Hi Vivek, On 29/08/2026 09:31, Vivek Ray wrote: > Hi Casey, > > Hope you're doing well. I wanted to follow up on this patch once more. > Wanted to make sure it hadn't been lost track of. > > You kindly gave it a Reviewed-by back on June 17, and I haven't seen any > further activity on it , Patchwork still showing it as "New". I just > wanted to check whether it's been picked up into your tree, or if > there's anything still needed from my end.
My apologies, I think this slipped through and didn't get picked in my last PR. I'll pick it up shortly. Thanks and kind regards, > > Thanks again for your time and review. > > Regards, > Vivek Ray > > On Wed, Jun 17, 2026 at 8:26 PM Casey Connolly > <[email protected] <mailto:[email protected]>> wrote: > > Hi Vivek, > > On 10/06/2026 15:54, Vivek Ray wrote: > > The configure_env() function copies root compatible strings into a > > 32-byte buffer prior to parsing. If a compatible string exceeds this > > limit, strlcpy() silently truncates it, producing a malformed fdtfile > > path and a silent boot failure with no indication of the root cause. > > > > Bump buf[] to 128 bytes and dt_path[] to 256 bytes to accommodate > longer > > compatible strings. Also, add explicit truncation checks after both > > strlcpy() calls and emit a log_warning() with the offending compatible > > string to make such failures easier to diagnose. > > > > Signed-off-by: Vivek Ray <[email protected] > <mailto:[email protected]>> > > Thanks for the quick respin! > > Reviewed-by: Casey Connolly <[email protected] > <mailto:[email protected]>> > > > > > --- > > > > Changes in V2: > > - Bump buf[] from 32 to 128 bytes and dt_path[] from 64 to 256 > (Casey Connolly) > > - Fix line size check, use > not >= (Casey Connolly) > > > > --- > > arch/arm/mach-snapdragon/board.c | 18 ++++++++++++++---- > > 1 file changed, 14 insertions(+), 4 deletions(-) > > > > diff --git a/arch/arm/mach-snapdragon/board.c b/arch/arm/mach- > snapdragon/board.c > > index 829a0109ac7..d9390d9b5d3 100644 > > --- a/arch/arm/mach-snapdragon/board.c > > +++ b/arch/arm/mach-snapdragon/board.c > > @@ -390,13 +390,14 @@ static void configure_env(void) > > { > > const char *first_compat, *last_compat; > > char *tmp; > > - char buf[32] = { 0 }; > > + char buf[128] = { 0 }; > > + int len = 0; > > /* > > * Most DTB filenames follow the scheme: qcom/<soc>- > [vendor]-<board>.dtb > > * The vendor is skipped when it's a Qualcomm reference > board, or the > > * db845c. > > */ > > - char dt_path[64] = { 0 }; > > + char dt_path[256] = { 0 }; > > int compat_count, ret; > > ofnode root; > > > > @@ -417,7 +418,11 @@ static void configure_env(void) > > return; > > } > > > > - strlcpy(buf, first_compat, sizeof(buf) - 1); > > + /* A safety check to avoid silent failure due to name > truncation */ > > + len = strlcpy(buf, first_compat, sizeof(buf) - 1); > > + if (len > sizeof(buf) - 1) > > + log_warning("compatible '%s' got truncated, fdtfile > name too long\n", > > + first_compat); > > tmp = buf; > > > > /* The Qualcomm reference boards (RBx, HDK, etc) */ > > @@ -468,7 +473,12 @@ static void configure_env(void) > > > > /* Copy the last compat (e.g. "qcom,sdm845") into buf */ > > memset(buf, 0, sizeof(buf)); > > - strlcpy(buf, last_compat, sizeof(buf) - 1); > > + > > + /* check for name truncation */ > > + len = strlcpy(buf, last_compat, sizeof(buf) - 1); > > + if (len > sizeof(buf) - 1) > > + log_warning("compatible '%s' got truncated, > fdtfile name too long\n", > > + last_compat); > > tmp = buf; > > > > /* strsep() is destructive, it replaces the comma > with a \0 */ > > -- > // Casey (she/her) > -- // Casey (she/her)
