642698] Hardware name: XFUSION 2288H V6/BC13MBSBC, BIOS 1.29 11/25/2022
>> [121037.642706] RIP: 0010:console_unlock+0x283/0x350
>> [121037.642709] Code: 00 e8 01 15 00 00 55 9d 45 84 f6 0f 84 46 ff ff ff e8 71 f8 ff ff 85 c0 0f 85 e8 fd ff ff e9 34 ff ff ff e8 df 14 00 00 55 9d <8b> 44 24 04 85 c0 0f 84 f8 fd ff ff e8 9c 84 99 00 e9 ee fd ff ff
>> [121037.642710] RSP: 0018:ff607726b60abc00 EFLAGS: 00000246
>> [121037.642711] RAX: 0000000000000000 RBX: ffffffffaef699cc RCX: 0000000000000008
>> [121037.642712] RDX: 0000000000000000 RSI: 0000000000000004 RDI: ffffffffaef6c228
>> [121037.642713] RBP: 0000000000000246 R08: ffffffffadabd540 R09: 0000000000aaaaaa
>> [121037.642713] R10: 0000000000000001 R11: ff2cf962c6a5b550 R12: ff2cf962c2700000
>> [121037.642714] R13: 0000000000000000 R14: ffffffffaef6c228 R15: 000000000000008c
>> [121037.642715] FS:  00007f2021c07640(0000) GS:ff2cf9e0be700000(0000) knlGS:0000000000000000
>> [121037.642715] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [121037.642716] CR2: 0000000000e3b7a0 CR3: 0000003f41266005 CR4: 0000000000771ee0
>> [121037.642716] PKRU: 55555554
>> [121037.642717] Call Trace:
>> [121037.642722]  vprintk_emit+0x118/0x120
>> [121037.642727]  printk+0x58/0x73
>> [121037.642731]  bpf_unpriv_handler+0xbf/0x180
>> [121037.642733]  ? proc_taint+0x1d0/0x1d0
>> [121037.642738]  proc_sys_call_handler+0x13e/0x250
>> [121037.642742]  new_sync_read+0x10d/0x1b0
>> [121037.642745]  vfs_read+0x14e/0x1b0
>> [121037.642747]  ksys_read+0x5f/0xe0
>> [121037.642750]  do_syscall_64+0x3d/0x80
>> [121037.642753]  entry_SYSCALL_64_after_hwframe+0x61/0xc6
>> 
>> Fixes: 0de05d056afd ("x86/speculation: Warn about eIBRS + LFENCE + Unprivileged eBPF + SMT")
>> Fixes: 44a3918c8245 ("x86/speculation: Include unprivileged eBPF status in Spectre v2 mitigation reporting")
>> Signed-off-by: Qiang Liu <liuq131@...natelecom.cn>
>> ---
>>  arch/x86/kernel/cpu/bugs.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>> 
>> diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
>> index 36dcfc5105be..22fb2113cbe6 100644
>> --- a/arch/x86/kernel/cpu/bugs.c
>> +++ b/arch/x86/kernel/cpu/bugs.c
>> @@ -1809,11 +1809,11 @@ void unpriv_ebpf_notify(int new_state)
>>  
>>  	switch (spectre_v2_enabled) {
>>  	case SPECTRE_V2_EIBRS:
>> -		pr_err(SPECTRE_V2_EIBRS_EBPF_MSG);
>> +		pr_err_ratelimited(SPECTRE_V2_EIBRS_EBPF_MSG);
>>  		break;
>>  	case SPECTRE_V2_EIBRS_LFENCE:
>>  		if (sched_smt_active())
>> -			pr_err(SPECTRE_V2_EIBRS_LFENCE_EBPF_SMT_MSG);
>> +			pr_err_ratelimited(SPECTRE_V2_EIBRS_LFENCE_EBPF_SMT_MSG);
>
>Just a thought, pr_err_once() might be sufficient, repeated warnings
>doesn't add much value.
>
Thank you for your valuable suggestion. However, I still believe pr_err_ratelimited
would be better, as it ensures relevant error messages are printed during each test
execution, preventing the test tool from incorrectly assuming the test passed.

>>  		break;
>>  	default:
>>  		break;
>> -- 
>> 2.46.0
>>