Re: [RFC PATCH] mmc: cqhci: keep CQE enabled if halt times out
Adrian Hunter <[email protected]>
| Newsgroups | org.kernel.vger.linux-mmc,org.kernel.vger.linux-kernel |
|---|---|
| Organization | Intel Finland Oy, Registered Address: c/o Alberga Business Park, 6 krs, Bertel Jungin Aukio 5, 02600 Espoo, Business Identity Code: 0357606 - 4, Domiciled in Helsinki |
| Message-ID | <[email protected]> |
On 23/06/2026 17:04, Pengpeng Hou wrote: > cqhci_off() asks the controller to halt before post_disable(), clearing > mmc->cqe_on and freeing CQE resources. If the halt wait times out, the > function still continues through that success path even though the > hardware may still be using the command queue. CQHCI should be idle when cqhci_off() is used, so it should halt immediately. If not, it is a hardware error, but carrying on is a reasonable option because I/O errors will trigger recovery later. > > Make cqhci_off() report whether the halt completed and leave the CQE > state/resources intact when the controller remains stuck on. This keeps > the existing host-specific ->disable() ordering and is intended as an > RFC patch because the cqe_disable callback itself has no error return. Is there a real problem that you are trying to fix? > > Signed-off-by: Pengpeng Hou <[email protected]> > --- > drivers/mmc/host/cqhci-core.c | 17 +++++++++++------ > 1 file changed, 11 insertions(+), 6 deletions(-) > > diff --git a/drivers/mmc/host/cqhci-core.c b/drivers/mmc/host/cqhci-core.c > index 178277d90..84846207c 100644 > --- a/drivers/mmc/host/cqhci-core.c > +++ b/drivers/mmc/host/cqhci-core.c > @@ -372,14 +372,14 @@ static u32 cqhci_read_ctl(struct cqhci_host *cq_host) > return cqhci_readl(cq_host, CQHCI_CTL); > } > > -static void cqhci_off(struct mmc_host *mmc) > +static bool cqhci_off(struct mmc_host *mmc) > { > struct cqhci_host *cq_host = mmc->cqe_private; > u32 reg; > int err; > > if (!cq_host->enabled || !mmc->cqe_on || cq_host->recovery_halt) > - return; > + return true; > > if (cq_host->ops->disable) > cq_host->ops->disable(mmc, false); > @@ -388,15 +388,19 @@ static void cqhci_off(struct mmc_host *mmc) > > err = readx_poll_timeout(cqhci_read_ctl, cq_host, reg, > reg & CQHCI_HALT, 0, CQHCI_OFF_TIMEOUT); > - if (err < 0) > + if (err < 0) { > pr_err("%s: cqhci: CQE stuck on\n", mmc_hostname(mmc)); > - else > - pr_debug("%s: cqhci: CQE off\n", mmc_hostname(mmc)); > + return false; > + } > + > + pr_debug("%s: cqhci: CQE off\n", mmc_hostname(mmc)); > > if (cq_host->ops->post_disable) > cq_host->ops->post_disable(mmc); > > mmc->cqe_on = false; > + > + return true; > } > > static void cqhci_disable(struct mmc_host *mmc) > @@ -406,7 +410,8 @@ static void cqhci_disable(struct mmc_host *mmc) > if (!cq_host->enabled) > return; > > - cqhci_off(mmc); > + if (!cqhci_off(mmc)) > + return; > > __cqhci_disable(cq_host); >