https://bugzilla.kernel.org/show_bug.cgi?id=215740 --- Comment #35 from Alan Stern (stern@xxxxxxxxxxxxxxxxxxx) --- So as most of the people subscribing to this bug have seen from the email discussion, it looks like the bug ultimately lies in the alignment of kmalloc() buffers and it probably will be fixed with a kernel patch. Equally important, however, is the question of why the people experiencing this problem turned on CONFIG_DMA_API_DEBUG in the first place. That setting is meant only for kernel testing and it drastically reduces performance. Redhat certainly should not enable it in any of their distribution kernels (and in fact it is not set in my copy of Fedora 38). Andrew, Ferry, any comments? -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
