What I noticed is when dma_map_page() function called in do_async_xor() it causing **src_list getting corrupted. Look like d-cache flush is causing some issue. Thanks, Marri -----Original Message----- From: linux-raid-owner@xxxxxxxxxxxxxxx [mailto:linux-raid-owner@xxxxxxxxxxxxxxx] On Behalf Of Tirumala Reddy Marri Sent: Friday, June 26, 2009 11:35 AM To: linux-raid@xxxxxxxxxxxxxxx Cc: Dan Williams Subject: Strange issue with async_xor All, I am using not open sourced(will open source it in future)ADMA driver for PowerPC procerror. I tried to upgrade to newer kernel to get all latest MD and async changes. I see some arguments are messed up between do_async_xor() and ppc4xx_adma_prep_dma_mq_xor(). From the code I see same number of arguments. But when I step through debugger src_cnt & len arguments got corrupted. Is anything changed with async_xor() or dmaengine.c ? I Looked at Dan williams driver and everything looks ok. Any ideas ? Regards, Marri (gdb) bt #0 ppc460ex_adma_prep_dma_mq_xor (chan=0xdd61c4a4, dst=0x5af7, src=0x0, src_cnt=492167168, len=3714768144, flags=3) at drivers/dma/ppc460ex-adma.c:3113 #1 0xc01196e4 in async_xor (dest=<value optimized out>, src_list=0xdd6addbc, offset=<value optimized out>, src_cnt=3, len=4096, flags=ASYNC_TX_XOR_ZERO_DST, depend_tx=0x0, cb_fn=0xc01e42a8 <ops_complete_compute>, cb_param=0xd8986000) at crypto/async_tx/async_xor.c:93 #2 0xc01e4bac in ops_run_compute5 (sh=<value optimized out>) at drivers/md/raid5.c:641 #3 0xc01e6240 in raid_run_ops (sh=0xd8986000, ops_request=2) at drivers/md/raid5.c:1147 #4 0xc01e6734 in handle_stripe5 (sh=0xd8986000) at drivers/md/raid5.c:2945 #5 0xc01e786c in handle_stripe (sh=0xdd61c4a4, tmp_page=0x5af7) at drivers/md/raid5.c:3233 #6 0xc01e8a44 in raid5d (mddev=0xdd74e800) at drivers/md/raid5.c:4082 #7 0xc01f4908 in md_thread (arg=<value optimized out>) at drivers/md/md.c:5295 #8 0xc00386c8 in kthread (_create=<value optimized out>) at kernel/kthread.c:83 #9 0xc000da8c in kernel_thread () Backtrace stopped: previous frame inner to this frame (corrupt stack?) (gdb) p *src_list No symbol "src_list" in current context. (gdb) f 1 #1 0xc01196e4 in async_xor (dest=<value optimized out>, src_list=0xdd6addbc, offset=<value optimized out>, src_cnt=3, len=4096, flags=ASYNC_TX_XOR_ZERO_DST, depend_tx=0x0, cb_fn=0xc01e42a8 <ops_complete_compute>, cb_param=0xd8986000) at crypto/async_tx/async_xor.c:93 93 tx = dma->device_prep_dma_xor(chan, dma_dest, &dma_src[src_off], (gdb) p *src_list $13 = (struct page *) 0xe (gdb) -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
