From: Ding Hui <dinghui@sangfor.com.cn> To: tony.luck@intel.com, bp@alien8.de, bp@suse.de, naoya.horiguchi@nec.com, osalvador@suse.de, peterz@infradead.org Cc: linux-edac@vger.kernel.org, linux-kernel@vger.kernel.org, tglx@linutronix.de, mingo@redhat.com, x86@kernel.org, hpa@zytor.com, youquan.song@intel.com, huangcun@sangfor.com.cn, stable@vger.kernel.org, Ding Hui <dinghui@sangfor.com.cn> Subject: [PATCH v1] x86/mce: Fix endless loop when run task works after #MC Date: Mon, 5 Jul 2021 20:59:21 +0800 [thread overview] Message-ID: <20210705125921.936-1-dinghui@sangfor.com.cn> (raw) Recently we encounter multi #MC on the same task when it's task_work_run() has not been called, current->mce_kill_me was added to task_works list more than once, that make a circular linked task_works, so task_work_run() will do a endless loop. More seriously, the SIGBUS signal can not be delivered to the userspace task which tigger the #MC and I met #MC flood. I borrowed mce_kill_me.func to check whether current->mce_kill_me has been added to task_works, prevent duplicate addition. When work function be called, the task_works must has been taken, so it is safe to be cleared in callback. Fixed: commit 5567d11c21a1 ("x86/mce: Send #MC singal from task work") Cc: <stable@vger.kernel.org> #v5.8+ Signed-off-by: Ding Hui <dinghui@sangfor.com.cn> --- arch/x86/kernel/cpu/mce/core.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c index 22791aadc085..32fb9ded6b85 100644 --- a/arch/x86/kernel/cpu/mce/core.c +++ b/arch/x86/kernel/cpu/mce/core.c @@ -1250,6 +1250,7 @@ static void __mc_scan_banks(struct mce *m, struct pt_regs *regs, struct mce *fin static void kill_me_now(struct callback_head *ch) { + WRITE_ONCE(ch->func, NULL); force_sig(SIGBUS); } @@ -1259,6 +1260,8 @@ static void kill_me_maybe(struct callback_head *cb) int flags = MF_ACTION_REQUIRED; int ret; + WRITE_ONCE(cb->func, NULL); + pr_err("Uncorrected hardware memory error in user-access at %llx", p->mce_addr); if (!p->mce_ripv) @@ -1289,17 +1292,20 @@ static void kill_me_maybe(struct callback_head *cb) static void queue_task_work(struct mce *m, int kill_current_task) { + struct callback_head ch; + current->mce_addr = m->addr; current->mce_kflags = m->kflags; current->mce_ripv = !!(m->mcgstatus & MCG_STATUS_RIPV); current->mce_whole_page = whole_page(m); if (kill_current_task) - current->mce_kill_me.func = kill_me_now; + ch.func = kill_me_now; else - current->mce_kill_me.func = kill_me_maybe; + ch.func = kill_me_maybe; - task_work_add(current, ¤t->mce_kill_me, TWA_RESUME); + if (!cmpxchg(¤t->mce_kill_me.func, NULL, ch.func)) + task_work_add(current, ¤t->mce_kill_me, TWA_RESUME); } /* -- 2.17.1
reply other threads:[~2021-07-05 13:08 UTC|newest] Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20210705125921.936-1-dinghui@sangfor.com.cn \ --to=dinghui@sangfor.com.cn \ --cc=bp@alien8.de \ --cc=bp@suse.de \ --cc=hpa@zytor.com \ --cc=huangcun@sangfor.com.cn \ --cc=linux-edac@vger.kernel.org \ --cc=linux-kernel@vger.kernel.org \ --cc=mingo@redhat.com \ --cc=naoya.horiguchi@nec.com \ --cc=osalvador@suse.de \ --cc=peterz@infradead.org \ --cc=stable@vger.kernel.org \ --cc=tglx@linutronix.de \ --cc=tony.luck@intel.com \ --cc=x86@kernel.org \ --cc=youquan.song@intel.com \ --subject='Re: [PATCH v1] x86/mce: Fix endless loop when run task works after #MC' \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).