From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gnu.wildebeest.org (gnu.wildebeest.org [45.83.234.184]) by sourceware.org (Postfix) with ESMTPS id F118C3858C41 for ; Tue, 10 Oct 2023 15:23:10 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org F118C3858C41 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=klomp.org Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=klomp.org Received: from r6.localdomain (82-217-174-174.cable.dynamic.v4.ziggo.nl [82.217.174.174]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by gnu.wildebeest.org (Postfix) with ESMTPSA id EFC2E300B302; Tue, 10 Oct 2023 17:23:09 +0200 (CEST) Received: by r6.localdomain (Postfix, from userid 1000) id B367034031B; Tue, 10 Oct 2023 17:23:09 +0200 (CEST) Message-ID: Subject: Re: [PATCH 04/16] libelf: Fix deadlock in elf_cntl From: Mark Wielaard To: elfutils-devel@sourceware.org Cc: hsm2@rice.edu Date: Tue, 10 Oct 2023 17:23:09 +0200 In-Reply-To: <20231010134300.53830-4-mark@klomp.org> References: <301fac87e83ebbbd677750579ae9a3429b461bdf.camel@klomp.org> <20231010134300.53830-1-mark@klomp.org> <20231010134300.53830-4-mark@klomp.org> Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable User-Agent: Evolution 3.48.4 (3.48.4-1.fc38) MIME-Version: 1.0 X-Spam-Status: No, score=-3033.6 required=5.0 tests=BAYES_00,GIT_PATCH_0,JMQ_SPF_NEUTRAL,KAM_DMARC_STATUS,RCVD_IN_BARRACUDACENTRAL,SPF_HELO_NONE,SPF_PASS,TXREP autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org List-Id: Hi Heather, On Tue, 2023-10-10 at 15:42 +0200, Mark Wielaard wrote: > From: Heather McIntyre >=20 > * libelf/elf_cntl.c (elf_cntl): Move rwlock_wrlock, rwlock_unlock, > inside case switch statements. >=20 > Signed-off-by: Heather S. McIntyre > Signed-off-by: Mark Wielaard > --- > libelf/elf_cntl.c | 11 +++++++---- > 1 file changed, 7 insertions(+), 4 deletions(-) >=20 > diff --git a/libelf/elf_cntl.c b/libelf/elf_cntl.c > index 04aa9132..64087c7d 100644 > --- a/libelf/elf_cntl.c > +++ b/libelf/elf_cntl.c > @@ -48,13 +48,16 @@ elf_cntl (Elf *elf, Elf_Cmd cmd) > return -1; > } > =20 > - rwlock_wrlock (elf->lock); > + > =20 > switch (cmd) > { > case ELF_C_FDREAD: > + rwlock_rdlock (elf->lock); > + int addr_isnull =3D elf->map_address =3D=3D NULL; > + rwlock_unlock(elf->lock); > /* If not all of the file is in the memory read it now. */ > - if (elf->map_address =3D=3D NULL && __libelf_readall (elf) =3D=3D = NULL) > + if (addr_isnull && __libelf_readall (elf) =3D=3D NULL) > { > /* We were not able to read everything. */ > result =3D -1; Can't we just rely on if (__libelf_readall (elf) =3D=3D NULL)? __libelf_readall already does locking and will return non-NULL if elf- >map_address is already set. So it looks like the extra check (and locking) to check addr_isnull is redundant and just make the code more complex. > @@ -64,7 +67,9 @@ elf_cntl (Elf *elf, Elf_Cmd cmd) > =20 > case ELF_C_FDDONE: > /* Mark the file descriptor as not usable. */ > + rwlock_wrlock (elf->lock); > elf->fildes =3D -1; > + rwlock_unlock (elf->lock); > break; > =20 > default: This looks correct. All other accesses to elf->fildes seem to be done under the elf->lock too. > @@ -73,7 +78,5 @@ elf_cntl (Elf *elf, Elf_Cmd cmd) > break; > } > =20 > - rwlock_unlock (elf->lock); > - > return result; > }