Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 12EF8C004D4 for ; Sat, 21 Jan 2023 13:37:24 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229714AbjAUNhX (ORCPT ); Sat, 21 Jan 2023 08:37:23 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:50044 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229699AbjAUNhW (ORCPT ); Sat, 21 Jan 2023 08:37:22 -0500 Received: from cosmopolitan.snafu.de (cosmopolitan.snafu.de [IPv6:2001:1560:3:255::151]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 01DA249430 for ; Sat, 21 Jan 2023 05:37:18 -0800 (PST) X-Trace: 4c7c73656261737469616e2e6b6c6f736b6140736e6166752e64657c38342e3233 2e3235342e3230357c31704a4533532d3030304c58302d49797c31363734333038 323330 Received: from cosmopolitan.snafu.de ([10.151.10.19] helo=localhost) by cosmopolitan.snafu.de with esmtpa (Exim 4.94.2) id 1pJE3S-000LX0-Iy; Sat, 21 Jan 2023 14:37:12 +0100 MIME-Version: 1.0 Date: Sat, 21 Jan 2023 14:37:10 +0100 From: sebastian.kloska@snafu.de To: Adrian Hunter Cc: linux-mmc@vger.kernel.org Subject: Re: Regression 5.14.21 vs 5.15.1: mmc blocked: mmc0: cqhci: timeout for tag # In-Reply-To: References: <05add7ba-f2f8-7cbe-3449-4a3128c096b5@intel.com> <2ae56408f156f4f728cd1e94863c0795@snafu.de> <08bd329f23b8fac7a2249034804ac3c7@snafu.de> User-Agent: Roundcube Webmail/1.4.13 Message-ID: <2faae5dfb4999fba2da80ae7d18af899@snafu.de> X-Sender: sebastian.kloska@snafu.de Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-VISP-ShouldScan: 1 X-VISP-Virus-Check: clean X-VISP-Spam-Score: 0.7 (/) X-VISP-Spam-Report: This message has been scanned on "cosmopolitan.snafu.de" to identify if it is considered spam or not. Contact the support hotline for details. Content analysis details: (0.7 points) pts rule name description ---- ---------------------- -------------------------------------------------- 0.5 JMQ_SPF_NEUTRAL ASKDNS: SPF set to ?all [snafu.de TXT:v=spf1 mx ip4:84.23.254.128/25] [ip6:2001:1560:3:255::/64 ip6:2001:1560:3:254::/64 ip4:213.73.113.34/32 ip6:2001:1560:a000:2:50:ff:fe00:27ed/128 include:_spf.snafu.de ?all] -1.0 ALL_TRUSTED Passed through trusted hosts only via SMTP 0.1 TW_GF BODY: Odd Letter Triples with GF 0.1 TW_XF BODY: Odd Letter Triples with XF 0.2 KAM_DMARC_NONE DKIM has Failed or SPF has failed on the message and the domain has no DMARC policy 0.0 KAM_DMARC_STATUS Test Rule for DKIM or SPF Failure with Strict Alignment 0.8 KAM_ASCII_DIVIDERS Email that uses ascii formatting dividers and possible spam tricks X-VISP-Spam-Max-Score: +++++ X-SA-Exim-Connect-IP: 84.23.254.205 X-SA-Exim-Mail-From: sebastian.kloska@snafu.de X-SA-Exim-Scanned: No (on cosmopolitan.snafu.de); SAEximRunCond expanded to false Precedence: bulk List-ID: X-Mailing-List: linux-mmc@vger.kernel.org Am 20.01.2023 13:08, schrieb Adrian Hunter: > On 20/01/23 13:14, sebastian.kloska@snafu.de wrote: >> So no mail with attachment on linux-mmc@vger.kernel.org >> but I linked the dmesgs on >> https://bugzilla.kernel.org/show_bug.cgi?id=216255 >> > > I did receive the logs but it is good to attach them to the bug anyway. > >>   >> Am 19.01.2023 11:28, schrieb Adrian Hunter: >>> On 19/01/23 11:46, sebastian.kloska@snafu.de >>>  wrote: >>>> Am 19.01.2023 10:15, schrieb Adrian Hunter: >>>>> On 18/01/23 23:24, sebastian.kloska@snafu.de >>>>>  wrote: >>>>>> Hi, >>>>>> >>>>>> Hardware: Acer Swift-1 SF114-34-P6U1 >>>>>> SD Host controller: Intel Corporation Device 4dc4 (rev 01) 8086:4dc4 >>>>>> eMMC: Kingston A29128 >>>>>> >>>>>> * A couple of seconds after boot access to the eMMC is completely >>>>>>   blocked. Can be triggered by extensive I/O (e.g, with badblocks >>>>>> -w) >>>>>> >>>>>> * Tested with stable branch linux-5.14.21 vs. linux-5.15.1 >>>>>>   Latest highest version including the bug is a fedora 6.1.6 >>>>> >>>>> There is very little difference in drivers/mmc between 5.14.21 >>>>> and 5.15.1. Are you sure the same issue is not present in >>>>> 5.14.21? >>>> >>>> Pretty sure. Using 5.14.21 day in day out 5.15.1 fails on me after >>>> ~10min >>>> The "slowed down by pr_debug" version sometimes lives for >2h >>>> (without stress >>>> via badblocks etc) >>>> >> So all of the following is now done on the latest stable, Which >> currently >> is 6.2.0-rc4+ > > That is a release candidate (rc). The latest stable kernel > is 6.1.7. > I know -just thought I should pull master/HEAD from the latest stable >>>>> >>>>> You could try disabling runtime PM to see if that helps: >>>>> >>>>>     echo on > /sys/class/mmc_host/mmc0/device/power/control >>>>> >> Didn't help >> >>>> Will try >>>>> Also it should be possible to disable cqhci by adding to the kernel >>>>> command line: >>>>> >>>>>     sdhci.debug_quirks=0x40a0000 >>>>> >> >> Weird behavior: Seems to make it worse. System often freezes >> during early boot-up. >> >> HOWEVER: If I slow down the driver via dynamic debugging it seems to >> be stable >> Failed to crash it over a couple of hours. You can say it fixes the >> issue if you >> are willing to accept that 80% of one CPU core is busy with logging ;-) > > That is interesting. Logging can have more side effects than you might > expect. > For example the logs might be written back to storage by syslog, which > causes > more logging, and more writing etc etc. That in turn is bumping up the > CPU > load. > > Increasing the CPU load will decrease the chance of the CPU entering a > low-power > state (C-state). > > And there was an issue with C-state transitions affecting Intel SDHCI > controllers, > so we need to check that possibility. You could try kernel command > line option: > > intel_idle.max_cstate=1 > This -- Sir -- works just wonderfully !! Don't see any changes in /sys/devices/system/cpu/cpu*/cpuidle/state*/disable,but may be that's expected. Kernel runs just fine. works for stable/6.1.7 and fedora/6.1.6 So for me personally this workaround is fine. Guess it will hurt my power consumption/battery live though but that's acceptable for me. Don't hesitate to contact if I should test patches for you. The source is checked out and ccache is filled. > You can see the C-states and whether they are disabled: > > grep -H . /sys/devices/system/cpu/cpu*/cpuidle/state*/disable > > "disable" is writable by root, so you can disable them that way also. > > In the meantime I will take a look at the logs. > >> >>>> Will try >>>>>> >>>>>> * Spiked the code with a lot of pr_debug messages. As a side >>>>>> effect the driver >>>>>>   seems to be much more stable, so to me it appears like a timing >>>>>> issue where >>>>>>   the driver fails to wait for a specific state change which seems >>>>>> to appear >>>>>>   less often when  slowed down by logging >>>>> >>>>> Ideally it would be good to see a full log with dynamic >>>>> debugging enabled for the latest kernel. >>>> >>>> Latest torvalds/linux or stable/linux ? >>> >>> Stable would be fine. >>> >>>> >>>>> >>>>>     To enable mmc debug via kernel command line: >>>>> >>>>>         dyndbg="file drivers/mmc/core/* +p;file drivers/mmc/host/* >>>>> +p" >>>>> >> >> This for whatever reason did not work for me on the kernel cmdline >> but I could trigger it via echo "..." >/dynamic_debug/control >> so the attached dmesgs are missing early logs. >> >>>>>     Kernel must be configured: >>>>> >>>>>         CONFIG_DYNAMIC_DEBUG=y >>>>> >>>>>> >>>>>> * Would love to have a decent reference documentation to the SD >>>>>> Controller >>>>>>   especially a detailed description of the behavior of the >>>>>> register map in >>>>>>  mmc/host/cqhci.h All intel documents I could find seem to describe >>>>>>  similar but not quite identical register maps >>>>> >>>>> The JEDEC eMMC specification describes the CQHCI registers. >>>>> It looks like you need to register to get access to them >>>>> at www.jedec.org.  >>>>>  AFAICT older specs, which would be just >>>>> fine, are free to download after registering. >>>>> >> >> Uhh -- but the price for registration is hefty. THX anyway >> >>>> THX >>>>> You can get SD Host Controller Simplified Specification >>>>> from www.sdcard.org . >>>>> >> >> I attach to dmesg logs >> * One with dynamic debugging and without sdhci.debug_quirks=0x40a0000 >> * The other without dynamic debugging (as said I failed to crash the >> system) and sdhci.debug_quirks=0x40a0000 >> >> Hope I filtered properly for the relevant stuff >> The second one contains what seems to me a ton of kernel Oops but >> that might be due to the preliminary status of 6.2.0-rc4 >> >> THX >> >>>> THX >>>>>> >>>>>> Regards >>>>>> >>>>>> Sebastian >>>>>> >>>>>> * dmesg >>>>>> >>>>>> [  347.583082] mmc0: cqhci: timeout for tag 3, qcnt >>>>> >>>>> The timeout is pretty long, so it seems like the controller >>>>> has gotten stuck.  It won't reset the command or data >>>>> circuits either, which is impossible if it is behaving >>>>> correctly. >>>>> >>>>>> [  347.583086] mmc0: cqhci: ============ CQHCI REGISTER DUMP >>>>>> =========== >>>>>> [  347.583114] mmc0: cqhci: Caps:      0x000030c0 | Version:  >>>>>> 0x00000510 >>>>>> [  347.583117] mmc0: cqhci: Config:    0x00001101 | Control:  >>>>>> 0x00000000 >>>>>> [  347.583120] mmc0: cqhci: Int stat:  0x00000000 | Int enab: >>>>>> 0x00000016 >>>>>> [  347.583123] mmc0: cqhci: Int sig:   0x00000016 | Int Coal: >>>>>> 0x00000000 >>>>>> [  347.583157] mmc0: cqhci: TDL base:  0x7c2b5000 | TDL up32: >>>>>> 0x00000001 >>>>>> [  347.583160] mmc0: cqhci: Doorbell:  0x0000000f | TCN:      >>>>>> 0x00000000 >>>>>> [  347.583163] mmc0: cqhci: Dev queue: 0x00000008 | Dev Pend: >>>>>> 0x00000008 >>>>>> [  347.583166] mmc0: cqhci: Task clr:  0x00000000 | SSC1:     >>>>>> 0x00010008 >>>>>> [  347.583169] mmc0: cqhci: SSC2:      0x00000001 | DCMD rsp: >>>>>> 0x00000800 >>>>>> [  347.583172] mmc0: cqhci: RED mask:  0xfdf9a080 | TERRI:    >>>>>> 0x00000000 >>>>>> [  347.583175] mmc0: cqhci: Resp idx:  0x0000002f | Resp arg: >>>>>> 0x00000900 >>>>>> [  347.583176] mmc0: sdhci: ============ SDHCI REGISTER DUMP >>>>>> =========== >>>>>> [  347.583182] mmc0: sdhci: Sys addr:  0x00000080 | Version:  >>>>>> 0x00001002 >>>>>> [  347.583185] mmc0: sdhci: Blk size:  0x00000200 | Blk cnt:  >>>>>> 0x00000070 >>>>>> [  347.583188] mmc0: sdhci: Argument:  0x00030000 | Trn mode: >>>>>> 0x00000023 >>>>>> [  347.583191] mmc0: sdhci: Present:   0x1fff0106 | Host ctl: >>>>>> 0x0000003c >>>>>> [  347.583194] mmc0: sdhci: Power:     0x0000000b | Blk gap:  >>>>>> 0x00000080 >>>>>> [  347.583197] mmc0: sdhci: Wake-up:   0x00000000 | Clock:    >>>>>> 0x00000007 >>>>>> [  347.583200] mmc0: sdhci: Timeout:   0x0000000e | Int stat: >>>>>> 0x00000000 >>>>>> [  347.583203] mmc0: sdhci: Int enab:  0x02ff4000 | Sig enab: >>>>>> 0x02ff4000 >>>>>> [  347.583206] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: >>>>>> 0x00000000 >>>>>> [  347.583209] mmc0: sdhci: Caps:      0x546ec881 | Caps_1:   >>>>>> 0x80000807 >>>>>> [  347.583212] mmc0: sdhci: Cmd:       0x00002f3a | Max curr: >>>>>> 0x00000000 >>>>>> [  347.583214] mmc0: sdhci: Resp[0]:   0x00000900 | Resp[1]:  >>>>>> 0x00000000 >>>>>> [  347.583217] mmc0: sdhci: Resp[2]:   0x00000000 | Resp[3]:  >>>>>> 0x00000000 >>>>>> [  347.583219] mmc0: sdhci: Host ctl2: 0x0000000d >>>>>> [  347.583223] mmc0: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: >>>>>> 0x000000011eef1218 >>>>>> [  347.583224] mmc0: sdhci: >>>>>> ============================================ >>>>>> [  347.583229] mmc0: running CQE recovery >>>>>> [  347.690108] mmc0: Reset 0x2 never completed. >>>>>> [  347.690109] mmc0: sdhci: ============ SDHCI REGISTER DUMP >>>>>> =========== >>>>>> [  347.690115] mmc0: sdhci: Sys addr:  0x00000080 | Version:  >>>>>> 0x00001002 >>>>>> [  347.690118] mmc0: sdhci: Blk size:  0x00000200 | Blk cnt:  >>>>>> 0x00000070 >>>>>> [  347.690121] mmc0: sdhci: Argument:  0x00030000 | Trn mode: >>>>>> 0x00000023 >>>>>> [  347.690124] mmc0: sdhci: Present:   0x1fff0106 | Host ctl: >>>>>> 0x0000003c >>>>>> [  347.690127] mmc0: sdhci: Power:     0x0000000b | Blk gap:  >>>>>> 0x00000080 >>>>>> [  347.690130] mmc0: sdhci: Wake-up:   0x00000000 | Clock:    >>>>>> 0x00000007 >>>>>> [  347.690133] mmc0: sdhci: Timeout:   0x0000000e | Int stat: >>>>>> 0x00000000 >>>>>> [  347.690136] mmc0: sdhci: Int enab:  0x00ff0003 | Sig enab: >>>>>> 0x00ff0003 >>>>>> [  347.690139] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: >>>>>> 0x00000000 >>>>>> [  347.690141] mmc0: sdhci: Caps:      0x546ec881 | Caps_1:   >>>>>> 0x80000807 >>>>>> [  347.690144] mmc0: sdhci: Cmd:       0x00002f3a | Max curr: >>>>>> 0x00000000 >>>>>> [  347.690147] mmc0: sdhci: Resp[0]:   0x00000900 | Resp[1]:  >>>>>> 0x00000000 >>>>>> [  347.690150] mmc0: sdhci: Resp[2]:   0x00000000 | Resp[3]:  >>>>>> 0x00000000 >>>>>> [  347.690152] mmc0: sdhci: Host ctl2: 0x0000000d >>>>>> [  347.690156] mmc0: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: >>>>>> 0x000000011eef1218 >>>>>> [  347.690157] mmc0: sdhci: >>>>>> ============================================ >>>>>> [  347.790190] mmc0: Reset 0x4 never completed. >>>>>> [  347.790191] mmc0: sdhci: ============ SDHCI REGISTER DUMP >>>>>> =========== >>>>>> [  347.790193] mmc0: sdhci: Sys addr:  0x00000080 | Version:  >>>>>> 0x00001002 >>>>>> [  347.790196] mmc0: sdhci: Blk size:  0x00000200 | Blk cnt:  >>>>>> 0x00000070 >>>>>> [  347.790199] mmc0: sdhci: Argument:  0x00030000 | Trn mode: >>>>>> 0x00000023 >>>>>> [  347.790202] mmc0: sdhci: Present:   0x1fff0106 | Host ctl: >>>>>> 0x0000003c >>>>>> [  347.790205] mmc0: sdhci: Power:     0x0000000b | Blk gap:  >>>>>> 0x00000080 >>>>>> [  347.790208] mmc0: sdhci: Wake-up:   0x00000000 | Clock:    >>>>>> 0x00000007 >>>>>> [  347.790211] mmc0: sdhci: Timeout:   0x0000000e | Int stat: >>>>>> 0x00000000 >>>>>> [  347.790213] mmc0: sdhci: Int enab:  0x00ff0003 | Sig enab: >>>>>> 0x00ff0003 >>>>>> [  347.790216] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: >>>>>> 0x00000000 >>>>>> [  347.790219] mmc0: sdhci: Caps:      0x546ec881 | Caps_1:   >>>>>> 0x80000807 >>>>>> [  347.790222] mmc0: sdhci: Cmd:       0x00002f3a | Max curr: >>>>>> 0x00000000 >>>>>> [  347.790225] mmc0: sdhci: Resp[0]:   0x00000900 | Resp[1]:  >>>>>> 0x00000000 >>>>>> [  347.790228] mmc0: sdhci: Resp[2]:   0x00000000 | Resp[3]:  >>>>>> 0x00000000 >>>>>> [  347.790230] mmc0: sdhci: Host ctl2: 0x0000000d >>>>>> [  347.790234] mmc0: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: >>>>>> 0x000000011eef1218 >>>>>> [  347.790234] mmc0: sdhci: >>>>>> ============================================ >>>>>> [  347.803056] mmc0: Controller never released inhibit bit(s). >>>>>> [  347.803058] mmc0: sdhci: ============ SDHCI REGISTER DUMP >>>>>> =========== >>>>>> [  347.803077] mmc0: sdhci: Sys addr:  0x00000080 | Version:  >>>>>> 0x00001002 >>>>>> [  347.803080] mmc0: sdhci: Blk size:  0x00000200 | Blk cnt:  >>>>>> 0x00000070 >>>>>> [  347.803083] mmc0: sdhci: Argument:  0x00030000 | Trn mode: >>>>>> 0x00000023 >>>>>> [  347.803086] mmc0: sdhci: Present:   0x1fff0106 | Host ctl: >>>>>> 0x0000003c >>>>>> [  347.803089] mmc0: sdhci: Power:     0x0000000b | Blk gap:  >>>>>> 0x00000080 >>>>>> [  347.803092] mmc0: sdhci: Wake-up:   0x00000000 | Clock:    >>>>>> 0x00000007 >>>>>> [  347.803095] mmc0: sdhci: Timeout:   0x0000000e | Int stat: >>>>>> 0x00000000 >>>>>> [  347.803098] mmc0: sdhci: Int enab:  0x00ff0003 | Sig enab: >>>>>> 0x00ff0003 >>>>>> [  347.803101] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: >>>>>> 0x00000000 >>>>>> [  347.803103] mmc0: sdhci: Caps:      0x546ec881 | Caps_1:   >>>>>> 0x80000807 >>>>>> [  347.803106] mmc0: sdhci: Cmd:       0x00002f3a | Max curr: >>>>>> 0x00000000 >>>>>> [  347.803109] mmc0: sdhci: Resp[0]:   0x00000900 | Resp[1]:  >>>>>> 0x00000000 >>>>>> [  347.803112] mmc0: sdhci: Resp[2]:   0x00000000 | Resp[3]:  >>>>>> 0x00000000 >>>>>> [  347.803114] mmc0: sdhci: Host ctl2: 0x0000000d >>>>>> [  347.803118] mmc0: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: >>>>>> 0x000000011eef1218 >>>>>> [  347.803119] mmc0: sdhci: >>>>>> ============================================ >>>>>>  -- >>>>>> >> -- >> --