On 21/01/2020 11.44, Cédric Le Goater wrote: > On 1/21/20 11:02 AM, Thomas Huth wrote: >> On 21/01/2020 10.28, mreza...@redhat.com wrote: >>> From: Miroslav Rezanina <mreza...@redhat.com> >>> >>> Compiler reports uninitialized warning for cmd_flags variable. >>> >>> Adding NULL initialization to prevent this warning. >>> >>> Signed-off-by: Miroslav Rezanina <mreza...@redhat.com> >>> --- >>> hw/i2c/aspeed_i2c.c | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/hw/i2c/aspeed_i2c.c b/hw/i2c/aspeed_i2c.c >>> index 2da04a4..445182a 100644 >>> --- a/hw/i2c/aspeed_i2c.c >>> +++ b/hw/i2c/aspeed_i2c.c >>> @@ -400,7 +400,7 @@ static bool aspeed_i2c_check_sram(AspeedI2CBus *bus) >>> >>> static void aspeed_i2c_bus_cmd_dump(AspeedI2CBus *bus) >>> { >>> - g_autofree char *cmd_flags; >>> + g_autofree char *cmd_flags = NULL; >>> uint32_t count; >>> >>> if (bus->cmd & (I2CD_RX_BUFF_ENABLE | I2CD_RX_BUFF_ENABLE)) { >> >> Reviewed-by: Thomas Huth <th...@redhat.com> >> >> ... maybe someone with enough Perl-foo (i.e. not me ;-)) should add a >> check to our check_patch.pl script so that it complains when new code is >> introduced that uses g_autofree without initializing the variable... > > weird. The cmd_flags variable is assigned just after and used > in a trace.
I don't know, but my guess is that you could compile with tracing disabled - in that case gcc might maybe optimize the assignment away, too... ? Thomas