From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:35437) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VIsVu-00040d-Jd for qemu-devel@nongnu.org; Sun, 08 Sep 2013 23:51:55 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VIsVl-0000lt-Fg for qemu-devel@nongnu.org; Sun, 08 Sep 2013 23:51:46 -0400 Received: from e28smtp08.in.ibm.com ([122.248.162.8]:60588) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VIsVk-0000lX-TJ for qemu-devel@nongnu.org; Sun, 08 Sep 2013 23:51:37 -0400 Received: from /spool/local by e28smtp08.in.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Mon, 9 Sep 2013 09:09:39 +0530 Received: from d28relay04.in.ibm.com (d28relay04.in.ibm.com [9.184.220.61]) by d28dlp01.in.ibm.com (Postfix) with ESMTP id 8A8BCE004F for ; Mon, 9 Sep 2013 09:22:15 +0530 (IST) Received: from d28av01.in.ibm.com (d28av01.in.ibm.com [9.184.220.63]) by d28relay04.in.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id r893pNXJ43057366 for ; Mon, 9 Sep 2013 09:21:24 +0530 Received: from d28av01.in.ibm.com (localhost [127.0.0.1]) by d28av01.in.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id r893pPnb005767 for ; Mon, 9 Sep 2013 09:21:25 +0530 Message-ID: <522D4564.3060100@linux.vnet.ibm.com> Date: Mon, 09 Sep 2013 11:49:56 +0800 From: Lei Li MIME-Version: 1.0 References: <1378495308-24560-1-git-send-email-otubo@linux.vnet.ibm.com> <1378495308-24560-2-git-send-email-otubo@linux.vnet.ibm.com> In-Reply-To: <1378495308-24560-2-git-send-email-otubo@linux.vnet.ibm.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCHv2 1/3] seccomp: adding blacklist support List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eduardo Otubo Cc: pmoore@redhat.com, coreyb@linux.vnet.ibm.com, qemu-devel@nongnu.org On 09/07/2013 03:21 AM, Eduardo Otubo wrote: > Adding a system call blacklist right before the vcpus starts. This filter is > composed by the system calls that can't be executed after the guests are up. > This list should be refined as the whitelist is, with as much testing as we can > do using virt-test. > > Signed-off-by: Eduardo Otubo > --- > include/sysemu/seccomp.h | 5 ++++- > qemu-seccomp.c | 57 ++++++++++++++++++++++++++++++++++++++---------- > vl.c | 16 +++++++++++++- > 3 files changed, 64 insertions(+), 14 deletions(-) > > diff --git a/include/sysemu/seccomp.h b/include/sysemu/seccomp.h > index 1189fa2..551ad12 100644 > --- a/include/sysemu/seccomp.h > +++ b/include/sysemu/seccomp.h > @@ -15,8 +15,11 @@ > #ifndef QEMU_SECCOMP_H > #define QEMU_SECCOMP_H > > +#define WHITELIST 0 > +#define BLACKLIST 1 > + > #include > #include "qemu/osdep.h" > > -int seccomp_start(void); > +int seccomp_start(int state); > #endif > diff --git a/qemu-seccomp.c b/qemu-seccomp.c > index 37d38f8..5e85eb5 100644 > --- a/qemu-seccomp.c > +++ b/qemu-seccomp.c > @@ -21,7 +21,7 @@ struct QemuSeccompSyscall { > uint8_t priority; > }; > > -static const struct QemuSeccompSyscall seccomp_whitelist[] = { > +static const struct QemuSeccompSyscall whitelist[] = { > { SCMP_SYS(timer_settime), 255 }, > { SCMP_SYS(timer_gettime), 254 }, > { SCMP_SYS(futex), 253 }, > @@ -221,32 +221,65 @@ static const struct QemuSeccompSyscall seccomp_whitelist[] = { > { SCMP_SYS(arch_prctl), 240 } > }; > > -int seccomp_start(void) > +static const struct QemuSeccompSyscall blacklist[] = { > + { SCMP_SYS(execve), 255 } > +}; > + > +static int process_list(scmp_filter_ctx *ctx, > + const struct QemuSeccompSyscall *list, > + unsigned int list_size, uint32_t action) > { > int rc = 0; > unsigned int i = 0; > - scmp_filter_ctx ctx; > > - ctx = seccomp_init(SCMP_ACT_KILL); > - if (ctx == NULL) { > - goto seccomp_return; > - } > + for (i = 0; i < list_size; i++) { > + rc = seccomp_rule_add(ctx, action, list[i].num, 0); > + if (rc < 0) { > + goto seccomp_return; > + } > > - for (i = 0; i < ARRAY_SIZE(seccomp_whitelist); i++) { > - rc = seccomp_rule_add(ctx, SCMP_ACT_ALLOW, seccomp_whitelist[i].num, 0); > + rc = seccomp_syscall_priority(ctx, list[i].num, > + list[i].priority); > if (rc < 0) { > goto seccomp_return; > } > - rc = seccomp_syscall_priority(ctx, seccomp_whitelist[i].num, > - seccomp_whitelist[i].priority); > + } > + > +seccomp_return: > + return rc; > +} > + > +int seccomp_start(int list_type) > +{ > + int rc = 0; > + scmp_filter_ctx ctx; > + > + switch (list_type) { > + case WHITELIST: > + ctx = seccomp_init(SCMP_ACT_KILL); > + if (ctx == NULL) { > + goto seccomp_return; > + } > + rc = process_list(ctx, whitelist, ARRAY_SIZE(whitelist), SCMP_ACT_ALLOW); > if (rc < 0) { > goto seccomp_return; > } > + break; > + case BLACKLIST: > + ctx = seccomp_init(SCMP_ACT_ALLOW); > + if (ctx == NULL) { > + goto seccomp_return; > + } > + rc = process_list(ctx, blacklist, ARRAY_SIZE(blacklist), SCMP_ACT_KILL); > + break; > + default: > + rc = -1; > + goto seccomp_return; > } > > rc = seccomp_load(ctx); > > - seccomp_return: > +seccomp_return: > seccomp_release(ctx); > return rc; > } > diff --git a/vl.c b/vl.c > index b4b119a..02f7486 100644 > --- a/vl.c > +++ b/vl.c > @@ -179,6 +179,7 @@ int main(int argc, char **argv) > #define MAX_VIRTIO_CONSOLES 1 > #define MAX_SCLP_CONSOLES 1 > > +static bool enable_blacklist = false; > static const char *data_dir[16]; > static int data_dir_idx; > const char *bios_name = NULL; > @@ -1033,11 +1034,13 @@ static int parse_sandbox(QemuOpts *opts, void *opaque) > /* FIXME: change this to true for 1.3 */ > if (qemu_opt_get_bool(opts, "enable", false)) { > #ifdef CONFIG_SECCOMP > - if (seccomp_start() < 0) { > + if (seccomp_start(WHITELIST) < 0) { > qerror_report(ERROR_CLASS_GENERIC_ERROR, > "failed to install seccomp syscall filter in the kernel"); > return -1; > } > + > + enable_blacklist = true; > #else > qerror_report(ERROR_CLASS_GENERIC_ERROR, > "sandboxing request but seccomp is not compiled into this build"); > @@ -1765,12 +1768,23 @@ void vm_state_notify(int running, RunState state) > } > } > > +static void install_seccomp_blacklist(void) > +{ > + if (enable_blacklist) { > + if (seccomp_start(BLACKLIST) < 0) { > + qerror_report(ERROR_CLASS_GENERIC_ERROR, > + "failed to install seccomp syscall second level filter in the kernel"); > + } > + } > +} > + > void vm_start(void) > { > if (!runstate_is_running()) { > cpu_enable_ticks(); > runstate_set(RUN_STATE_RUNNING); > vm_state_notify(1, RUN_STATE_RUNNING); > + install_seccomp_blacklist(); Hi Eduardo, Looks good to me, especially the implementation of blacklist than the previous one that the second whitelist is the same as the first one except that two system calls. Just one question: If the seccomp_start(WHITELIST) failed, it will lead to the QEMU exit(1). But seems that if this seccomp_start(BLACKLIST) failed, it won't, just report an error that the second level filter in kernel failed to install. Is this the expected behaver? > resume_all_vcpus(); > monitor_protocol_event(QEVENT_RESUME, NULL); > } -- Lei