* [PATCH 0/1] mmc: Reduce tuning log output @ 2022-11-11 8:42 Wenchao Chen 2022-11-11 8:42 ` [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning Wenchao Chen 0 siblings, 1 reply; 9+ messages in thread From: Wenchao Chen @ 2022-11-11 8:42 UTC (permalink / raw) To: ulf.hansson, adrian.hunter; +Cc: linux-mmc, linux-kernel, megoo.tang, lzx.stg From: Wenchao Chen <wenchao.chen@unisoc.com> During the HS200 Tuning process, a lot of "mmc0: Got data interrupt 0x00200000 even though no data operation was in progress." logs were printed. From the cmd in the log, it is CMD21. [ 4.547365][ C0] mmc0: Got data interrupt 0x00200000 even though no data operation was in progress. [ 4.564088][ C0] mmc0: sdhci: ============ SDHCI REGISTER DUMP =========== [ 4.571213][ C0] mmc0: sdhci: Sys addr: 0x00000001 | Version: 0x00000004 [ 4.578339][ C0] mmc0: sdhci: Blk size: 0x00000080 | Blk cnt: 0x00000000 [ 4.585466][ C0] mmc0: sdhci: Argument: 0x00000000 | Trn mode: 0x00000013 [ 4.592595][ C0] mmc0: sdhci: Present: 0x01f000f0 | Host ctl: 0x00000030 [ 4.599723][ C0] mmc0: sdhci: Power: 0x00000000 | Blk gap: 0x00000000 [ 4.606849][ C0] mmc0: sdhci: Wake-up: 0x00000000 | Clock: 0x00000007 [ 4.613978][ C0] mmc0: sdhci: Timeout: 0x0000000c | Int stat: 0x00000000 [ 4.621106][ C0] mmc0: sdhci: Int enab: 0x037f000b | Sig enab: 0x037f000b [ 4.628233][ C0] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000 [ 4.635360][ C0] mmc0: sdhci: Caps: 0x1e6d0080 | Caps_1: 0x08000007 [ 4.642489][ C0] mmc0: sdhci: Cmd: 0x0000153a | Max curr: 0x000003fc [ 4.649616][ C0] mmc0: sdhci: Resp[0]: 0x00000900 | Resp[1]: 0x00000900 [ 4.656743][ C0] mmc0: sdhci: Resp[2]: 0x00000000 | Resp[3]: 0x00000000 [ 4.663881][ C0] mmc0: sdhci: Host ctl2: 0x00003005 [ 4.669002][ C0] mmc0: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0x00000000f0200200 [ 4.811901][ C0] mmc0: sdhci: ============================================ ....... [ 4.819456][ C0] mmc0: Got data interrupt 0x00200000 even though no data operation was in progress. [ 4.835111][ C0] mmc0: sdhci: ============ SDHCI REGISTER DUMP =========== [ 4.835116][ C0] mmc0: sdhci: Sys addr: 0x00000001 | Version: 0x00000004 [ 4.870574][ C0] mmc0: sdhci: Blk size: 0x00000080 | Blk cnt: 0x00000000 [ 4.870580][ C0] mmc0: sdhci: Argument: 0x00000000 | Trn mode: 0x00000013 [ 4.920122][ C0] mmc0: sdhci: Present: 0x01f000f0 | Host ctl: 0x00000030 [ 4.927249][ C0] mmc0: sdhci: Power: 0x00000000 | Blk gap: 0x00000000 [ 4.934377][ C0] mmc0: sdhci: Wake-up: 0x00000000 | Clock: 0x00000007 [ 4.941506][ C0] mmc0: sdhci: Timeout: 0x0000000c | Int stat: 0x00000000 [ 4.948631][ C0] mmc0: sdhci: Int enab: 0x037f000b | Sig enab: 0x037f000b [ 4.955759][ C0] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000 [ 4.962887][ C0] mmc0: sdhci: Caps: 0x1e6d0080 | Caps_1: 0x08000007 [ 4.970014][ C0] mmc0: sdhci: Cmd: 0x0000153a | Max curr: 0x000003fc [ 4.977143][ C0] mmc0: sdhci: Resp[0]: 0x00000900 | Resp[1]: 0x00000900 [ 4.984270][ C0] mmc0: sdhci: Resp[2]: 0x000ls00900 | Resp[3]: 0x00000000 [ 4.991398][ C0] mmc0: sdhci: Host ctl2: 0x00003005 [ 4.996526][ C0] mmc0: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0x00000000f0200200 [ 4.811901][ C0] mmc0: sdhci: ============================================ Wenchao Chen (1): mmc: sdhci: Fixed too many logs being printed during tuning drivers/mmc/host/sdhci.c | 4 ++++ 1 file changed, 4 insertions(+) -- 2.17.1 ^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning 2022-11-11 8:42 [PATCH 0/1] mmc: Reduce tuning log output Wenchao Chen @ 2022-11-11 8:42 ` Wenchao Chen 2022-11-16 15:20 ` Ulf Hansson 2022-11-16 16:00 ` Adrian Hunter 0 siblings, 2 replies; 9+ messages in thread From: Wenchao Chen @ 2022-11-11 8:42 UTC (permalink / raw) To: ulf.hansson, adrian.hunter; +Cc: linux-mmc, linux-kernel, megoo.tang, lzx.stg From: Wenchao Chen <wenchao.chen@unisoc.com> During the HS200 tuning process, too many tuning errors are printed in the log. Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com> --- drivers/mmc/host/sdhci.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index fef03de85b99..a503b54305eb 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -3401,6 +3401,10 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) if (host->pending_reset) return; + command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND)); + if (command == MMC_SEND_TUNING_BLOCK || command == MMC_SEND_TUNING_BLOCK_HS200) + return; + pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", mmc_hostname(host->mmc), (unsigned)intmask); sdhci_err_stats_inc(host, UNEXPECTED_IRQ); -- 2.17.1 ^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning 2022-11-11 8:42 ` [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning Wenchao Chen @ 2022-11-16 15:20 ` Ulf Hansson 2022-11-18 9:22 ` Wenchao Chen 2022-11-16 16:00 ` Adrian Hunter 1 sibling, 1 reply; 9+ messages in thread From: Ulf Hansson @ 2022-11-16 15:20 UTC (permalink / raw) To: Wenchao Chen; +Cc: adrian.hunter, linux-mmc, linux-kernel, megoo.tang, lzx.stg On Fri, 11 Nov 2022 at 09:42, Wenchao Chen <wenchao.chen666@gmail.com> wrote: > > From: Wenchao Chen <wenchao.chen@unisoc.com> > > During the HS200 tuning process, too many tuning errors are printed in > the log. > > Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com> > --- > drivers/mmc/host/sdhci.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > index fef03de85b99..a503b54305eb 100644 > --- a/drivers/mmc/host/sdhci.c > +++ b/drivers/mmc/host/sdhci.c > @@ -3401,6 +3401,10 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) > if (host->pending_reset) > return; > > + command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND)); > + if (command == MMC_SEND_TUNING_BLOCK || command == MMC_SEND_TUNING_BLOCK_HS200) > + return; > + Please use mmc_op_tuning() instead of open-coding this. I also need an ack from Adrian to apply this. > pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", > mmc_hostname(host->mmc), (unsigned)intmask); > sdhci_err_stats_inc(host, UNEXPECTED_IRQ); > -- > 2.17.1 > Kind regards Uffe ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning 2022-11-16 15:20 ` Ulf Hansson @ 2022-11-18 9:22 ` Wenchao Chen 0 siblings, 0 replies; 9+ messages in thread From: Wenchao Chen @ 2022-11-18 9:22 UTC (permalink / raw) To: Ulf Hansson; +Cc: adrian.hunter, linux-mmc, linux-kernel, megoo.tang, lzx.stg On Wed, Nov 16, 2022 at 11:21 PM Ulf Hansson <ulf.hansson@linaro.org> wrote: > > On Fri, 11 Nov 2022 at 09:42, Wenchao Chen <wenchao.chen666@gmail.com> wrote: > > > > From: Wenchao Chen <wenchao.chen@unisoc.com> > > > > During the HS200 tuning process, too many tuning errors are printed in > > the log. > > > > Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com> > > --- > > drivers/mmc/host/sdhci.c | 4 ++++ > > 1 file changed, 4 insertions(+) > > > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > > index fef03de85b99..a503b54305eb 100644 > > --- a/drivers/mmc/host/sdhci.c > > +++ b/drivers/mmc/host/sdhci.c > > @@ -3401,6 +3401,10 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) > > if (host->pending_reset) > > return; > > > > + command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND)); > > + if (command == MMC_SEND_TUNING_BLOCK || command == MMC_SEND_TUNING_BLOCK_HS200) > > + return; > > + > > Please use mmc_op_tuning() instead of open-coding this. I also need an > ack from Adrian to apply this. Thank you for your review, I will update the next version. > > > pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", > > mmc_hostname(host->mmc), (unsigned)intmask); > > sdhci_err_stats_inc(host, UNEXPECTED_IRQ); > > -- > > 2.17.1 > > > > Kind regards > Uffe ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning 2022-11-11 8:42 ` [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning Wenchao Chen 2022-11-16 15:20 ` Ulf Hansson @ 2022-11-16 16:00 ` Adrian Hunter 2022-11-18 9:25 ` Wenchao Chen 1 sibling, 1 reply; 9+ messages in thread From: Adrian Hunter @ 2022-11-16 16:00 UTC (permalink / raw) To: Wenchao Chen, ulf.hansson; +Cc: linux-mmc, linux-kernel, megoo.tang, lzx.stg On 11/11/22 10:42, Wenchao Chen wrote: > From: Wenchao Chen <wenchao.chen@unisoc.com> > > During the HS200 tuning process, too many tuning errors are printed in > the log. > > Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com> > --- > drivers/mmc/host/sdhci.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > index fef03de85b99..a503b54305eb 100644 > --- a/drivers/mmc/host/sdhci.c > +++ b/drivers/mmc/host/sdhci.c > @@ -3401,6 +3401,10 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) > if (host->pending_reset) > return; > > + command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND)); > + if (command == MMC_SEND_TUNING_BLOCK || command == MMC_SEND_TUNING_BLOCK_HS200) > + return; Normally we wouldn't get here even if a request got an error because then we either reset the data circuit which should stop further interrupts, or set host->pending_reset. Can you elaborate on what is going wrong? > + > pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", > mmc_hostname(host->mmc), (unsigned)intmask); > sdhci_err_stats_inc(host, UNEXPECTED_IRQ); ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning 2022-11-16 16:00 ` Adrian Hunter @ 2022-11-18 9:25 ` Wenchao Chen 2022-11-18 15:57 ` Adrian Hunter 0 siblings, 1 reply; 9+ messages in thread From: Wenchao Chen @ 2022-11-18 9:25 UTC (permalink / raw) To: Adrian Hunter; +Cc: ulf.hansson, linux-mmc, linux-kernel, megoo.tang, lzx.stg On Thu, Nov 17, 2022 at 12:00 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > On 11/11/22 10:42, Wenchao Chen wrote: > > From: Wenchao Chen <wenchao.chen@unisoc.com> > > > > During the HS200 tuning process, too many tuning errors are printed in > > the log. > > > > Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com> > > --- > > drivers/mmc/host/sdhci.c | 4 ++++ > > 1 file changed, 4 insertions(+) > > > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > > index fef03de85b99..a503b54305eb 100644 > > --- a/drivers/mmc/host/sdhci.c > > +++ b/drivers/mmc/host/sdhci.c > > @@ -3401,6 +3401,10 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) > > if (host->pending_reset) > > return; > > > > + command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND)); > > + if (command == MMC_SEND_TUNING_BLOCK || command == MMC_SEND_TUNING_BLOCK_HS200) > > + return; > > Normally we wouldn't get here even if a request got an error because > then we either reset the data circuit which should stop further > interrupts, or set host->pending_reset. > > Can you elaborate on what is going wrong? > Hi adrian 1. In the process of tuning, not all tuning values are good, some tuning values may cause errors, and these errors print too many logs. Of course, we reset the cmdline or dataline on error. 2. use host->pending_reset = true static void __sdhci_finish_mrq(struct sdhci_host *host, struct mmc_request *mrq) { ... if (sdhci_needs_reset(host, mrq)) host->pending_reset = true; ... } But intmask = 0x00200000 static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) { ... if (!host->data) { struct mmc_command *data_cmd = host->data_cmd; if (data_cmd && (data_cmd->flags & MMC_RSP_BUSY)) { if (intmask & SDHCI_INT_DATA_TIMEOUT) { //#define SDHCI_INT_DATA_TIMEOUT 0x00100000 host->data_cmd = NULL; data_cmd->error = -ETIMEDOUT; sdhci_err_stats_inc(host, CMD_TIMEOUT); __sdhci_finish_mrq(host, data_cmd->mrq);//<= return; } if (intmask & SDHCI_INT_DATA_END) { //#define SDHCI_INT_DATA_END 0x00000002 host->data_cmd = NULL; /* * Some cards handle busy-end interrupt * before the command completed, so make * sure we do things in the proper order. */ if (host->cmd == data_cmd) return; __sdhci_finish_mrq(host, data_cmd->mrq);//<= return; } } ... if (host->pending_reset) return; pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", mmc_hostname(host->mmc), (unsigned)intmask); sdhci_err_stats_inc(host, UNEXPECTED_IRQ); sdhci_dumpregs(host); return; } ... } > > + > > pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", > > mmc_hostname(host->mmc), (unsigned)intmask); > > sdhci_err_stats_inc(host, UNEXPECTED_IRQ); > ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning 2022-11-18 9:25 ` Wenchao Chen @ 2022-11-18 15:57 ` Adrian Hunter 2022-11-21 3:11 ` Wenchao Chen 0 siblings, 1 reply; 9+ messages in thread From: Adrian Hunter @ 2022-11-18 15:57 UTC (permalink / raw) To: Wenchao Chen; +Cc: ulf.hansson, linux-mmc, linux-kernel, megoo.tang, lzx.stg On 18/11/22 11:25, Wenchao Chen wrote: > On Thu, Nov 17, 2022 at 12:00 AM Adrian Hunter <adrian.hunter@intel.com> wrote: >> >> On 11/11/22 10:42, Wenchao Chen wrote: >>> From: Wenchao Chen <wenchao.chen@unisoc.com> >>> >>> During the HS200 tuning process, too many tuning errors are printed in >>> the log. >>> >>> Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com> >>> --- >>> drivers/mmc/host/sdhci.c | 4 ++++ >>> 1 file changed, 4 insertions(+) >>> >>> diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c >>> index fef03de85b99..a503b54305eb 100644 >>> --- a/drivers/mmc/host/sdhci.c >>> +++ b/drivers/mmc/host/sdhci.c >>> @@ -3401,6 +3401,10 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) >>> if (host->pending_reset) >>> return; >>> >>> + command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND)); >>> + if (command == MMC_SEND_TUNING_BLOCK || command == MMC_SEND_TUNING_BLOCK_HS200) >>> + return; >> >> Normally we wouldn't get here even if a request got an error because >> then we either reset the data circuit which should stop further >> interrupts, or set host->pending_reset. >> >> Can you elaborate on what is going wrong? >> > Hi adrian > 1. In the process of tuning, not all tuning values are good, some > tuning values may cause errors, and these errors print too many logs. > Of course, we reset the cmdline or dataline on error. > 2. use host->pending_reset = true > static void __sdhci_finish_mrq(struct sdhci_host *host, struct mmc_request *mrq) > { > ... > if (sdhci_needs_reset(host, mrq)) > host->pending_reset = true; > ... > } > > But intmask = 0x00200000 > static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) > { > ... > if (!host->data) { > struct mmc_command *data_cmd = host->data_cmd; > > if (data_cmd && (data_cmd->flags & MMC_RSP_BUSY)) { > if (intmask & SDHCI_INT_DATA_TIMEOUT) { //#define > SDHCI_INT_DATA_TIMEOUT 0x00100000 > host->data_cmd = NULL; > data_cmd->error = -ETIMEDOUT; > sdhci_err_stats_inc(host, CMD_TIMEOUT); > __sdhci_finish_mrq(host, data_cmd->mrq);//<= > return; > } > if (intmask & SDHCI_INT_DATA_END) { //#define SDHCI_INT_DATA_END 0x00000002 > > host->data_cmd = NULL; > /* > * Some cards handle busy-end interrupt > * before the command completed, so make > * sure we do things in the proper order. > */ > if (host->cmd == data_cmd) > return; > > __sdhci_finish_mrq(host, data_cmd->mrq);//<= > return; > } > } > ... > if (host->pending_reset) > return; > > pr_err("%s: Got data interrupt 0x%08x even though no data operation > was in progress.\n", > mmc_hostname(host->mmc), (unsigned)intmask); > sdhci_err_stats_inc(host, UNEXPECTED_IRQ); > sdhci_dumpregs(host); > > return; > } > ... > } Not really following that I'm sorry. Are you saying you get a data interrupt after he data circuit is reset? > >>> + >>> pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", >>> mmc_hostname(host->mmc), (unsigned)intmask); >>> sdhci_err_stats_inc(host, UNEXPECTED_IRQ); >> ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning 2022-11-18 15:57 ` Adrian Hunter @ 2022-11-21 3:11 ` Wenchao Chen 2022-11-30 6:18 ` Adrian Hunter 0 siblings, 1 reply; 9+ messages in thread From: Wenchao Chen @ 2022-11-21 3:11 UTC (permalink / raw) To: Adrian Hunter; +Cc: ulf.hansson, linux-mmc, linux-kernel, megoo.tang, lzx.stg On Fri, Nov 18, 2022 at 11:57 PM Adrian Hunter <adrian.hunter@intel.com> wrote: > > On 18/11/22 11:25, Wenchao Chen wrote: > > On Thu, Nov 17, 2022 at 12:00 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > >> > >> On 11/11/22 10:42, Wenchao Chen wrote: > >>> From: Wenchao Chen <wenchao.chen@unisoc.com> > >>> > >>> During the HS200 tuning process, too many tuning errors are printed in > >>> the log. > >>> > >>> Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com> > >>> --- > >>> drivers/mmc/host/sdhci.c | 4 ++++ > >>> 1 file changed, 4 insertions(+) > >>> > >>> diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > >>> index fef03de85b99..a503b54305eb 100644 > >>> --- a/drivers/mmc/host/sdhci.c > >>> +++ b/drivers/mmc/host/sdhci.c > >>> @@ -3401,6 +3401,10 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) > >>> if (host->pending_reset) > >>> return; > >>> > >>> + command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND)); > >>> + if (command == MMC_SEND_TUNING_BLOCK || command == MMC_SEND_TUNING_BLOCK_HS200) > >>> + return; > >> > >> Normally we wouldn't get here even if a request got an error because > >> then we either reset the data circuit which should stop further > >> interrupts, or set host->pending_reset. > >> > >> Can you elaborate on what is going wrong? > >> > > Hi adrian > > 1. In the process of tuning, not all tuning values are good, some > > tuning values may cause errors, and these errors print too many logs. > > Of course, we reset the cmdline or dataline on error. > > 2. use host->pending_reset = true > > static void __sdhci_finish_mrq(struct sdhci_host *host, struct mmc_request *mrq) > > { > > ... > > if (sdhci_needs_reset(host, mrq)) > > host->pending_reset = true; > > ... > > } > > > > But intmask = 0x00200000 > > static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) > > { > > ... > > if (!host->data) { > > struct mmc_command *data_cmd = host->data_cmd; > > > > if (data_cmd && (data_cmd->flags & MMC_RSP_BUSY)) { > > if (intmask & SDHCI_INT_DATA_TIMEOUT) { //#define > > SDHCI_INT_DATA_TIMEOUT 0x00100000 > > host->data_cmd = NULL; > > data_cmd->error = -ETIMEDOUT; > > sdhci_err_stats_inc(host, CMD_TIMEOUT); > > __sdhci_finish_mrq(host, data_cmd->mrq);//<= > > return; > > } > > if (intmask & SDHCI_INT_DATA_END) { //#define SDHCI_INT_DATA_END 0x00000002 > > > > host->data_cmd = NULL; > > /* > > * Some cards handle busy-end interrupt > > * before the command completed, so make > > * sure we do things in the proper order. > > */ > > if (host->cmd == data_cmd) > > return; > > > > __sdhci_finish_mrq(host, data_cmd->mrq);//<= > > return; > > } > > } > > ... > > if (host->pending_reset) > > return; > > > > pr_err("%s: Got data interrupt 0x%08x even though no data operation > > was in progress.\n", > > mmc_hostname(host->mmc), (unsigned)intmask); > > sdhci_err_stats_inc(host, UNEXPECTED_IRQ); > > sdhci_dumpregs(host); > > > > return; > > } > > ... > > } > > Not really following that I'm sorry. > > Are you saying you get a data interrupt after he data circuit is reset? > No. During the tuning process, sdhci_data_irq does not skip printing because the interrupt state is 0x00200000 (DATA CRC ERROR). > > > >>> + > >>> pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", > >>> mmc_hostname(host->mmc), (unsigned)intmask); > >>> sdhci_err_stats_inc(host, UNEXPECTED_IRQ); > >> > ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning 2022-11-21 3:11 ` Wenchao Chen @ 2022-11-30 6:18 ` Adrian Hunter 0 siblings, 0 replies; 9+ messages in thread From: Adrian Hunter @ 2022-11-30 6:18 UTC (permalink / raw) To: Wenchao Chen; +Cc: ulf.hansson, linux-mmc, linux-kernel, megoo.tang, lzx.stg On 21/11/22 05:11, Wenchao Chen wrote: > On Fri, Nov 18, 2022 at 11:57 PM Adrian Hunter <adrian.hunter@intel.com> wrote: >> >> On 18/11/22 11:25, Wenchao Chen wrote: >>> On Thu, Nov 17, 2022 at 12:00 AM Adrian Hunter <adrian.hunter@intel.com> wrote: >>>> >>>> On 11/11/22 10:42, Wenchao Chen wrote: >>>>> From: Wenchao Chen <wenchao.chen@unisoc.com> >>>>> >>>>> During the HS200 tuning process, too many tuning errors are printed in >>>>> the log. >>>>> >>>>> Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com> >>>>> --- >>>>> drivers/mmc/host/sdhci.c | 4 ++++ >>>>> 1 file changed, 4 insertions(+) >>>>> >>>>> diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c >>>>> index fef03de85b99..a503b54305eb 100644 >>>>> --- a/drivers/mmc/host/sdhci.c >>>>> +++ b/drivers/mmc/host/sdhci.c >>>>> @@ -3401,6 +3401,10 @@ static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) >>>>> if (host->pending_reset) >>>>> return; >>>>> >>>>> + command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND)); >>>>> + if (command == MMC_SEND_TUNING_BLOCK || command == MMC_SEND_TUNING_BLOCK_HS200) >>>>> + return; >>>> >>>> Normally we wouldn't get here even if a request got an error because >>>> then we either reset the data circuit which should stop further >>>> interrupts, or set host->pending_reset. >>>> >>>> Can you elaborate on what is going wrong? >>>> >>> Hi adrian >>> 1. In the process of tuning, not all tuning values are good, some >>> tuning values may cause errors, and these errors print too many logs. >>> Of course, we reset the cmdline or dataline on error. >>> 2. use host->pending_reset = true >>> static void __sdhci_finish_mrq(struct sdhci_host *host, struct mmc_request *mrq) >>> { >>> ... >>> if (sdhci_needs_reset(host, mrq)) >>> host->pending_reset = true; >>> ... >>> } >>> >>> But intmask = 0x00200000 >>> static void sdhci_data_irq(struct sdhci_host *host, u32 intmask) >>> { >>> ... >>> if (!host->data) { >>> struct mmc_command *data_cmd = host->data_cmd; >>> >>> if (data_cmd && (data_cmd->flags & MMC_RSP_BUSY)) { >>> if (intmask & SDHCI_INT_DATA_TIMEOUT) { //#define >>> SDHCI_INT_DATA_TIMEOUT 0x00100000 >>> host->data_cmd = NULL; >>> data_cmd->error = -ETIMEDOUT; >>> sdhci_err_stats_inc(host, CMD_TIMEOUT); >>> __sdhci_finish_mrq(host, data_cmd->mrq);//<= >>> return; >>> } >>> if (intmask & SDHCI_INT_DATA_END) { //#define SDHCI_INT_DATA_END 0x00000002 >>> >>> host->data_cmd = NULL; >>> /* >>> * Some cards handle busy-end interrupt >>> * before the command completed, so make >>> * sure we do things in the proper order. >>> */ >>> if (host->cmd == data_cmd) >>> return; >>> >>> __sdhci_finish_mrq(host, data_cmd->mrq);//<= >>> return; >>> } >>> } >>> ... >>> if (host->pending_reset) >>> return; >>> >>> pr_err("%s: Got data interrupt 0x%08x even though no data operation >>> was in progress.\n", >>> mmc_hostname(host->mmc), (unsigned)intmask); >>> sdhci_err_stats_inc(host, UNEXPECTED_IRQ); >>> sdhci_dumpregs(host); >>> >>> return; >>> } >>> ... >>> } >> >> Not really following that I'm sorry. >> >> Are you saying you get a data interrupt after he data circuit is reset? >> > > No. During the tuning process, sdhci_data_irq does not skip printing > because the interrupt state is 0x00200000 (DATA CRC ERROR). Sorry I am still confused. Which driver is this? I was assuming you were using mmc_send_tuning() because the only interrupt enabled for regular SDHCI tuning is SDHCI_INT_DATA_AVAIL (0x00000020). So are you saying you are getting status 0x00200000 even though the mask is 0x00000020 ? > >>> >>>>> + >>>>> pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n", >>>>> mmc_hostname(host->mmc), (unsigned)intmask); >>>>> sdhci_err_stats_inc(host, UNEXPECTED_IRQ); >>>> >> ^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2022-11-30 6:19 UTC | newest] Thread overview: 9+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2022-11-11 8:42 [PATCH 0/1] mmc: Reduce tuning log output Wenchao Chen 2022-11-11 8:42 ` [PATCH 1/1] mmc: sdhci: Fixed too many logs being printed during tuning Wenchao Chen 2022-11-16 15:20 ` Ulf Hansson 2022-11-18 9:22 ` Wenchao Chen 2022-11-16 16:00 ` Adrian Hunter 2022-11-18 9:25 ` Wenchao Chen 2022-11-18 15:57 ` Adrian Hunter 2022-11-21 3:11 ` Wenchao Chen 2022-11-30 6:18 ` Adrian Hunter
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).