From patchwork Thu Sep 8 16:33:10 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: hotran X-Patchwork-Id: 9321773 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id A48E26077F for ; Thu, 8 Sep 2016 16:34:07 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 949E029906 for ; Thu, 8 Sep 2016 16:34:07 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 8939029908; Thu, 8 Sep 2016 16:34:07 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-1.8 required=2.0 tests=BAYES_00,DKIM_SIGNED, T_DKIM_INVALID autolearn=no version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id B9F6829906 for ; Thu, 8 Sep 2016 16:34:06 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S938956AbcIHQeG (ORCPT ); Thu, 8 Sep 2016 12:34:06 -0400 Received: from mail-pa0-f43.google.com ([209.85.220.43]:36695 "EHLO mail-pa0-f43.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S938429AbcIHQeF (ORCPT ); Thu, 8 Sep 2016 12:34:05 -0400 Received: by mail-pa0-f43.google.com with SMTP id id6so18898810pad.3 for ; Thu, 08 Sep 2016 09:34:04 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=apm.com; s=apm; h=from:to:cc:subject:date:message-id; bh=Pkpt0CzBfW/6pFC3g09zyGAoAwDRVbgReM/X5YFD3rE=; b=mUUc027QQJCj2Di14K/LPEp6KC2OQ7S7t0MU/d2pBV1rY2rEK3MNnMk4wc22pBZD6/ PUC/Cz9/GIU1vvI1ogqkHWRpdBZUsu0joiuuCik2lMRv3aj5gt/FtMKJIwesYyJpYJQV 9863P00UeSSVz/J+rc2ysDxW5qBY3rJKqJhFM= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:from:to:cc:subject:date:message-id; bh=Pkpt0CzBfW/6pFC3g09zyGAoAwDRVbgReM/X5YFD3rE=; b=Xu0rSKiByzbSTX1g/NB2T1eeVw3ayQiERudyTrma93nfDMhMGKKMmHiqSKQQOp9T5J VNu779EPJZH6i9m4me9NjsxyY639iuRHXcsiQEMJnMgFftMmJMzY7IfBW2FDjxBMkGFD mTUTI/DHoCFA6ct+8oaPQNRqJZbJo5PoZf/Sn0g3V5Cov7Y/FES8GHuykm/KEW3cOwOI KsiYeQaOQwl8uX+F3h2vwbyIsbC5b6qXM6vnSI5Sdm4PIc2ecfx13NPHQkYnQin+qo7O PM1dtODnS1Fcfke7fyJFN7FYbvyB4x7mfr4L/TvZY9+jDwDOtAUW/7zcjqpyFTprc101 iSog== X-Gm-Message-State: AE9vXwPANUhwT/iP3jxuORdw64WUNC6noLWU73W7A9cpBjx2qYhoCeiuRWDvYot6vcYpvyvr X-Received: by 10.66.79.138 with SMTP id j10mr960522pax.60.1473352444341; Thu, 08 Sep 2016 09:34:04 -0700 (PDT) Received: from hotran_localhost.amcc.com ([206.80.4.98]) by smtp.gmail.com with ESMTPSA id p74sm41743875pfk.68.2016.09.08.09.34.02 (version=TLS1_2 cipher=ECDHE-RSA-AES128-SHA bits=128/128); Thu, 08 Sep 2016 09:34:03 -0700 (PDT) From: Hoan Tran To: Guenter Roeck , Jean Delvare Cc: linux-hwmon@vger.kernel.org, linux-kernel@vger.kernel.org, Itaru Kitayama , lho@apm.com, Duc Dang , Hoan Tran Subject: [PATCH v3] hwmon: xgene: Fix crash when alarm occurs before driver probe Date: Thu, 8 Sep 2016 09:33:10 -0700 Message-Id: <1473352390-30781-1-git-send-email-hotran@apm.com> X-Mailer: git-send-email 1.9.1 Sender: linux-hwmon-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-hwmon@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP The system crashes during probing xgene-hwmon driver when temperature alarm interrupt occurs before. It's because - xgene_hwmon_probe() requests mailbox channel which also enables the mailbox interrupt. - As temperature alarm interrupt is pending, ISR runs and crashes when accesses into invalid resourse as unmapped PCC shared memory. This patch fixes this issue by saving this alarm message and scheduling a bottom handler after xgene_hwmon_probe() finish. Signed-off-by: Hoan Tran Reported-by: Itaru Kitayama --- v3 * Replace IS_ERR by IS_ERR_OR_NULL v2 * Check hwmon_dev and resp_pending to determine the driver is not ready * Remove init_flag and rx_pending * Replace EBUSY by ENODEV v1 * Initial drivers/hwmon/xgene-hwmon.c | 69 ++++++++++++++++++++++++++++++++------------- 1 file changed, 50 insertions(+), 19 deletions(-) diff --git a/drivers/hwmon/xgene-hwmon.c b/drivers/hwmon/xgene-hwmon.c index bc78a5d..aa44579 100644 --- a/drivers/hwmon/xgene-hwmon.c +++ b/drivers/hwmon/xgene-hwmon.c @@ -465,13 +465,34 @@ static void xgene_hwmon_evt_work(struct work_struct *work) } } +static int xgene_hwmon_rx_ready(struct xgene_hwmon_dev *ctx, void *msg) +{ + if (IS_ERR_OR_NULL(ctx->hwmon_dev) && !ctx->resp_pending) { + /* Enqueue to the FIFO */ + kfifo_in_spinlocked(&ctx->async_msg_fifo, msg, + sizeof(struct slimpro_resp_msg), + &ctx->kfifo_lock); + return -ENODEV; + } + + return 0; +} + /* * This function is called when the SLIMpro Mailbox received a message */ static void xgene_hwmon_rx_cb(struct mbox_client *cl, void *msg) { struct xgene_hwmon_dev *ctx = to_xgene_hwmon_dev(cl); - struct slimpro_resp_msg amsg; + + /* + * While the driver registers with the mailbox framework, an interrupt + * can be pending before the probe function completes its + * initialization. If such condition occurs, just queue up the message + * as the driver is not ready for servicing the callback. + */ + if (xgene_hwmon_rx_ready(ctx, msg) < 0) + return; /* * Response message format: @@ -500,12 +521,8 @@ static void xgene_hwmon_rx_cb(struct mbox_client *cl, void *msg) return; } - amsg.msg = ((u32 *)msg)[0]; - amsg.param1 = ((u32 *)msg)[1]; - amsg.param2 = ((u32 *)msg)[2]; - /* Enqueue to the FIFO */ - kfifo_in_spinlocked(&ctx->async_msg_fifo, &amsg, + kfifo_in_spinlocked(&ctx->async_msg_fifo, msg, sizeof(struct slimpro_resp_msg), &ctx->kfifo_lock); /* Schedule the bottom handler */ schedule_work(&ctx->workq); @@ -520,6 +537,15 @@ static void xgene_hwmon_pcc_rx_cb(struct mbox_client *cl, void *msg) struct acpi_pcct_shared_memory *generic_comm_base = ctx->pcc_comm_addr; struct slimpro_resp_msg amsg; + /* + * While the driver registers with the mailbox framework, an interrupt + * can be pending before the probe function completes its + * initialization. If such condition occurs, just queue up the message + * as the driver is not ready for servicing the callback. + */ + if (xgene_hwmon_rx_ready(ctx, &amsg) < 0) + return; + msg = generic_comm_base + 1; /* Check if platform sends interrupt */ if (!xgene_word_tst_and_clr(&generic_comm_base->status, @@ -596,6 +622,17 @@ static int xgene_hwmon_probe(struct platform_device *pdev) platform_set_drvdata(pdev, ctx); cl = &ctx->mbox_client; + spin_lock_init(&ctx->kfifo_lock); + mutex_init(&ctx->rd_mutex); + + rc = kfifo_alloc(&ctx->async_msg_fifo, + sizeof(struct slimpro_resp_msg) * ASYNC_MSG_FIFO_SIZE, + GFP_KERNEL); + if (rc) + goto out_mbox_free; + + INIT_WORK(&ctx->workq, xgene_hwmon_evt_work); + /* Request mailbox channel */ cl->dev = &pdev->dev; cl->tx_done = xgene_hwmon_tx_done; @@ -676,17 +713,6 @@ static int xgene_hwmon_probe(struct platform_device *pdev) ctx->usecs_lat = PCC_NUM_RETRIES * cppc_ss->latency; } - spin_lock_init(&ctx->kfifo_lock); - mutex_init(&ctx->rd_mutex); - - rc = kfifo_alloc(&ctx->async_msg_fifo, - sizeof(struct slimpro_resp_msg) * ASYNC_MSG_FIFO_SIZE, - GFP_KERNEL); - if (rc) - goto out_mbox_free; - - INIT_WORK(&ctx->workq, xgene_hwmon_evt_work); - ctx->hwmon_dev = hwmon_device_register_with_groups(ctx->dev, "apm_xgene", ctx, @@ -697,17 +723,22 @@ static int xgene_hwmon_probe(struct platform_device *pdev) goto out; } + /* + * Schedule the bottom handler if there is a pending message. + */ + schedule_work(&ctx->workq); + dev_info(&pdev->dev, "APM X-Gene SoC HW monitor driver registered\n"); return 0; out: - kfifo_free(&ctx->async_msg_fifo); -out_mbox_free: if (acpi_disabled) mbox_free_channel(ctx->mbox_chan); else pcc_mbox_free_channel(ctx->mbox_chan); +out_mbox_free: + kfifo_free(&ctx->async_msg_fifo); return rc; }