XDMA: 调用dma_to_device往fpga内存写操作时，报错timeout，并且SG engine status: BUSY

[wangzhl-code]

log如下：
[ 916.624808] xdma:xdma_xfer_submit: xfer 0x00000000da55982a,2097152, s 0x1 timed out, ep 0x200000.
[ 916.633692] xdma:engine_reg_dump: 0-H2C0-MM: ioread32(0x00000000f050ec13) = 0x1fc00006 (id).
[ 916.642121] xdma:engine_reg_dump: 0-H2C0-MM: ioread32(0x00000000bc1b0568) = 0x00000001 (status).
[ 916.650897] xdma:engine_reg_dump: 0-H2C0-MM: ioread32(0x00000000f364767c) = 0x00f83e1f (control)
[ 916.659673] xdma:engine_reg_dump: 0-H2C0-MM: ioread32(0x00000000c61bbec9) = 0xf9b40000 (first_desc_lo)
[ 916.668969] xdma:engine_reg_dump: 0-H2C0-MM: ioread32(0x00000000aaed0a9e) = 0x00000000 (first_desc_hi)
[ 916.678265] xdma:engine_reg_dump: 0-H2C0-MM: ioread32(0x000000003ce8a09c) = 0x0000003f (first_desc_adjacent).
[ 916.688169] xdma:engine_reg_dump: 0-H2C0-MM: ioread32(0x000000001a1282d9) = 0x00000000 (completed_desc_count).
[ 916.698160] xdma:engine_reg_dump: 0-H2C0-MM: ioread32(0x0000000084e157a3) = 0x00f83e1e (interrupt_enable_mask)
[ 916.708152] xdma:engine_status_dump: SG engine 0-H2C0-MM status: 0x00000001: BUSY
[ 916.715623] xdma:transfer_abort: abort transfer 0x00000000da55982a, desc 512, engine desc queued 0.
有哪位大佬知道这是什么原因吗？

[chenlimingcn]

一样的问题，请问解决了么

[wangzhl-code]

暂时还没有，不过这是偶现的，只要出现了不管怎么reset，哪怕rescan pcie或者重新加载驱动都不行。但是只要是正常了，怎么造都可以。。。

…

---原始邮件--- 发件人: "Rimon ***@***.***> 发送时间: 2025年7月8日(周二) 下午5:34 收件人: ***@***.***>; 抄送: ***@***.******@***.***>; 主题: Re: [Xilinx/dma_ip_drivers] XDMA: 调用dma_to_device往fpga内存写操作时，报错timeout，并且SG engine status: BUSY (Issue #345) chenlimingcn left a comment (Xilinx/dma_ip_drivers#345) 一样的问题，请问解决了么 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>