Message ID | 1359078224-3327-1-git-send-email-airlied@gmail.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Thu, Jan 24, 2013 at 8:43 PM, Dave Airlie <airlied@gmail.com> wrote: > Okay so Alan's patch handled the case where there was no registered fbcon, > however the other path entered in set_con2fb_map pit. > > In there we called fbcon_takeover, but we also took the console lock in a couple > of places. So push the console lock out to the callers of set_con2fb_map, > > this means fbmem and switcheroo needed to take the lock around the fb notifier > entry points that lead to this. > > This should fix the efifb regression seen by Maarten. > > Signed-off-by: Dave Airlie <airlied@redhat.com> Linus might not be taking this for 3.8 because of lack of testing, but that doesn't mean the problems don't exist in 3.8 (and older). Maybe throw a CC for stable on the patch? That way it works its way back when it does get merged. josh -- To unsubscribe from this list: send the line "unsubscribe linux-fbdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Jan 25, 2013 at 2:43 AM, Dave Airlie <airlied@gmail.com> wrote: > Okay so Alan's patch handled the case where there was no registered fbcon, > however the other path entered in set_con2fb_map pit. > > In there we called fbcon_takeover, but we also took the console lock in a couple > of places. So push the console lock out to the callers of set_con2fb_map, > > this means fbmem and switcheroo needed to take the lock around the fb notifier > entry points that lead to this. > > This should fix the efifb regression seen by Maarten. > > Signed-off-by: Dave Airlie <airlied@redhat.com> > --- > drivers/gpu/vga/vga_switcheroo.c | 3 +++ > drivers/video/console/fbcon.c | 11 ++++++++--- > drivers/video/fbmem.c | 2 ++ > 3 files changed, 13 insertions(+), 3 deletions(-) > [cut] > ret = vgasr_priv.handler->switchto(new_client->id); > diff --git a/drivers/video/console/fbcon.c b/drivers/video/console/fbcon.c > index 2aef9ca..2e2d959 100644 > --- a/drivers/video/console/fbcon.c > +++ b/drivers/video/console/fbcon.c > @@ -842,6 +842,8 @@ static void con2fb_init_display(struct vc_data *vc, struct fb_info *info, > * > * Maps a virtual console @unit to a frame buffer device > * @newidx. > + * > + * This should be called with the console lock held. > */ > static int set_con2fb_map(int unit, int newidx, int user) > { What about throwing a WARN_CONSOLE_UNLOCKED(); in here to make sure this new rule is obeyed? -Daniel
diff --git a/drivers/gpu/vga/vga_switcheroo.c b/drivers/gpu/vga/vga_switcheroo.c index fa60add..cf787e1 100644 --- a/drivers/gpu/vga/vga_switcheroo.c +++ b/drivers/gpu/vga/vga_switcheroo.c @@ -25,6 +25,7 @@ #include <linux/fb.h> #include <linux/pci.h> +#include <linux/console.h> #include <linux/vga_switcheroo.h> #include <linux/vgaarb.h> @@ -337,8 +338,10 @@ static int vga_switchto_stage2(struct vga_switcheroo_client *new_client) if (new_client->fb_info) { struct fb_event event; + console_lock(); event.info = new_client->fb_info; fb_notifier_call_chain(FB_EVENT_REMAP_ALL_CONSOLE, &event); + console_unlock(); } ret = vgasr_priv.handler->switchto(new_client->id); diff --git a/drivers/video/console/fbcon.c b/drivers/video/console/fbcon.c index 2aef9ca..2e2d959 100644 --- a/drivers/video/console/fbcon.c +++ b/drivers/video/console/fbcon.c @@ -842,6 +842,8 @@ static void con2fb_init_display(struct vc_data *vc, struct fb_info *info, * * Maps a virtual console @unit to a frame buffer device * @newidx. + * + * This should be called with the console lock held. */ static int set_con2fb_map(int unit, int newidx, int user) { @@ -859,7 +861,7 @@ static int set_con2fb_map(int unit, int newidx, int user) if (!search_for_mapped_con() || !con_is_bound(&fb_con)) { info_idx = newidx; - return fbcon_takeover(0); + return do_fbcon_takeover(0); } if (oldidx != -1) @@ -867,7 +869,6 @@ static int set_con2fb_map(int unit, int newidx, int user) found = search_fb_in_map(newidx); - console_lock(); con2fb_map[unit] = newidx; if (!err && !found) err = con2fb_acquire_newinfo(vc, info, unit, oldidx); @@ -894,7 +895,6 @@ static int set_con2fb_map(int unit, int newidx, int user) if (!search_fb_in_map(info_idx)) info_idx = newidx; - console_unlock(); return err; } @@ -3019,6 +3019,7 @@ static inline int fbcon_unbind(void) } #endif /* CONFIG_VT_HW_CONSOLE_BINDING */ +/* called with console_lock held */ static int fbcon_fb_unbind(int idx) { int i, new_idx = -1, ret = 0; @@ -3045,6 +3046,7 @@ static int fbcon_fb_unbind(int idx) return ret; } +/* called with console_lock held */ static int fbcon_fb_unregistered(struct fb_info *info) { int i, idx; @@ -3082,6 +3084,7 @@ static int fbcon_fb_unregistered(struct fb_info *info) return 0; } +/* called with console_lock held */ static void fbcon_remap_all(int idx) { int i; @@ -3126,6 +3129,7 @@ static inline void fbcon_select_primary(struct fb_info *info) } #endif /* CONFIG_FRAMEBUFFER_DETECT_PRIMARY */ +/* called with console_lock held */ static int fbcon_fb_registered(struct fb_info *info) { int ret = 0, i, idx; @@ -3278,6 +3282,7 @@ static int fbcon_event_notify(struct notifier_block *self, ret = fbcon_fb_unregistered(info); break; case FB_EVENT_SET_CONSOLE_MAP: + /* called with console lock held */ con2fb = event->data; ret = set_con2fb_map(con2fb->console - 1, con2fb->framebuffer, 1); diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c index 070b9a1..dc61c12 100644 --- a/drivers/video/fbmem.c +++ b/drivers/video/fbmem.c @@ -1177,8 +1177,10 @@ static long do_fb_ioctl(struct fb_info *info, unsigned int cmd, event.data = &con2fb; if (!lock_fb_info(info)) return -ENODEV; + console_lock(); event.info = info; ret = fb_notifier_call_chain(FB_EVENT_SET_CONSOLE_MAP, &event); + console_unlock(); unlock_fb_info(info); break; case FBIOBLANK:
Okay so Alan's patch handled the case where there was no registered fbcon, however the other path entered in set_con2fb_map pit. In there we called fbcon_takeover, but we also took the console lock in a couple of places. So push the console lock out to the callers of set_con2fb_map, this means fbmem and switcheroo needed to take the lock around the fb notifier entry points that lead to this. This should fix the efifb regression seen by Maarten. Signed-off-by: Dave Airlie <airlied@redhat.com> --- drivers/gpu/vga/vga_switcheroo.c | 3 +++ drivers/video/console/fbcon.c | 11 ++++++++--- drivers/video/fbmem.c | 2 ++ 3 files changed, 13 insertions(+), 3 deletions(-)