| Message ID | 1555387409-9018-1-git-send-email-hofrat@osadl.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [V2] ARM: imx legacy: warn on failed allocation | expand |
diff --git a/arch/arm/mach-imx/mach-mx27ads.c b/arch/arm/mach-imx/mach-mx27ads.c index 6dd7f57..cba6876 100644 --- a/arch/arm/mach-imx/mach-mx27ads.c +++ b/arch/arm/mach-imx/mach-mx27ads.c @@ -247,6 +247,10 @@ static void __init mx27ads_regulator_init(void) struct gpio_chip *vchip; vchip = kzalloc(sizeof(*vchip), GFP_KERNEL); + if (!vchip) + pr_err("vchip allocation failure in %s()\n", + __func__); + vchip->owner = THIS_MODULE; vchip->label = "LCD"; vchip->base = MX27ADS_LCD_GPIO;
Even in init the allocation can fail and thus should report by pr_err() so that the cause can be easily identified. Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org> Link: https://lkml.org/lkml/2019/4/14/152 --- Problem located with an experimental coccinelle script V2: Russell King <linux@armlinux.org.uk> pointed out that the use of WARN_ON() would result in a stack trace followed by the oops due to dereferencing of the NULL pointer and so make it even less likely that users would uncover the actual cause - so drop the WARN_ON() and use a short pr_err() message that points to the oops cause directly. Note that this will trigger a checkpatch WARNING: "WARNING: Possible unnecessary 'out of memory' message" but comparing the oops with an without the one-line pr_err I would argue that it makes sense to include it (first line here is from pr_err()): <snip> [ 8061.514840] shared page allocation failure in hello_init() [ 8113.563239] BUG: unable to handle kernel NULL pointer dereference at 0000000000000000 [ 8113.563250] #PF error: [WRITE] [ 8113.563255] PGD 8000000129993067 P4D 8000000129993067 PUD 129992067 PMD 0 [ 8113.563267] Oops: 0002 [#1] SMP PTI [ 8113.563276] CPU: 2 PID: 2656 Comm: bash Tainted: G W O 5.0.0-rc3livepatchtest-next-20190123+ #4 [ 8113.563280] Hardware name: Quanta TWH/TWH, BIOS QU221 10/14/2011 [ 8113.563292] RIP: 0010:foo_store+0x3a/0x90 [hello_chardev] ... <snip> Patch was compile-tested with: imx_v4_v5_defconfig (implies CONFIG_MACH_MX27ADS=y) (with some unrelated sparse warnings about unimplemented syscalls) Patch is against 5.1-rc4 (localversion-next is 20190415) arch/arm/mach-imx/mach-mx27ads.c | 4 ++++ 1 file changed, 4 insertions(+)