Re: mmc: dw_mmc: warning with CONFIG_DMA_API_DEBUG

From: Shawn Lin
Date: Mon Jun 20 2016 - 21:52:35 EST


On 2016/6/20 16:34, Seung-Woo Kim wrote:
Hi folks,

During booting test on my Exynos5422 based Odroid-XU3, kernel compiled
with CONFIG_DMA_API_DEBUG reported following warning:

------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at lib/dma-debug.c:1096 check_unmap+0x7bc/0xb38
dwmmc_exynos 12200000.mmc: DMA-API: device driver tries to free DMA memory it has not allocated [device address=0x000000006d9d2200]

Thanks for this report and fix.

DTO(the same as IDMAC-RI/TI) interrupts may or may not come together with DATA_ERR. If DATA_ERR occur without geting DTO, we should issue
CMD12 manually to generate DTO. It's a ugly deisgn for dwmmc but from
the vendor's ask.

So you should never think we complete the xfer without
checking DATA_ERR. This way you got the warning.

So could you try this one:

--- a/drivers/mmc/host/dw_mmc.c
+++ b/drivers/mmc/host/dw_mmc.c
@@ -2474,7 +2474,8 @@ static irqreturn_t dw_mci_interrupt(int irq, void *dev_id)
mci_writel(host, IDSTS64, SDMMC_IDMAC_INT_TI |

SDMMC_IDMAC_INT_RI);
mci_writel(host, IDSTS64, SDMMC_IDMAC_INT_NI);
- host->dma_ops->complete((void *)host);
+ if (!test_bit(EVENT_DATA_ERROR, &host->pending_events))
+ host->dma_ops->complete((void *)host);
}
} else {
pending = mci_readl(host, IDSTS);
@@ -2482,7 +2483,8 @@ static irqreturn_t dw_mci_interrupt(int irq, void *dev_id)
mci_writel(host, IDSTS, SDMMC_IDMAC_INT_TI |

SDMMC_IDMAC_INT_RI);
mci_writel(host, IDSTS, SDMMC_IDMAC_INT_NI);
- host->dma_ops->complete((void *)host);
+ if (!test_bit(EVENT_DATA_ERROR, &host->pending_events))
+ host->dma_ops->complete((void *)host);
}
}


[size=128 bytes]
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.7.0-rc4 #26
Hardware name: SAMSUNG EXYNOS (Flattened Device Tree)
[<c0112b4c>] (unwind_backtrace) from [<c010d888>] (show_stack+0x20/0x24)
[<c010d888>] (show_stack) from [<c03fab0c>] (dump_stack+0x80/0x94)
[<c03fab0c>] (dump_stack) from [<c0123548>] (__warn+0xf8/0x110)
[<c0123548>] (__warn) from [<c01235a8>] (warn_slowpath_fmt+0x48/0x50)
[<c01235a8>] (warn_slowpath_fmt) from [<c042ac90>] (check_unmap+0x7bc/0xb38)
[<c042ac90>] (check_unmap) from [<c042b25c>] (debug_dma_unmap_sg+0x118/0x148)
[<c042b25c>] (debug_dma_unmap_sg) from [<c077512c>] (dw_mci_dma_cleanup+0x7c/0xb8)
[<c077512c>] (dw_mci_dma_cleanup) from [<c0773f24>] (dw_mci_stop_dma+0x40/0x50)
[<c0773f24>] (dw_mci_stop_dma) from [<c0777d04>] (dw_mci_tasklet_func+0x130/0x3b4)
[<c0777d04>] (dw_mci_tasklet_func) from [<c0129760>] (tasklet_action+0xb4/0x150)
[<c0129760>] (tasklet_action) from [<c0101674>] (__do_softirq+0xe4/0x3cc)
[<c0101674>] (__do_softirq) from [<c0129030>] (irq_exit+0xd0/0x10c)
[<c0129030>] (irq_exit) from [<c01778a0>] (__handle_domain_irq+0x90/0xfc)
[<c01778a0>] (__handle_domain_irq) from [<c0101548>] (gic_handle_irq+0x64/0xa8)
[<c0101548>] (gic_handle_irq) from [<c010e3d4>] (__irq_svc+0x54/0x90)
Exception stack(0xc1101ef8 to 0xc1101f40)
1ee0: 00000001 00000000
1f00: 00000000 c011b600 c1100000 c110753c 00000000 c11c3984 c11074d4 c1107548
1f20: 00000000 c1101f54 c1101f58 c1101f48 c010a1fc c010a200 60000013 ffffffff
[<c010e3d4>] (__irq_svc) from [<c010a200>] (arch_cpu_idle+0x48/0x4c)
[<c010a200>] (arch_cpu_idle) from [<c01669d8>] (default_idle_call+0x30/0x3c)
[<c01669d8>] (default_idle_call) from [<c0166d3c>] (cpu_startup_entry+0x358/0x3b4)
[<c0166d3c>] (cpu_startup_entry) from [<c0aa6ab8>] (rest_init+0x94/0x98)
[<c0aa6ab8>] (rest_init) from [<c1000d58>] (start_kernel+0x3a4/0x3b0)
[<c1000d58>] (start_kernel) from [<4000807c>] (0x4000807c)
---[ end trace 256f83eed365daf0 ]---

The warning occurs because after complete callback function,
dw_mci_dmac_complete_dma() is called, then dw_mci_stop_dma() is called
again. So it causes dma_unmap_sg() is called twice for same sg. It
occurs during clock setting at booting time.

Simply, clearing host->using_dma flag on dw_mci_dmac_complete_dma() and
dw_mci_stop_dma() like following fixes the issue, but I am not sure
this approach is proper.
---
diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
index 2cc6123..a71c94b 100644
--- a/drivers/mmc/host/dw_mmc.c
+++ b/drivers/mmc/host/dw_mmc.c
@@ -388,6 +388,7 @@ static void dw_mci_stop_dma(struct dw_mci *host)
if (host->using_dma) {
host->dma_ops->stop(host);
host->dma_ops->cleanup(host);
+ host->using_dma = 0;
}

/* Data transfer was stopped by the interrupt handler */
@@ -455,6 +456,7 @@ static void dw_mci_dmac_complete_dma(void *arg)
DMA_FROM_DEVICE);

host->dma_ops->cleanup(host);
+ host->using_dma = 0;

/*
* If the card was removed, data will be NULL. No point in trying to
@@ -943,8 +945,6 @@ static int dw_mci_submit_data_dma(struct dw_mci *host, struct mmc_data *data)
int sg_len;
u32 temp;

- host->using_dma = 0;
-
/* If we don't have a channel, we can't do DMA */
if (!host->use_dma)
return -ENODEV;
---

Best Regards,
- Seung-Woo Kim

--
To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html



--
Best Regards
Shawn Lin