From: Greg Kroah-Hartman gregkh@linuxfoundation.org
commit ed04fe8a0e87d7b5ea17d47f4ac9ec962b24814a upstream.
This reverts commit 1d84353d205a953e2381044953b7fa31c8c9702d.
Because of recent interactions with developers from @umn.edu, all commits from them have been recently re-reviewed to ensure if they were correct or not.
Upon review, this commit was found to be incorrect for the reasons below, so it must be reverted. It will be fixed up "correctly" in a later kernel change.
The original commit here, while technically correct, did not fully handle all of the reported issues that the commit stated it was fixing, so revert it until it can be "fixed" fully.
Note, ioremap() probably will never fail for old hardware like this, and if anyone actually used this hardware (a PowerMac era PCI display card), they would not be using fbdev anymore.
Cc: Kangjie Lu kjlu@umn.edu Cc: Aditya Pakki pakki001@umn.edu Cc: Finn Thain fthain@telegraphics.com.au Cc: Bartlomiej Zolnierkiewicz b.zolnierkie@samsung.com Reviewed-by: Rob Herring robh@kernel.org Fixes: 1d84353d205a ("video: imsttfb: fix potential NULL pointer dereferences") Cc: stable stable@vger.kernel.org Link: https://lore.kernel.org/r/20210503115736.2104747-67-gregkh@linuxfoundation.o... Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/video/fbdev/imsttfb.c | 5 ----- 1 file changed, 5 deletions(-)
--- a/drivers/video/fbdev/imsttfb.c +++ b/drivers/video/fbdev/imsttfb.c @@ -1512,11 +1512,6 @@ static int imsttfb_probe(struct pci_dev info->fix.smem_start = addr; info->screen_base = (__u8 *)ioremap(addr, par->ramdac == IBM ? 0x400000 : 0x800000); - if (!info->screen_base) { - release_mem_region(addr, size); - framebuffer_release(info); - return -ENOMEM; - } info->fix.mmio_start = addr + 0x800000; par->dc_regs = ioremap(addr + 0x800000, 0x1000); par->cmap_regs_phys = addr + 0x840000;