wifi sdio -110 错误分析

硬件平台

mdm9x07 + qca9377

现象

连wifi 测试跑大数据 出现wifi 连接不上， 重新卸载加载wifi驱动 报-110 错误, 只有重启设备wifi 才恢复正常

日志

打开MMC kernel 调试日志

CONFIG_MMC_DEBUG=y

关闭mmc1

& sdhc_2 {

        *****

        status = "disabled"

}

<3>[   74.494041] mmc0: Timeout waiting for hardware interrupt.
<6>[   74.494075] sdhci: =========== REGISTER DUMP (mmc0)===========
<6>[   74.494094] sdhci: Sys addr: 0x00000000 | Version:  0x00002e02      BASE
<6>[   74.494111] sdhci: Blk size: 0x00000100 | Blk cnt:  0x00000000
<6>[   74.494129] sdhci: Argument: 0x9e549c06 | Trn mode: 0x00000023
<6>[   74.494146] sdhci: Present:  0x03f80101 | Host ctl: 0x00000016
<6>[   74.494163] sdhci: Power:    0x0000000d | Blk gap:  0x00000000
<6>[   74.494180] sdhci: Wake-up:  0x00000000 | Clock:    0x00000007      实际始终400K
<6>[   74.494197] sdhci: Timeout:  0x0000000d | Int stat: 0x00000000      超时时间0d
<6>[   74.494214] sdhci: Int enab: 0x03bf810b | Sig enab: 0x03bf010b
<6>[   74.494231] sdhci: AC12 err: 0x00000000 | Slot int: 0x00000000
<6>[   74.494248] sdhci: Caps:     0x262dc8b2 | Caps_1:   0x00008007
<6>[   74.494265] sdhci: Cmd:      0x0000353a | Max curr: 0x00000000      CMD 53
<6>[   74.494282] sdhci: Resp 1:   0x00000000 | Resp 0:   0x00001000
<6>[   74.494300] sdhci: Resp 3:   0x00000000 | Resp 2:   0x00000000
<6>[   74.494316] sdhci: Host ctl2: 0x0000400b
<6>[   74.494332] sdhci: ADMA Err: 0x00000003 | ADMA Ptr: 0x8f848008
<6>[   74.494349] ----------- VENDOR REGISTER DUMP -----------
<6>[   74.494367] Data cnt: 0x00000000 | Fifo cnt: 0x00000200 | Int sts: 0x000c0000
<6>[   74.494387] DLL cfg:  0x074d6400 | DLL sts:  0x000001a4 | SDCC ver: 0x1000002e
<6>[   74.494408] Vndr func: 0x00018a1e | Vndr adma err : addr0: 0x008c8400 addr1: 0x00000000
<6>[   74.494429] Vndr func2: 0xfa8018a8
<6>[   74.494475]  Test bus[0 to 3]: 0x0000c846 0x000020ce 0x00007018 0x01c002ff
<6>[   74.494495]  Test bus[4 to 7]: 0x00473fd8 0x0005c038 0x40000000 0xffa3ffff
<6>[   74.494514]  Test bus[8 to 11]: 0x4ffc3684 0x40800002 0x213be089 0x00001cc0
<6>[   74.494534]  Test bus[12 to 15]: 0x075f0508 0x843f11a0 0x0d001c48 0x00000a88
<6>[   74.494554]  Test bus[16 to 19]: 0x00022000 0x000f808c 0x00000000 0x00001c91
<6>[   74.494573]  Test bus[20 to 23]: 0x0241ff61 0x00054ea1 0x00000000 0x00000000
<6>[   74.494592]  Test bus[24 to 27]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494611]  Test bus[28 to 31]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494630]  Test bus[32 to 35]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494649]  Test bus[36 to 39]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494668]  Test bus[40 to 43]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494687]  Test bus[44 to 47]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494706]  Test bus[48 to 51]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494725]  Test bus[52 to 55]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494745]  Test bus[56 to 59]: 0x00000000 0x00000000 0x00000000 0x00000000
<6>[   74.494766] mmc0: clk: 200000000 clk-gated: 0 claimer: AR6K Async pwr: 12 host->irq = 8192
<6>[   74.494787] mmc0: rpmstatus[pltfm](runtime-suspend:usage_count:disable_depth)(0:1:0)
<6>[   74.494807] sdhci: ===========================================
<6>[   74.494827] mmc0: cf848000: DMA 0x8c928000, LEN 0x0600, Attr=0x23
<6>[   74.494845] mmc0: bytes to transfer: 1536 transferred: 0
<7>[   74.494966] mmc0: req done (CMD53): -84: 00000000 00000000 00000000 00000000
<7>[   74.494976] mmc0:     0 bytes transferred: -110
<4>[   74.495067] AR6000: SDIO bus operation failed! MMC stack returned : -84 
<4>[   74.495088] __HIFReadWrite, addr:0X012A4E, len:00001536, Write, Async

 此处 为等待发送cmd 53 后，没有接收到中断，超时，延长等待中断时间

vim host/sdhci.c

void sdhci_send_command(struct sdhci_host *host, struct mmc_command *cmd)
{
        ...

        //add by svk@20230822
        pr_debug("%s: salvikie cmd->opcode=%d timeout=%lu\n", mmc_hostname(host->mmc), cmd->opcode, (timeout - jiffies) / HZ);
        if(cmd->opcode == 53)
        {
                timeout += 1000 * HZ;
                pr_debug("%s: salvikie cmd->opcode=%d set timeout=%lu\n", mmc_hostname(host->mmc), cmd->opcode, (timeout - jiffies) / HZ);
        }
        //end by svk
        mod_timer(&host->timer, timeout);
        ....
}

收到一个异常中断，然后卡在那里

<7>[  116.624325] mmc0: starting CMD53 arg 9410d004 flags 000001b5
<7>[  116.624344] mmc0:     blksz 4 blocks 1 flags 00000100 tsac 1000 ms nsac 0
<7>[  116.624369] mmc0: salvikie cmd->busy_timeout=0
<7>[  116.624378] mmc0: salvikie cmd->opcode=53 timeout=10
<7>[  116.624386] mmc0: salvikie cmd->opcode=53 set timeout=1010
<7>[  116.624399] mmc0: salvikie data->timeout_clks=50000 host->clock=200000000
<7>[  116.624409] mmc0: salvikie current_timeout=163 target_timeout=1000000
<7>[  116.624418] mmc0: func[sdhci_set_timeout] salvikie count=0xd
<7>[  116.624447] sdhci [sdhci_irq()]: *** mmc0 got interrupt: 0x00028000

降低clock

void mmc_set_ios(struct mmc_host *host)
{
        struct mmc_ios *ios = &host->ios;

        pr_debug("%s: clock %uHz busmode %u powermode %u cs %u Vdd %u "
                "width %u timing %u\n",
                 mmc_hostname(host), ios->clock, ios->bus_mode,
                 ios->power_mode, ios->chip_select, ios->vdd,
                 ios->bus_width, ios->timing);

        //add by svk
        if(ios->clock == 200000000)
        {
                //ios->clock = 400000;
                //ios->clock = 20000000;
                ios->clock = 100000000;
                pr_debug("%s: salvikie change clock to %d/1001 K\n", mmc_hostname(host), ios->clock);
        }
        //end by svk
         
        ....       
}

降到20M 稳定很多

后续，debugfs 设置最大时钟，卸载WiFi驱动 完全断电（mmc）