--------- Forwarded message ---------- From: Conor O <falling174fps@xxxxxxxxx> Date: Tue, Mar 11, 2014 at 4:36 PM Subject: Addition to https://www.kernel.org/doc/Documentation/fb/deferred_io.txt To: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> Dear Geert, Having run into this problem myself (http://stackoverflow.com/questions/22285151/kernel-panic-using-deferred-io-on-kmalloced-buffer) I'd like to suggest an addendum to the deferred_io.txt document to note the difference in treatment between a framebuffer allocated with vmalloc and one allocated with, say, kmalloc. I've pasted in the diff below. Thanks, Conor O'Rourke. $ diff deferred_io.txt deferred_io_changes.txt 75a76,89 > > 5. Bear in mind the following: > > - deferred_io changes info->fp_ops->fb_mmap to its own function that > sets up fault handlers etc. In the fault handler it checks bounds > against info->fix.smem_len > > - when it gets the page written, it checks the address in info->screen_base > to see if that address is one from vmalloc. If so it uses that address > as the one to get the page number. If not, the assumption is that the > address of interest is the physical framebuffer address stored in > info->fix.smem_start. You can get an physical address from, say, a > kmalloced pointer by using virt_to_phys(). > -- To unsubscribe from this list: send the line "unsubscribe linux-fbdev" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
