On 7 February 2017 at 14:49, Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> wrote: > On Tue, Feb 07, 2017 at 02:38:16PM +0000, Emil Velikov wrote: >> On 7 February 2017 at 14:29, Daniel Vetter <daniel.vetter@xxxxxxxx> wrote: >> > Noticed that everyone duplicates the same logic here and we could safe >> > a few lines per driver. Yay for lots of drivers to make such tiny >> > refactors worth-while! >> > >> > v2: Forgot to git add everything :( >> > >> Hmm afaict this patch inlines drm_fb_helper_release_fbi within >> drm_fb_helper_fini yet it is missing: >> - removal of the (now unused ?) drm_fb_helper_release_fbi >> - the leaks which now occur in the error paths. > > The error cleanup is a bit unobvious. The fbi is allocated during the > create/initial_fb callback, which is after a successful > framebuffer_init. The caller must be prepared to do a framebuffer_fini > (and so release_fbi will be handled if required) on failure or success. Silly me was assuming that drm_fb_helper_init() executes the fb_probe callback, while in fact it is drm_fb_helper_initial_config() that does so. I should read the docs more carefully ;-) Thanks gents, Emil _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
