[PATCH] posix: execvpe: skip $PATH components that are too long

[Pádraig Brady]

Adhemerval Zanella Netto wrote:
On 08/11/25 16:51, Pádraig Brady wrote:
 > > * posix/execvpe.c (__execvpe_common): Rather than error out
 > > with ENAMETOOLONG, just ignore and try the next path.
 > > Note we know the FILE length is <= NAME_MAX, so the ENAMETOOLONG
 > > almost certainly pertains to the current $PATH entry.
 > > ---
 > >  posix/execvpe.c | 3 +++
 > >  1 file changed, 3 insertions(+)
 > >
 > > diff --git a/posix/execvpe.c b/posix/execvpe.c
 > > index c139dfe8fd..0fb09b9631 100644
 > > --- a/posix/execvpe.c
 > > +++ b/posix/execvpe.c
 > > @@ -156,6 +156,9 @@ __execvpe_common (const char *file, char *const argv[], char *const envp[],
 > >  	  /* Those errors indicate the file is missing or not executable
 > >  	     by us, in which case we want to just try the next path
 > >  	     directory.  */
 > > +	  case ENAMETOOLONG:
 > > +	  /* We've already verified that the FILE length is < NAME_MAX,
 > > +	     so this implies a path component is too long, so skip it.  */
 > >  	  case ENODEV:
 > >  	  case ETIMEDOUT:
 > >  	  /* Some strange filesystems like AFS return even
 >
 > I think if Paul's latest comment holds that this is indeed an issue, we should
 > add a bug report and a regression test for this issue.

Cool. I've created https://sourceware.org/PR33626
and attached an updated patch to include a test case
(also attached here).

cheers,
Padraig
-------------- next part --------------
A non-text attachment was scrubbed...
Name: glibc-execvpe-enametoolong.patch
Type: text/x-patch
Size: 3777 bytes
Desc: not available
URL: <https://sourceware.org/pipermail/libc-alpha/attachments/20251113/525f49dd/attachment-0001.bin>