Received: by 2002:ac0:a5b6:0:0:0:0:0 with SMTP id m51-v6csp4380883imm; Wed, 30 May 2018 04:44:00 -0700 (PDT) X-Google-Smtp-Source: ADUXVKLEj3lKb0SjhD9eyQ71Loj8HLHDZCaXwG7grpjy/YRMsxduDkgUm4/L/B6XdnM5Fz+xImMB X-Received: by 2002:a62:df89:: with SMTP id d9-v6mr2430323pfl.147.1527680640457; Wed, 30 May 2018 04:44:00 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1527680640; cv=none; d=google.com; s=arc-20160816; b=R/Pyl+7qg37+UBClUFnm7UtwUCocVOHyFgJiyDgzir/b/yJhsOJ/q9Nf/ewN0IEVAJ Y4yfgAlLHMW8OEtwET1FAWuehs64JsLIHVPIrV51BUlG/xDQBCNw4hm9gAPPPdxrYhpa sWDjSz3HVFbWT4bumR7Q1S5x1YMG9t2MBSVR702u4dkJUF2Ynl/Mw3PkKIFhrh38coe0 RyIGhAw7DEg6OUcSSNp5wc2HIqZ5j5mmogsJtbfslrC6Hz1g0R7ebZ7L2oV7p6m4UiOL kWrNBPFQjFXCrtv1PJrwc89BxpTQFSZWPZXohr8IEFV9UuguXTJHtG7qr3HgoOQjyY+C rQDA== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:sender:user-agent:in-reply-to :content-disposition:mime-version:references:message-id:subject:cc :to:from:date:arc-authentication-results; bh=3jMn9iVH8GB3gJVhtPx+IWUlv1Jc4thjMfmLx+Cz2/4=; b=CV/e2+09s/+oYdGtoAPhD5RBayGniQ5k15Btd4w1GejKkFF/gfgpeNWlMXIhs8P3LP Kpezh2w4E32NobPg/7OYdtjK5PsPJzrxcCV4xqO368oMeiOOG9AEbf6M8ydrnscMKU2u gCvbTj3WLNp/GJuv0IQcU6PVxweTx1hEH821BT1CvRbDui17WoVGnv6sbaAFcHs1/vY2 2zpQYESZqGKdxO5Et3QFShdAOmdGxtf3jO1/JGZ7pPKNcEoktysqPLtbxHW3CqZTi6eh HoHhio1SVrdBeUqoOc8BEvtEJAJBbBOnHlITedsAf+BHzDrhDkSmrfXTp9QjP+KDuA9d rBww== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=intel.com Return-Path: Received: from vger.kernel.org (vger.kernel.org. [209.132.180.67]) by mx.google.com with ESMTP id v8-v6si34360648plo.306.2018.05.30.04.43.46; Wed, 30 May 2018 04:44:00 -0700 (PDT) Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67; Authentication-Results: mx.google.com; spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=intel.com Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753201AbeE3Llb (ORCPT + 99 others); Wed, 30 May 2018 07:41:31 -0400 Received: from mga03.intel.com ([134.134.136.65]:9328 "EHLO mga03.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753113AbeE3Ll2 (ORCPT ); Wed, 30 May 2018 07:41:28 -0400 X-Amp-Result: UNSCANNABLE X-Amp-File-Uploaded: False Received: from orsmga006.jf.intel.com ([10.7.209.51]) by orsmga103.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 30 May 2018 04:41:27 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.49,460,1520924400"; d="scan'208";a="45906704" Received: from raj-desk2.iind.intel.com ([10.223.107.30]) by orsmga006.jf.intel.com with ESMTP; 30 May 2018 04:41:24 -0700 Date: Wed, 30 May 2018 17:03:29 +0530 From: Rajneesh Bhardwaj To: "David E. Box" Cc: andy@infradead.org, vishwanath.somayaji@intel.com, dvhart@infradead.org, kyle.d.pelton@linux.intel.com, platform-driver-x86@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH V2] platform/x86: intel_pmc_core: Add CNP SLPS0 debug registers Message-ID: <20180530113329.GA24455@raj-desk2.iind.intel.com> References: <20180525011056.25132-1-david.e.box@linux.intel.com> <20180528070032.GA6321@raj-desk2.iind.intel.com> <1527677592.5247.24.camel@linux.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1527677592.5247.24.camel@linux.intel.com> User-Agent: Mutt/1.5.24 (2015-08-30) Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, May 30, 2018 at 03:53:12AM -0700, David E. Box wrote: Hi Dave, > Hi Rajneesh, > > On Mon, 2018-05-28 at 12:30 +0530, Rajneesh Bhardwaj wrote: > > On Thu, May 24, 2018 at 06:10:56PM -0700, David E. Box wrote: > > > > Thanks for sending this, Dave. Few comments below. > > > > > Adds debugfs access to registers in the Cannon Point PCH PMC that > > > are > > > > Please use Cannonlake PCH. > > > > > useful for debugging #SLP_S0 signal assertion and other low power > > > related > > > > assertion failures and other low power debug. > > > > > activities. Device pm states are latched in these registers > > > whenever the > > > package enters C10 and can be read from slp_s0_debug_status. The pm > > > states may also be latched by writing 1 to slp_s0_dbg_latch which > > > will > > > immediately capture the current state on the next read of > > > slp_s0_debug_status. Also while in intel_pmc_core.h clean up > > > spacing. > > > > > > > Initially, unless the latch bit is not set the debugfs file does not > > show > > any information as expected but once you write Y to latch file, the > > debugfs > > file continues to show blockers even though you write N back there or > > unload > > - reload the driver. Please fix this. > > See comments below > > > > > > Signed-off-by: David E. Box > > > --- > > > > > > V2: > > > Per Andy Shevchenko: > > > - Clear latch bit after use > > > - Pass pmc_dev as parameter > > > - Use DEFINE_SHOW_ATTRIBUTE macro > > > > > > drivers/platform/x86/intel_pmc_core.c | 112 > > > ++++++++++++++++++++++++++++++++++ > > > drivers/platform/x86/intel_pmc_core.h | 27 +++++--- > > > 2 files changed, 132 insertions(+), 7 deletions(-) > > > > > > diff --git a/drivers/platform/x86/intel_pmc_core.c > > > b/drivers/platform/x86/intel_pmc_core.c > > > index 43bbe74..ed40999 100644 > > > --- a/drivers/platform/x86/intel_pmc_core.c > > > +++ b/drivers/platform/x86/intel_pmc_core.c > > > @@ -196,9 +196,64 @@ static const struct pmc_bit_map > > > cnp_pfear_map[] = { > > > {} > > > }; > > > > > > +static const struct pmc_bit_map cnp_slps0_dbg0_map[] = { > > > + {"AUDIO_D3", BIT(0)}, > > > + {"OTG_D3", BIT(1)}, > > > + {"XHCI_D3", BIT(2)}, > > > + {"LPIO_D3", BIT(3)}, > > > + {"SDX_D3", BIT(4)}, > > > + {"SATA_D3", BIT(5)}, > > > + {"UFS0_D3", BIT(6)}, > > > + {"UFS1_D3", BIT(7)}, > > > + {"EMMC_D3", BIT(8)}, > > > +}; > > > + > > > +static const struct pmc_bit_map cnp_slps0_dbg1_map[] = { > > > + {"SDIO_PLL_OFF", BIT(0)}, > > > + {"USB2_PLL_OFF", BIT(1)}, > > > + {"AUDIO_PLL_OFF", BIT(2)}, > > > + {"OC_PLL_OFF", BIT(3)}, > > > > Please rename to ISCLK_OC as it is the preferred nomenclature for > > debug. > > Okay > > > > > > + {"MAIN_PLL_OFF", BIT(4)}, > > > > Ditto, please use ISCLK_MAIN. > > > > > + {"XOSC_OFF", BIT(5)}, > > > + {"LPC_CLKS_GATED", BIT(6)}, > > > + {"PCIE_CLKREQS_IDLE", BIT(7)}, > > > + {"AUDIO_ROSC_OFF", BIT(8)}, > > > + {"HPET_XOSC_CLK_REQ", BIT(9)}, > > > + {"PMC_ROSC_SLOW_CLK", BIT(10)}, > > > + {"AON2_ROSC_GATED", BIT(11)}, > > > + {"CLKACKS_DEASSERTED", BIT(12)}, > > > +}; > > > + > > > +static const struct pmc_bit_map cnp_slps0_dbg2_map[] = { > > > + {"MPHY_CORE_GATED", BIT(0)}, > > > + {"CSME_GATED", BIT(1)}, > > > + {"USB2_SUS_GATED", BIT(2)}, > > > + {"DYN_FLEX_IO_IDLE", BIT(3)}, > > > + {"GBE_NO_LINK", BIT(4)}, > > > + {"THERM_SEN_DISABLED", BIT(5)}, > > > + {"PCIE_LOW_POWER", BIT(6)}, > > > + {"ISH_VNNAON_REQ_ACT", BIT(7)}, > > > + {"ISH_VNN_REQ_ACT", BIT(8)}, > > > + {"CNV_VNNAON_REQ_ACT", BIT(9)}, > > > + {"CNV_VNN_REQ_ACT", BIT(10)}, > > > + {"NPK_VNNON_REQ_ACT", BIT(11)}, > > > + {"PMSYNC_STATE_IDLE", BIT(12)}, > > > + {"ALST_GT_THRES", BIT(13)}, > > > + {"PMC_ARC_PG_READY", BIT(14)}, > > > +}; > > > + > > > +static const struct slps0_dbg_map cnp_slps0_dbg_maps[] = { > > > + {cnp_slps0_dbg0_map, ARRAY_SIZE(cnp_slps0_dbg0_map)}, > > > + {cnp_slps0_dbg1_map, ARRAY_SIZE(cnp_slps0_dbg1_map)}, > > > + {cnp_slps0_dbg2_map, ARRAY_SIZE(cnp_slps0_dbg2_map)}, > > > +}; > > > + > > > static const struct pmc_reg_map cnp_reg_map = { > > > .pfear_sts = cnp_pfear_map, > > > .slp_s0_offset = CNP_PMC_SLP_S0_RES_COUNTER_OFFSET, > > > + .slps0_dbg_maps = cnp_slps0_dbg_maps, > > > + .slps0_dbg_offset = CNP_PMC_SLPS0_DBG_OFFSET, > > > + .slps0_dbg_num = ARRAY_SIZE(cnp_slps0_dbg_maps), > > > .ltr_ignore_offset = CNP_PMC_LTR_IGNORE_OFFSET, > > > .regmap_length = CNP_PMC_MMIO_REG_LEN, > > > .ppfear0_offset = CNP_PMC_HOST_PPFEAR0A, > > > @@ -252,6 +307,8 @@ static int pmc_core_check_read_lock_bit(void) > > > } > > > > > > #if IS_ENABLED(CONFIG_DEBUG_FS) > > > +static bool slps0_dbg_latch; > > > + > > > static void pmc_core_display_map(struct seq_file *s, int index, > > > u8 pf_reg, const struct > > > pmc_bit_map *pf_map) > > > { > > > @@ -481,6 +538,52 @@ static const struct file_operations > > > pmc_core_ltr_ignore_ops = { > > > .release = single_release, > > > }; > > > > > > +static void pmc_core_slps0_dbg_latch(struct pmc_dev *pmcdev, bool > > > reset) > > > +{ > > > + const struct pmc_reg_map *map = pmcdev->map; > > > + u32 fd; > > > + > > > + mutex_lock(&pmcdev->lock); > > > + > > > + if (!reset && !slps0_dbg_latch) > > > + goto out_unlock; > > > + > > > + fd = pmc_core_reg_read(pmcdev, map->slps0_dbg_offset); > > > + reset ? (fd &= ~CNP_PMC_LATCH_SLPS0_EVENTS) : > > > + (fd |= CNP_PMC_LATCH_SLPS0_EVENTS); > > > + pmc_core_reg_write(pmcdev, map->slps0_dbg_offset, fd); > > > + > > > + slps0_dbg_latch = 0; > > > + > > > +out_unlock: > > > + mutex_unlock(&pmcdev->lock); > > > +} > > > + > > > +static int pmc_core_slps0_dbg_show(struct seq_file *s, void > > > *unused) > > > +{ > > > + struct pmc_dev *pmcdev = s ? s->private : &pmc; > > > + const struct slps0_dbg_map *maps = pmcdev->map- > > > >slps0_dbg_maps; > > > + int num_slps0_dbg_regs = pmcdev->map->slps0_dbg_num; > > > + int i, offset; > > > + u32 data; > > > + > > > + pmc_core_slps0_dbg_latch(pmcdev, false); > > > + offset = pmcdev->map->slps0_dbg_offset; > > > + while (num_slps0_dbg_regs--) { > > > + data = pmc_core_reg_read(pmcdev, offset); > > > + offset += 4; > > > + for (i = 0; i < maps->size; i++) > > > + seq_printf(s, "SLP_S0_DBG: %-32s\tState: > > > %s\n", > > > + maps->slps0_dbg_sts[i].name, > > > + data & maps- > > > >slps0_dbg_sts[i].bit_mask ? > > > + "Yes" : "No"); > > > > In current form, it looks pretty similar to > > pch_ip_power_gating_status > > output. Since it helps in debug, please use Blocks_slp_s0: instead of > > State: > > > > That may not be true. What blocks or doesn't can be configured by the > OEM. IIUC, The IPs exposed by these registers are the ones that the PMC always looks at before making a decision to assert SLP_S0#. I dont think they can be changed by OEMs in determining the S0ix flows. > > > > + maps++; > > > + } > > > + pmc_core_slps0_dbg_latch(pmcdev, true); > > > + return 0; > > > +} > > > +DEFINE_SHOW_ATTRIBUTE(pmc_core_slps0_dbg); > > > > When the latch bit is not set, can we format the output of this > > attribute > > stating clearly whether the current blockers are "live" or the "last > > captured at PC10 entry" ? > > There is no "live". It's always the last captured state which is why > the status doesn't update the way you expect above. With V2 the latch > is automatically reset after every read. > > When the latch isn't set, any updates would indeed be from PC10 > transitions, but you can't be sure this isn't the last forced latch > state unless you monitor PC10 residency and assume that the values were > updated when the count changes. Even with the latch, if there's a PC10 > transition that occurs after the bit is written but before the > registers are read, the values may be from that transition, not the > latch. All default values are printed as "No". Here, we can rather print a message that unless latch is set or Package enters c10 (which is unlikely unless the display is off) the debugfs would not show any blockers. In the current form this patch prints the output when we read the first time before setting the latch. SLP_S0_DBG: AUDIO_D3 State: No SLP_S0_DBG: OTG_D3 State: No ---- xxx ---- snip ---- xxx ---- Once the latch bit is set the SLP_S0_DBG_X registers show proper values but after explicitely setting latch as "N", the stale values are printed. The system on which i tested never enters PC10 so once the latch is unset, the values should be shown as default i.e. NO but that is not always correct as you mentioned above. IMO, a better approch would be to set/unset the latch internally when the user does a read on SLP_S0_DEBUG_X registers rather than exposing a seperate debugfs file because a write of 0 to the latch bit has no effect but it is a requirement for the next read. > > > > + > > > static void pmc_core_dbgfs_unregister(struct pmc_dev *pmcdev) > > > { > > > debugfs_remove_recursive(pmcdev->dbgfs_dir); > > > @@ -514,6 +617,15 @@ static int pmc_core_dbgfs_register(struct > > > pmc_dev *pmcdev) > > > 0444, dir, pmcdev, > > > &pmc_core_mphy_pg_ops); > > > > > > + if (pmcdev->map->slps0_dbg_maps) { > > > + debugfs_create_file("slp_s0_debug_status", 0444, > > > > slp_s0_dbg_latch > > > > > + dir, pmcdev, > > > + &pmc_core_slps0_dbg_fops); > > > + > > > + debugfs_create_bool("slp_s0_dbg_latch", 0644, > > > + dir, &slps0_dbg_latch); > > > + } > > > + > > > return 0; > > > } > > > #else > > > diff --git a/drivers/platform/x86/intel_pmc_core.h > > > b/drivers/platform/x86/intel_pmc_core.h > > > index 5fa5f97..97e4b5a 100644 > > > --- a/drivers/platform/x86/intel_pmc_core.h > > > +++ b/drivers/platform/x86/intel_pmc_core.h > > > @@ -124,27 +124,35 @@ enum ppfear_regs { > > > #define SPT_PMC_BIT_MPHY_CMN_LANE3 BIT(3) > > > > > > /* Cannonlake Power Management Controller register offsets */ > > > -#define CNP_PMC_SLP_S0_RES_COUNTER_OFFSET 0x193C > > > -#define CNP_PMC_LTR_IGNORE_OFFSET 0x1B0C > > > -#define CNP_PMC_PM_CFG_OFFSET 0x1818 > > > +#define CNP_PMC_SLP_S0_RES_COUNTER_OFFSET 0x193C > > > +#define CNP_PMC_LTR_IGNORE_OFFSET 0x1B0C > > > +#define CNP_PMC_PM_CFG_OFFSET 0x1818 > > > +#define CNP_PMC_SLPS0_DBG_OFFSET 0x10B4 > > > /* Cannonlake: PGD PFET Enable Ack Status Register(s) start */ > > > -#define CNP_PMC_HOST_PPFEAR0A 0x1D90 > > > +#define CNP_PMC_HOST_PPFEAR0A 0x1D90 > > > > > > -#define CNP_PMC_MMIO_REG_LEN 0x2000 > > > -#define CNP_PPFEAR_NUM_ENTRIES 8 > > > -#define CNP_PMC_READ_DISABLE_BIT 22 > > > +#define CNP_PMC_MMIO_REG_LEN 0x2000 > > > +#define CNP_PPFEAR_NUM_ENTRIES 8 > > > +#define CNP_PMC_READ_DISABLE_BIT 22 > > > +#define CNP_PMC_LATCH_SLPS0_EVENTS BIT(31) > > > > > > struct pmc_bit_map { > > > const char *name; > > > u32 bit_mask; > > > }; > > > > > > +struct slps0_dbg_map { > > > + const struct pmc_bit_map *slps0_dbg_sts; > > > + int size; > > > +}; > > > + > > > /** > > > * struct pmc_reg_map - Structure used to define parameter unique > > > to a > > > PCH family > > > * @pfear_sts: Maps name of IP block to PPFEAR* bit > > > * @mphy_sts: Maps name of MPHY lane to MPHY status > > > lane status bit > > > * @pll_sts: Maps name of PLL to corresponding bit > > > status > > > + * @slps0_dbg_maps: Array of SLP_S0_DBG* registers > > > containing debug info > > > * @slp_s0_offset: PWRMBASE offset to read SLP_S0 residency > > > * @ltr_ignore_offset: PWRMBASE offset to read/write LTR > > > ignore bit > > > * @regmap_length: Length of memory to map from PWRMBASE > > > address to access > > > @@ -153,6 +161,8 @@ struct pmc_bit_map { > > > * PPFEAR > > > * @pm_cfg_offset: PWRMBASE offset to PM_CFG register > > > * @pm_read_disable_bit: Bit index to read PMC_READ_DISABLE > > > + * @slps0_dbg_offset: PWRMBASE offset to SLP_S0_DEBUG_REG* > > > + * @slps0_dbg_num: Number of SLP_S0_DEBUG_REG registers > > > * > > > * Each PCH has unique set of register offsets and bit indexes. > > > This structure > > > * captures them to have a common implementation. > > > @@ -161,6 +171,7 @@ struct pmc_reg_map { > > > const struct pmc_bit_map *pfear_sts; > > > const struct pmc_bit_map *mphy_sts; > > > const struct pmc_bit_map *pll_sts; > > > + const struct slps0_dbg_map *slps0_dbg_maps; > > > const u32 slp_s0_offset; > > > const u32 ltr_ignore_offset; > > > const int regmap_length; > > > @@ -168,6 +179,8 @@ struct pmc_reg_map { > > > const int ppfear_buckets; > > > const u32 pm_cfg_offset; > > > const int pm_read_disable_bit; > > > + const u32 slps0_dbg_offset; > > > + const int slps0_dbg_num; > > > }; > > > > > > /** > > > -- > > > 2.7.4 > > > > > > > -- Best Regards, Rajneesh