- dev - passt

[PATCH v3] feat: Add cli option '--pass-fds' for pasta mode.
by Richard Lawrence 23 Jul '26

23 Jul '26

From: Richard Lawrence <rarensu(a)tamu.edu> When pasta mode is used to launch an executable (`pasta [COMMAND]`) and that executable accepts inputs in the form of arbitrary file descriptors (such as `bwrap`), then passt should not stand in the way of the parent process handing off those file descriptors to the child process. See bug 204 for additional discussion. The `pass-fds` option accepts a comma-separated list of file descriptor numbers. `conf_pass_fds()` parses the command line argument, then `isolate_fds()` skips closing the specified fds by calling `close_range()` on the gaps between them. Additionally, the tap fd is safely relocated to the lowest unused fd number which it at least 3, to avoid accidentally overwriting an existing fd. Signed-off-by: Richard Lawrence <rlawrence(a)tamu.edu> --- conf.c | 74 ++++++++++++++++++++++++++++++++++++++++++++++++++++- conf.h | 3 +++ isolation.c | 73 ++++++++++++++++++++++++++++++++++++++++++---------- passt.1 | 5 ++++ 4 files changed, 140 insertions(+), 15 deletions(-) diff --git a/conf.c b/conf.c index 0fcba5c..729816d 100644 --- a/conf.c +++ b/conf.c @@ -732,7 +732,9 @@ pasta_opts: " Don't copy all addresses to namespace\n" " --ns-mac-addr ADDR Set MAC address on tap interface\n" " --no-splice Disable inbound socket splicing\n" - " --splice-only Only enable loopback forwarding\n"); + " --splice-only Only enable loopback forwarding\n" + " --pass-fds FDS Comma-separated list of fds to pass to\n" + " the spawned command\n"); passt_exit(status); } @@ -1183,6 +1185,71 @@ int conf_tap_fd(int argc, char **argv) return val; } +/** + * conf_pass_fds() - Read fds as supplied by --pass-fds command line option + * @argc: Argument count + * @argv: Command line options + * @fds: Array where we store the parsed fds + * @max_fds: Maximum size of the array + * + * Return: number of parsed fds, or -1 if option not specified + */ +int conf_pass_fds(int argc, char **argv, int *fds, int max_fds) +{ + const struct option opt[] = { { "pass-fds", required_argument, NULL, 33 }, + { 0 }, }; + const char *fdsarg = NULL; + int name, fds_cnt = 0; + int old_opterr; + + old_opterr = opterr; + opterr = 0; + optind = 0; + do { + name = getopt_long(argc, argv, "-:", opt, NULL); + if (name == 33) + fdsarg = optarg; + } while (name != -1); + opterr = old_opterr; + + if (!fdsarg) + return -1; + + const char *orig_fdsarg = fdsarg; + + while (*fdsarg) { + unsigned long val; + char *endptr; + + val = strtoul(fdsarg, &endptr, 10); + if (fdsarg == endptr) { + die("Invalid character in --pass-fds option '%s' (near '%s')", + orig_fdsarg, fdsarg); + } + + if (val > INT_MAX) { + die("Invalid file descriptor in --pass-fds option '%s' (near '%s')", + orig_fdsarg, fdsarg); + } + + if (fds_cnt >= max_fds) + die("Too many file descriptors in --pass-fds"); + + fds[fds_cnt++] = (int)val; + + if (*endptr == ',') + fdsarg = endptr + 1; + else if (*endptr == '\0') + fdsarg = endptr; + else { + die("Invalid character in --pass-fds option '%s' (near '%s')", + orig_fdsarg, endptr); + } + } + + return fds_cnt; +} + /** * conf_addr() - Configure guest address with -a option * @c: Execution context @@ -1331,6 +1398,7 @@ void conf(struct ctx *c, int argc, char **argv) {"stats", required_argument, NULL, 31 }, {"conf-path", required_argument, NULL, 'c' }, {"chroot-fallback", no_argument, NULL, 32 }, + {"pass-fds", required_argument, NULL, 33 }, { 0 }, }; const char *optstring = "+dqfel:hs:c:F:I:p:P:m:a:n:M:g:i:o:D:S:H:461t:u:T:U:"; @@ -1572,6 +1640,10 @@ void conf(struct ctx *c, int argc, char **argv) case 32: c->chroot_fallback = true; break; + case 33: + if (c->mode != MODE_PASTA) + die("--pass-fds is for pasta mode only"); + break; case 'd': c->debug = 1; c->quiet = 0; diff --git a/conf.h b/conf.h index 19bf9bc..28a9acc 100644 --- a/conf.h +++ b/conf.h @@ -6,7 +6,10 @@ #ifndef CONF_H #define CONF_H +#define PASS_FDS_MAX 1024 + enum passt_modes conf_mode(int argc, char *argv[]); +int conf_pass_fds(int argc, char **argv, int *fds, int max_fds); int conf_tap_fd(int argc, char **argv); void conf(struct ctx *c, int argc, char **argv); void conf_listen_handler(struct ctx *c, uint32_t events); diff --git a/isolation.c b/isolation.c index 94cbe7f..25f599b 100644 --- a/isolation.c +++ b/isolation.c @@ -249,32 +249,77 @@ void isolate_initial(void) } /* - * isolate_fds() - Close leaked files, but not --fd, stdin, stdout, stderr + * isolate_fds() - Close leaked files, but not --fd, --pass-fds, standard streams * @argc: Argument count - * @argv: Command line options, as we need to skip any file given via --fd + * @argv: Command line options * * Should: - * - close all open files except for standard streams and the one from --fd + * - close all open files except for standard streams, --fd, and --pass-fds * - move the --fd descriptor out of the range 0-2 * * Return: new fd number for descriptor from --fd, or -1 if not specified */ int isolate_fds(int argc, char **argv) { - int fd, close_from = STDERR_FILENO + 1; + int fds[PASS_FDS_MAX + 1]; + int fds_cnt = 0; + int prev_fd; + int next_fd; + int tap_fd; + int rc = 0; + int i; + + tap_fd = conf_tap_fd(argc, argv); + if (tap_fd >= 0 && tap_fd < 3) { + /* Move the tap fd to a safer location */ + int new_fd = fcntl(tap_fd, F_DUPFD, STDERR_FILENO + 1); + + if (new_fd < 0) + die_perror("Could not relocate --fd descriptor"); + + close(tap_fd); + tap_fd = new_fd; + } + + if (tap_fd >= 0) + /* Keep the tap fd */ + fds[fds_cnt++] = tap_fd; + + rc = conf_pass_fds(argc, argv, fds + fds_cnt, PASS_FDS_MAX); + if (rc > 0) + /* Keep the pass-fds */ + fds_cnt += rc; - fd = conf_tap_fd(argc, argv); + rc = 0; - if (fd >= 0) { - /* Move the passed fd to a more convenient location */ - if (fd != close_from && - (dup2(fd, close_from) != close_from || - close(fd))) - die_perror("Could not move --fd descriptor"); - fd = close_from++; + /* Keep standard streams */ + prev_fd = STDERR_FILENO; + + while (1) { + next_fd = -1; + + /* Find the next-lowest fd to keep */ + for (i = 0; i < fds_cnt; i++) { + if (fds[i] > prev_fd && (next_fd == -1 || fds[i] < next_fd)) + next_fd = fds[i]; + } + + if (next_fd == -1) + break; + + if (next_fd > prev_fd + 1) { + /* Close fds between two kept fds */ + if (close_range(prev_fd + 1, next_fd - 1, CLOSE_RANGE_UNSHARE)) + rc = -1; + } + prev_fd = next_fd; } + + /* Close all other fds */ + if (close_range(prev_fd + 1, ~0U, CLOSE_RANGE_UNSHARE)) + rc = -1; - if (close_range(close_from, ~0U, CLOSE_RANGE_UNSHARE)) { + if (rc) { if (errno == ENOSYS || errno == EINVAL) { /* This probably means close_range() or the * CLOSE_RANGE_UNSHARE flag is not supported by the @@ -288,7 +333,7 @@ int isolate_fds(int argc, char **argv) } } - return fd; + return tap_fd; } /** diff --git a/passt.1 b/passt.1 index 995590a..bcd3aff 100644 --- a/passt.1 +++ b/passt.1 @@ -755,6 +755,11 @@ of local traffic in pasta\fR in the \fBNOTES\fR for more details. Do not create a tap device in the namespace. In this mode, \fIpasta\fR only forwards loopback traffic between namespaces. +.TP +.BR \-\-pass\-fds " " \fIfds\fR +Pass a comma-separated list of file descriptors to the spawned command. +These file descriptors will be kept open. + .SH EXAMPLES .SS \fBpasta -- 2.52.0

4 6

[PATCH 0/2] pasta: Do not configure ID mappings when invoked with --netns-only
by Dwayne B. Bent 22 Jul '26

22 Jul '26

This is a narrow fix for bug 216 where `pasta` attempts to configure user and group ID mappings when invoked with `--netns-only`. A new `bool` argument `config_idmaps` was added to `pasta_start_ns()` that guards the logic to configure user and group ID mappings. It is set to `false` when `netns_only` is `true`. OpenAI Codex was used to explore and understand the codebase. Code written myself. Test written by Codex guided and reviewed by me. Dwayne B. Bent (2): pasta: Regression test for bug 216 pasta: Do not configure ID mappings when invoked with --netns-only conf.c | 2 +- pasta.c | 24 ++++++++++++++---------- pasta.h | 3 ++- test/pasta_options/netns_only | 19 +++++++++++++++++++ test/run | 1 + 5 files changed, 37 insertions(+), 12 deletions(-) create mode 100644 test/pasta_options/netns_only -- 2.55.0

2 4

[PATCH] udp, icmp: Remove unused timer_run fields from protocol contexts
by Laurent Vivier 21 Jul '26

21 Jul '26

Since 260075bde769 ("tcp, udp, fwd: Run all port scanning from a single timer"), CALL_PROTO_HANDLER() is only used for TCP. UDP and ICMP timers now run from flow_defer_handler(), making the timer_run fields in struct udp_ctx and struct icmp_ctx dead stores. Remove the unused fields, drop struct icmp_ctx entirely (it has no remaining members), and stop initialising them in timer_init(). Signed-off-by: Laurent Vivier <lvivier(a)redhat.com> --- icmp.h | 7 ------- passt.c | 2 +- passt.h | 2 -- udp.h | 2 -- 4 files changed, 1 insertion(+), 12 deletions(-) diff --git a/icmp.h b/icmp.h index e8b928b61370..556260461995 100644 --- a/icmp.h +++ b/icmp.h @@ -19,12 +19,5 @@ int icmp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, struct iov_tail *data, const struct timespec *now); void icmp_init(void); -/** - * struct icmp_ctx - Execution context for ICMP routines - * @timer_run: Timestamp of most recent timer run - */ -struct icmp_ctx { - struct timespec timer_run; -}; #endif /* ICMP_H */ diff --git a/passt.c b/passt.c index b1d35d5bf988..677c8471a105 100644 --- a/passt.c +++ b/passt.c @@ -141,7 +141,7 @@ static void random_init(struct ctx *c) */ static void timer_init(struct ctx *c, const struct timespec *now) { - c->tcp.timer_run = c->udp.timer_run = c->icmp.timer_run = *now; + c->tcp.timer_run = *now; } /** diff --git a/passt.h b/passt.h index 3a072945c382..205d7eecede7 100644 --- a/passt.h +++ b/passt.h @@ -194,7 +194,6 @@ struct ip6_ctx { * @no_udp: Disable UDP operation * @udp: Context for UDP protocol handler * @no_icmp: Disable ICMP operation - * @icmp: Context for ICMP protocol handler * @mtu: MTU passed via DHCP/NDP * @no_dns: Do not source/use DNS servers for any purpose * @no_dns_search: Do not source/use domain search lists for any purpose @@ -279,7 +278,6 @@ struct ctx { int no_udp; struct udp_ctx udp; int no_icmp; - struct icmp_ctx icmp; int no_dns; int no_dns_search; diff --git a/udp.h b/udp.h index 9ea7de682caf..b50283eab198 100644 --- a/udp.h +++ b/udp.h @@ -26,14 +26,12 @@ void udp_update_l2_buf(const unsigned char *eth_d); * struct udp_ctx - Execution context for UDP * @scan_in: Port scanning state for inbound packets * @scan_out: Port scanning state for outbound packets - * @timer_run: Timestamp of most recent timer run * @timeout: Timeout for unidirectional flows (in s) * @stream_timeout: Timeout for stream-like flows (in s) */ struct udp_ctx { struct fwd_scan scan_in; struct fwd_scan scan_out; - struct timespec timer_run; int timeout; int stream_timeout; }; -- 2.54.0

3 3

[PATCH v5 1/7] dhcp: Refactor fill_one() to operate on a generic buffer
by Anshu Kumari 20 Jul '26

20 Jul '26

Change fill_one() to accept a buffer pointer and capacity instead of a struct msg pointer. This is a pure refactor with no behavior change, preparing for option overload support where fill_one() will also write into the file and sname fields. Link: https://bugs.passt.top/show_bug.cgi?id=192 Signed-off-by: Anshu Kumari <anskuma(a)redhat.com> --- v5: - Restore debug messages for skipped options in fill() - Add /* code and length of option */ comment in size check - Use buf + *offset instead of &buf[*offset] for memcpy v3: - Restored removed comments: "If we don't have space to write the option, then just skip" and "Move to option". v2: - Renamed parameter cap → size. --- dhcp.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) diff --git a/dhcp.c b/dhcp.c index c3c7422..bb72b72 100644 --- a/dhcp.c +++ b/dhcp.c @@ -131,28 +131,29 @@ struct msg { } __attribute__((__packed__)); /** - * fill_one() - Fill a single option in message - * @m: Message to fill + * fill_one() - Fill a single option into a buffer + * @buf: Buffer to write option + * @size: Usable size of @buf (excluding end marker) * @o: Option number - * @offset: Current offset within options field, updated on insertion + * @offset: Current offset within @buf, updated on insertion * - * Return: false if m has space to write the option, true otherwise + * Return: false if @buf has space to write the option, true otherwise */ -static bool fill_one(struct msg *m, int o, int *offset) +static bool fill_one(uint8_t *buf, size_t size, int o, int *offset) { size_t slen = opts[o].slen; /* If we don't have space to write the option, then just skip */ - if (*offset + 2 /* code and length of option */ + slen > OPT_MAX) + if (*offset + 2 /* code and length of option */ + slen > size) return true; - m->o[*offset] = o; - m->o[*offset + 1] = slen; + buf[*offset] = o; + buf[*offset + 1] = slen; /* Move to option */ *offset += 2; - memcpy(&m->o[*offset], opts[o].s, slen); + memcpy(buf + *offset, opts[o].s, slen); opts[o].sent = 1; *offset += slen; @@ -177,19 +178,19 @@ static int fill(struct msg *m) * Put it there explicitly, unless requested via option 55. */ if (opts[55].clen > 0 && !memchr(opts[55].c, 53, opts[55].clen)) - if (fill_one(m, 53, &offset)) - debug("DHCP: skipping option 53"); + if (fill_one(m->o, OPT_MAX, 53, &offset)) + debug("DHCP: skipping option 53"); for (i = 0; i < opts[55].clen; i++) { o = opts[55].c[i]; if (opts[o].slen != -1) - if (fill_one(m, o, &offset)) + if (fill_one(m->o, OPT_MAX, o, &offset)) debug("DHCP: skipping option %i", o); } for (o = 0; o < 255; o++) { if (opts[o].slen != -1 && !opts[o].sent) - if (fill_one(m, o, &offset)) + if (fill_one(m->o, OPT_MAX, o, &offset)) debug("DHCP: skipping option %i", o); } -- 2.54.0

2 1

[PATCH v2] feat: Add cli option '--pass-fds' for pasta mode.
by Richard Lawrence 19 Jul '26

19 Jul '26

From: Richard Lawrence <rarensu(a)tamu.edu> When pasta mode is used to launch an executable (`pasta [COMMAND]`) and that executable accepts inputs in the form of arbitrary file descriptors (such as `bwrap`), then passt should not stand in the way of the parent process handing off those file descriptors to the child process. See bug 204 for additional discussion. The `pass-fds` option accepts a comma-separated list of file descriptor numbers. `conf_pass_fds()` parses the command line argument, then `isolate_fds()` skips closing the specified fds by calling `close_range()` on the gaps between them. Additionally, the tap fd is safely relocated to the lowest unused fd number which it at least 3, to avoid accidentally overwriting an existing fd. Signed-off-by: Richard Lawrence <rlawrence(a)tamu.edu> --- conf.c | 66 ++++++++++++++++++++++++++++++++++++++++++++++++++- conf.h | 1 + isolation.c | 68 ++++++++++++++++++++++++++++++++++++++++++----------- passt.1 | 5 ++++ 4 files changed, 125 insertions(+), 15 deletions(-) diff --git a/conf.c b/conf.c index 0fcba5c..3246735 100644 --- a/conf.c +++ b/conf.c @@ -732,7 +732,9 @@ pasta_opts: " Don't copy all addresses to namespace\n" " --ns-mac-addr ADDR Set MAC address on tap interface\n" " --no-splice Disable inbound socket splicing\n" - " --splice-only Only enable loopback forwarding\n"); + " --splice-only Only enable loopback forwarding\n" + " --pass-fds FDS Comma-separated list of fds to pass to\n" + " the spawned command\n"); passt_exit(status); } @@ -1183,6 +1185,63 @@ int conf_tap_fd(int argc, char **argv) return val; } +/** + * conf_pass_fds() - Read fds as supplied by --pass-fds command line option + * @argc: Argument count + * @argv: Command line options + * @fds: Array where we store the parsed fds + * @max_fds: Maximum size of the array + * + * Return: number of parsed fds, or -1 if option not specified + */ +int conf_pass_fds(int argc, char **argv, int *fds, int max_fds) +{ + const struct option opt[] = { { "pass-fds", required_argument, NULL, 33 }, + { 0 }, }; + const char *fdsarg = NULL; + int name, fds_cnt = 0; + int old_opterr; + + old_opterr = opterr; + opterr = 0; + optind = 0; + do { + name = getopt_long(argc, argv, "-:", opt, NULL); + if (name == 33) + fdsarg = optarg; + } while (name != -1); + opterr = old_opterr; + + if (!fdsarg) + return -1; + + while (*fdsarg) { + unsigned long val; + char *endptr; + + val = strtoul(fdsarg, &endptr, 10); + if (fdsarg == endptr) + die("Invalid --pass-fds option: %s", fdsarg); + + if (val > INT_MAX) + die("Invalid file descriptor in --pass-fds: %lu", val); + + if (fds_cnt >= max_fds) + die("Too many file descriptors in --pass-fds"); + + fds[fds_cnt++] = (int)val; + + if (*endptr == ',') + fdsarg = endptr + 1; + else if (*endptr == '\0') + fdsarg = endptr; + else + die("Invalid character in --pass-fds option: %s", fdsarg); + } + + return fds_cnt; +} + /** * conf_addr() - Configure guest address with -a option * @c: Execution context @@ -1331,6 +1390,7 @@ void conf(struct ctx *c, int argc, char **argv) {"stats", required_argument, NULL, 31 }, {"conf-path", required_argument, NULL, 'c' }, {"chroot-fallback", no_argument, NULL, 32 }, + {"pass-fds", required_argument, NULL, 33 }, { 0 }, }; const char *optstring = "+dqfel:hs:c:F:I:p:P:m:a:n:M:g:i:o:D:S:H:461t:u:T:U:"; @@ -1572,6 +1632,10 @@ void conf(struct ctx *c, int argc, char **argv) case 32: c->chroot_fallback = true; break; + case 33: + if (c->mode != MODE_PASTA) + die("--pass-fds is for pasta mode only"); + break; case 'd': c->debug = 1; c->quiet = 0; diff --git a/conf.h b/conf.h index 19bf9bc..3489f35 100644 --- a/conf.h +++ b/conf.h @@ -7,6 +7,7 @@ #define CONF_H enum passt_modes conf_mode(int argc, char *argv[]); +int conf_pass_fds(int argc, char **argv, int *fds, int max_fds); int conf_tap_fd(int argc, char **argv); void conf(struct ctx *c, int argc, char **argv); void conf_listen_handler(struct ctx *c, uint32_t events); diff --git a/isolation.c b/isolation.c index 94cbe7f..0632f7f 100644 --- a/isolation.c +++ b/isolation.c @@ -249,32 +249,72 @@ void isolate_initial(void) } /* - * isolate_fds() - Close leaked files, but not --fd, stdin, stdout, stderr + * isolate_fds() - Close leaked files, but not --fd, --pass-fds, standard streams * @argc: Argument count - * @argv: Command line options, as we need to skip any file given via --fd + * @argv: Command line options * * Should: - * - close all open files except for standard streams and the one from --fd + * - close all open files except for standard streams, --fd, and --pass-fds * - move the --fd descriptor out of the range 0-2 * * Return: new fd number for descriptor from --fd, or -1 if not specified */ int isolate_fds(int argc, char **argv) { - int fd, close_from = STDERR_FILENO + 1; + int prev_fd = STDERR_FILENO; // Keep standard streams + int fds[1024 + 1]; + int fds_cnt = 0; + int tap_fd; + int rc = 0; + + tap_fd = conf_tap_fd(argc, argv); + if (tap_fd >= 0 && tap_fd < 3) { + /* Move the passed fd to a more convenient location */ + int new_fd = fcntl(tap_fd, F_DUPFD, 3); + + if (new_fd < 0) + die_perror("Could not relocate --fd descriptor"); - fd = conf_tap_fd(argc, argv); + close(tap_fd); + tap_fd = new_fd; + } - if (fd >= 0) { - /* Move the passed fd to a more convenient location */ - if (fd != close_from && - (dup2(fd, close_from) != close_from || - close(fd))) - die_perror("Could not move --fd descriptor"); - fd = close_from++; + if (tap_fd >= 0) + /* Keep the tap fd */ + fds[fds_cnt++] = tap_fd; + + rc = conf_pass_fds(argc, argv, fds + fds_cnt, 1024); + if (rc > 0) + /* Keep the pass-fds */ + fds_cnt += rc; + + rc = 0; + + while (1) { + int min_fd = -1; + + /* Find the next-lowest fd to keep */ + for (int i = 0; i < fds_cnt; i++) { + if (fds[i] > prev_fd && (min_fd == -1 || fds[i] < min_fd)) + min_fd = fds[i]; + } + + if (min_fd == -1) + break; + + if (min_fd > prev_fd + 1) { + /* Close fds between two kept fds */ + if (close_range(prev_fd + 1, min_fd - 1, CLOSE_RANGE_UNSHARE)) + rc = -1; + } + prev_fd = min_fd; } + + /* Close all other fds */ + if (close_range(prev_fd + 1, ~0U, CLOSE_RANGE_UNSHARE)) + rc = -1; - if (close_range(close_from, ~0U, CLOSE_RANGE_UNSHARE)) { + if (rc) { if (errno == ENOSYS || errno == EINVAL) { /* This probably means close_range() or the * CLOSE_RANGE_UNSHARE flag is not supported by the @@ -288,7 +328,7 @@ int isolate_fds(int argc, char **argv) } } - return fd; + return tap_fd; } /** diff --git a/passt.1 b/passt.1 index 995590a..bcd3aff 100644 --- a/passt.1 +++ b/passt.1 @@ -755,6 +755,11 @@ of local traffic in pasta\fR in the \fBNOTES\fR for more details. Do not create a tap device in the namespace. In this mode, \fIpasta\fR only forwards loopback traffic between namespaces. +.TP +.BR \-\-pass\-fds " " \fIfds\fR +Pass a comma-separated list of file descriptors to the spawned command. +These file descriptors will be kept open. + .SH EXAMPLES .SS \fBpasta -- 2.52.0

2 2

[PATCH] pif: Add message to static_assert for C11 compliance, fix build with gcc 8
by Stefano Brivio 19 Jul '26

19 Jul '26

Seen with gcc 8.5: --- In file included from passt.h:18, from qrap.c:36: pif.h: In function 'pif_type': pif.h:48:44: error: expected ',' before ')' token static_assert(sizeof("?") <= PIF_NAME_SIZE); ^ make: *** [Makefile:105: qrap] Error 1 --- As noted in ba84a3b17af8 ("treewide: Add messages to static_assert() calls"), static_assert() calls need a message to comply with C11. The form without message is supported starting from C23 only. Fixes: cbcd4284111f ("pif: Limit pif names to 128 bytes") Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com> --- pif.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/pif.h b/pif.h index 3a1e2e5..dedcaa5 100644 --- a/pif.h +++ b/pif.h @@ -45,7 +45,7 @@ static inline const char *pif_type(enum pif_type pt) return pif_type_str[pt]; else return "?"; - static_assert(sizeof("?") <= PIF_NAME_SIZE); + static_assert(sizeof("?") <= PIF_NAME_SIZE, "PIF_NAME_SIZE too small"); } static inline const char *pif_name(uint8_t pif) -- 2.43.0

2 2

[PATCH v2 0/6] Fix bug 215 and some related issues with fd handling
by David Gibson 19 Jul '26

19 Jul '26

Stefano, you called it correctly. While working on bug 215, as usual, I found a bunch of adjacent things to clean up. I did start by attempting the appoach you suggested for bug 215 - remembering which of the low fds were standard streams and avoiding closing them in __daemon(). It is indeed shorter, but only by 1-2 lines. Looking at possible interactions with other things, I became more and more convinced that leaving anything other than the standard streams in fds 0-2 was an accident waiting to happen. Much of the rest of the series is, for example, dealing with the possibility of --fd [012]. v2: * No longer fold fd closing into isolate_initial() * Some minor polish * Various improved comments and commit messages David Gibson (6): passt: Always close pidfile_fd, not just when daemonizing isolation: Move close_open_files() to isolate_fds() isolation, conf: Set c->fd_tap from early parse of --fd conf: Make conf_tap_fd() operate more like conf_mode() isolation: Move --fd descriptor to a number of our choosing main: Ensure fds 0-2 are populated conf.c | 31 ++++++++++++++++++++------- conf.h | 2 +- isolation.c | 60 ++++++++++++++++++++++++++++++++++++++++++++++------- isolation.h | 3 ++- passt.c | 28 +++++++++++++++++-------- util.c | 52 +--------------------------------------------- util.h | 1 - 7 files changed, 98 insertions(+), 79 deletions(-) -- 2.55.0

2 6

[PATCH] treewide: Sandbox qrap
by Stefano Brivio 19 Jul '26

19 Jul '26

No sunsetting or rightsizing for you, qrap: dig, bury, and cover. Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com> --- .gitignore | 1 - Makefile | 11 +- README.md | 4 - contrib/fedora/passt.spec | 2 - doc/demo.sh | 12 +- doc/passt_overview.excalidraw | 29 --- hooks/pre-push | 2 +- passt.1 | 235 ++++++++++--------- qrap.1 | 77 ------ qrap.c | 429 ---------------------------------- tap.c | 2 - test/distro/opensuse | 2 +- 12 files changed, 127 insertions(+), 679 deletions(-) delete mode 100644 qrap.1 delete mode 100644 qrap.c diff --git a/.gitignore b/.gitignore index 3e40d9f..71420b6 100644 --- a/.gitignore +++ b/.gitignore @@ -5,7 +5,6 @@ /pasta.avx2 /passt-repair /pesto -/qrap /pasta.1 /seccomp.h /seccomp_pesto.h diff --git a/Makefile b/Makefile index 5757aef..77474d6 100644 --- a/Makefile +++ b/Makefile @@ -39,13 +39,12 @@ PASST_SRCS = arch.c arp.c bitmap.c checksum.c conf.c dhcp.c dhcpv6.c \ parse.c passt.c pasta.c pcap.c pif.c repair.c serialise.c tap.c tcp.c \ tcp_buf.c tcp_splice.c tcp_vu.c udp.c udp_flow.c udp_vu.c util.c \ vhost_user.c virtio.c vu_common.c -QRAP_SRCS = qrap.c PASST_REPAIR_SRCS = passt-repair.c PESTO_SRCS = pesto.c bitmap.c fwd_rule.c inany.c ip.c lineread.c parse.c \ serialise.c -SRCS = $(PASST_SRCS) $(QRAP_SRCS) $(PASST_REPAIR_SRCS) $(PESTO_SRCS) +SRCS = $(PASST_SRCS) $(PASST_REPAIR_SRCS) $(PESTO_SRCS) -MANPAGES = passt.1 pasta.1 pesto.1 qrap.1 passt-repair.1 +MANPAGES = passt.1 pasta.1 pesto.1 passt-repair.1 PASST_HEADERS = arch.h arp.h bitmap.h checksum.h conf.h dhcp.h dhcpv6.h \ epoll_ctl.h flow.h fwd.h fwd_rule.h flow_table.h icmp.h icmp_flow.h \ @@ -54,7 +53,6 @@ PASST_HEADERS = arch.h arp.h bitmap.h checksum.h conf.h dhcp.h dhcpv6.h \ serialise.h siphash.h tap.h tcp.h tcp_buf.h tcp_conn.h tcp_internal.h \ tcp_splice.h tcp_vu.h udp.h udp_flow.h udp_internal.h udp_vu.h util.h \ vhost_user.h virtio.h vu_common.h -QRAP_HEADERS = arp.h ip.h passt.h util.h PASST_REPAIR_HEADERS = linux_dep.h PESTO_HEADERS = bitmap.h common.h fwd_rule.h inany.h ip.h log.h parse.h \ pesto.h serialise.h @@ -76,7 +74,7 @@ docdir ?= $(datarootdir)/doc/passt mandir ?= $(datarootdir)/man man1dir ?= $(mandir)/man1 -BASEBIN := passt qrap passt-repair pesto +BASEBIN := passt passt-repair pesto ifeq ($(TARGET_ARCH),x86_64) BASEBIN += passt.avx2 endif @@ -112,9 +110,6 @@ passt.avx2: $(PASST_SRCS) $(PASST_HEADERS) seccomp.h pasta.avx2 pasta.1 pasta: pasta%: passt% ln -sf $< $@ -qrap: BASE_CPPFLAGS += -DARCH=\"$(TARGET_ARCH)\" -qrap: $(QRAP_SRCS) $(QRAP_HEADERS) - passt-repair: $(PASST_REPAIR_SRCS) $(PASST_REPAIR_HEADERS) seccomp_repair.h pesto: BASE_CPPFLAGS += -DPESTO diff --git a/README.md b/README.md index 8fdc0a3..43caf49 100644 --- a/README.md +++ b/README.md @@ -376,10 +376,6 @@ speeding up local connections, and usually requiring NAT. _pasta_: _passt_ exchanges packets with _qemu_ via UNIX domain socket, using the `socket` back-end in qemu. This is supported since qemu 7.2. -For older versions, the [qrap](/passt/tree/qrap.c) wrapper can be used to -connect to a UNIX domain socket and to start qemu, which can now use the file -descriptor that's already opened. - This approach, compared to using a _tap_ device, doesn't require any security capabilities, as we don't need to create any interface. diff --git a/contrib/fedora/passt.spec b/contrib/fedora/passt.spec index 34838af..6c7530a 100644 --- a/contrib/fedora/passt.spec +++ b/contrib/fedora/passt.spec @@ -118,12 +118,10 @@ fi %doc %{_docdir}/%{name}/demo.sh %{_bindir}/passt %{_bindir}/pasta -%{_bindir}/qrap %{_bindir}/passt-repair %{_bindir}/pesto %{_mandir}/man1/passt.1* %{_mandir}/man1/pasta.1* -%{_mandir}/man1/qrap.1* %{_mandir}/man1/passt-repair.1* %{_mandir}/man1/pesto.1* %ifarch x86_64 diff --git a/doc/demo.sh b/doc/demo.sh index 6b0917c..b56669f 100755 --- a/doc/demo.sh +++ b/doc/demo.sh @@ -177,7 +177,7 @@ into_ns() { next [ "${KEY}" = "s" ] && ${SHELL} - cmd passt -P "${DEMO_DIR}/passt.pid" + cmd passt -P "${DEMO_DIR}/passt.pid" -s "${DEMO_DIR}/passt.socket" echo echo "...passt is running." next @@ -198,15 +198,12 @@ into_ns() { echo "can download and use mbuto (https://mbuto.sh/) to build a" echo "basic initramfs image. Otherwise, press 's' to skip this" echo "step, and start an existing virtual machine yourself." - echo "You'll need to use the qrap(1) wrapper, with qemu options" - echo "as reported above." next else echo "This script doesn't know, yet, how to run a virtual" echo "machine on your architecture (${__arch}). Please start an" - echo "existing virtual machine yourself, using the qrap(1)" - echo "wrapper, with qemu options as reported above." + echo "existing virtual machine yourself." echo fi @@ -225,12 +222,13 @@ into_ns() { next # shellcheck disable=SC2086 - cmd qrap 5 ${__qemu_arch} \ + cmd ${__qemu_arch} \ -smp "$(nproc)" -m 1024 \ -nographic -serial stdio -nodefaults -no-reboot -vga none \ -initrd "${DEMO_DIR}/demo.img" \ -kernel "/boot/vmlinuz-$(uname -r)" -append "console=ttyS0" \ - -net socket,fd=5 -net nic,model=virtio || : + -device virtio-net-pci,netdev=s \ + -netdev stream,id=s,server=off,addr.type=unix,addr.path="${DEMO_DIR}/passt.socket" || : } STTY_BACKUP="$(stty -g)" diff --git a/doc/passt_overview.excalidraw b/doc/passt_overview.excalidraw index 790728e..3217b3f 100644 --- a/doc/passt_overview.excalidraw +++ b/doc/passt_overview.excalidraw @@ -3595,35 +3595,6 @@ "baseline": 17, "textAlign": "left", "verticalAlign": "top" - }, - { - "type": "text", - "version": 1448, - "versionNonce": 1345391945, - "isDeleted": false, - "id": "c-esZiuHSm3wnTRlmIJZk", - "fillStyle": "hachure", - "strokeWidth": 1, - "strokeStyle": "dotted", - "roughness": 1, - "opacity": 100, - "angle": 4.969352159599495, - "x": 816.7081559166961, - "y": 215.74807066759342, - "strokeColor": "#2b8a3e", - "backgroundColor": "transparent", - "width": 211, - "height": 23, - "seed": 939668423, - "groupIds": [], - "strokeSharpness": "round", - "boundElementIds": [], - "fontSize": 20, - "fontFamily": 1, - "text": "(currently \"qrap\" wrapper)", - "baseline": 17, - "textAlign": "left", - "verticalAlign": "top" } ], "appState": { diff --git a/hooks/pre-push b/hooks/pre-push index 3e2d15f..8eebdbc 100755 --- a/hooks/pre-push +++ b/hooks/pre-push @@ -54,7 +54,7 @@ done cd .. make pkgs -scp passt passt.avx2 passt.1 qrap qrap.1 "${USER_HOST}:${BIN}" +scp passt passt.avx2 passt.1 "${USER_HOST}:${BIN}" scp pasta pasta.avx2 pasta.1 "${USER_HOST}:${BIN}" scp passt-repair passt-repair.1 "${USER_HOST}:${BIN}" scp pesto pesto.1 "${USER_HOST}:${BIN}" diff --git a/passt.1 b/passt.1 index 995590a..53e072a 100644 --- a/passt.1 +++ b/passt.1 @@ -580,8 +580,7 @@ Default is \fBnone\fR for \fBpasst\fR and \fBauto\fR for \fBpasta\fR. .TP .BR \-s ", " \-\-socket-path ", " \-\-socket " " \fIpath -Path for UNIX domain socket used by \fBqemu\fR(1) or \fBqrap\fR(1) to connect to -\fBpasst\fR. +Path for UNIX domain socket used by \fBqemu\fR(1) to connect to \fBpasst\fR. Default is to probe a free socket, not accepting connections, starting from \fI/tmp/passt_1.socket\fR to \fI/tmp/passt_64.socket\fR. @@ -762,129 +761,132 @@ forwards loopback traffic between namespaces. .RS .nf $ iperf3 -s -D -$ ./pasta -Outbound interface: eth0, namespace interface: eth0 -ARP: - address: 28:16:ad:39:a9:ea -DHCP: - assign: 192.168.1.118 - mask: 255.255.255.0 - router: 192.168.1.1 -NDP/DHCPv6: - assign: 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17 - router: fe80::62e3:27ff:fe33:2b01 -# -# dhclient -4 --no-pid -# dhclient -6 --no-pid + +$ pasta --config-net + # ip address show 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever - inet6 ::1/128 scope host + inet6 ::1/128 scope host proto kernel_lo valid_lft forever preferred_lft forever -2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 65520 qdisc pfifo_fast state UNKNOWN group default qlen 1000 - link/ether 5e:90:02:eb:b0:2a brd ff:ff:ff:ff:ff:ff - inet 192.168.1.118/24 brd 192.168.1.255 scope global eth0 +2: enp9s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 65520 qdisc fq_codel state UNKNOWN group default qlen 1000 + link/ether be:16:49:5a:65:dc brd ff:ff:ff:ff:ff:ff + inet 88.198.0.164/27 brd 88.198.0.191 scope global noprefixroute enp9s0 valid_lft forever preferred_lft forever - inet6 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17/128 scope global + inet6 2a01:4f8:222:904::2/64 scope global nodad valid_lft forever preferred_lft forever - inet6 2a02:6d40:3ca5:2001:5c90:2ff:feeb:b02a/64 scope global dynamic mngtmpaddr - valid_lft 3591sec preferred_lft 3591sec - inet6 fe80::5c90:2ff:feeb:b02a/64 scope link + inet6 fe80::bc16:49ff:fe5a:65dc/64 scope link nodad proto kernel_ll valid_lft forever preferred_lft forever + # ip route show -default via 192.168.1.1 dev eth0 -192.168.1.0/24 dev eth0 proto kernel scope link src 192.168.1.118 +default via 88.198.0.161 dev enp9s0 proto dhcp metric 1002 +88.198.0.160/27 dev enp9s0 proto dhcp scope link metric 1002 + # ip -6 route show -2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17 dev eth0 proto kernel metric 256 pref medium -2a02:6d40:3ca5:2001::/64 dev eth0 proto kernel metric 256 expires 3584sec pref medium -fe80::/64 dev eth0 proto kernel metric 256 pref medium -default via fe80::62e3:27ff:fe33:2b01 dev eth0 proto ra metric 1024 expires 3584sec pref medium +2a01:4f8:222:904::/64 dev enp9s0 proto kernel metric 256 pref medium +fe80::/64 dev enp9s0 proto kernel metric 256 pref medium +default via fe80::1 dev enp9s0 metric 1024 onlink pref medium + # iperf3 -c 127.0.0.1 -t1 Connecting to host 127.0.0.1, port 5201 -[ 5] local 127.0.0.1 port 51938 connected to 127.0.0.1 port 5201 +[ 5] local 127.0.0.1 port 59112 connected to 127.0.0.1 port 5201 [ ID] Interval Transfer Bitrate Retr Cwnd -[ 5] 0.00-1.00 sec 4.46 GBytes 38.3 Gbits/sec 0 3.93 MBytes +[ 5] 0.00-1.00 sec 3.89 GBytes 33.4 Gbits/sec 0 2.62 MBytes - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bitrate Retr -[ 5] 0.00-1.00 sec 4.46 GBytes 38.3 Gbits/sec 0 sender -[ 5] 0.00-1.41 sec 4.45 GBytes 27.1 Gbits/sec receiver +[ 5] 0.00-1.00 sec 3.89 GBytes 33.4 Gbits/sec 0 sender +[ 5] 0.00-1.00 sec 3.89 GBytes 33.4 Gbits/sec receiver iperf Done. + # iperf3 -c ::1 -t1 Connecting to host ::1, port 5201 -[ 5] local ::1 port 50108 connected to ::1 port 5201 +[ 5] local ::1 port 49138 connected to ::1 port 5201 [ ID] Interval Transfer Bitrate Retr Cwnd -[ 5] 0.00-1.00 sec 4.35 GBytes 37.4 Gbits/sec 0 4.99 MBytes +[ 5] 0.00-1.00 sec 4.32 GBytes 37.0 Gbits/sec 0 1.06 MBytes - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bitrate Retr -[ 5] 0.00-1.00 sec 4.35 GBytes 37.4 Gbits/sec 0 sender -[ 5] 0.00-1.41 sec 4.35 GBytes 26.4 Gbits/sec receiver +[ 5] 0.00-1.00 sec 4.32 GBytes 37.0 Gbits/sec 0 sender +[ 5] 0.00-1.00 sec 4.32 GBytes 37.0 Gbits/sec receiver iperf Done. -# ping -c1 -4 spaghetti.pizza -PING spaghetti.pizza (172.67.192.217) 56(84) bytes of data. -64 bytes from 172.67.192.217: icmp_seq=1 ttl=255 time=37.3 ms + +# ping -nc1 -4 spaghetti.pizza +PING spaghetti.pizza (5.161.47.86) 56(84) bytes of data. +64 bytes from 5.161.47.86: icmp_seq=1 ttl=255 time=99.4 ms --- spaghetti.pizza ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms -# ping -c1 -6 spaghetti.pizza -PING spaghetti.pizza(2606:4700:3034::6815:147a (2606:4700:3034::6815:147a)) 56 data bytes -64 bytes from 2606:4700:3034::6815:147a: icmp_seq=1 ttl=255 time=35.6 ms +rtt min/avg/max/mdev = 99.433/99.433/99.433/0.000 ms +# ping -nc1 -6 spaghetti.pizza +PING spaghetti.pizza (2a01:4f8:1c1e:d6f1::1) 56 data bytes +64 bytes from 2a01:4f8:1c1e:d6f1::1: icmp_seq=1 ttl=255 time=3.11 ms --- spaghetti.pizza ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms -rtt min/avg/max/mdev = 35.605/35.605/35.605/0.000 ms -# logout -$ - -.RE +rtt min/avg/max/mdev = 3.110/3.110/3.110/0.000 ms .fi +.RE .BR "Connect an existing user and network namespace" .RS .nf $ unshare -rUn -# echo $$ -2446678 +# echo $$ +1295969 .fi -.BR " [From another terminal]" + +.BR "[From another terminal]" .nf -$ ./pasta 2446678 -Outbound interface: eth0, namespace interface: eth0 -ARP: - address: 28:16:ad:39:a9:ea +$ pasta --config-net 1295969 +Template interface: enp9s0 (IPv4), enp9s0 (IPv6) +Namespace interface: enp9s0 +MAC: + host: 9a:55:9a:55:9a:55 + NAT to host 127.0.0.1: 88.198.0.161 DHCP: - assign: 192.168.1.118 - mask: 255.255.255.0 - router: 192.168.1.1 + assign: 88.198.0.164 + mask: 255.255.255.224 + router: 88.198.0.161 +DNS: + 185.12.64.1 + 185.12.64.2 + NAT to host ::1: fe80::1 NDP/DHCPv6: - assign: 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17 - router: fe80::62e3:27ff:fe33:2b01 - + assign: 2a01:4f8:222:904::2 + router: fe80::1 + our link-local: fe80::1 +Inbound forwarding rules (HOST): + TCP [*]:1-32767 => 1-32767 (best effort) (auto-scan) + TCP [*]:61000-65535 => 61000-65535 (best effort) (auto-scan) + UDP [*]:1-32767 => 1-32767 (best effort) (auto-scan) + UDP [*]:61000-65535 => 61000-65535 (best effort) (auto-scan) +Outbound forwarding rules (SPLICE): + TCP [*]%lo:1-32767 => 1-32767 (best effort) (auto-scan) + TCP [*]%lo:61000-65535 => 61000-65535 (best effort) (auto-scan) + UDP [*]%lo:1-32767 => 1-32767 (best effort) (auto-scan) + UDP [*]%lo:61000-65535 => 61000-65535 (best effort) (auto-scan) .fi -.BR " [Back to the original terminal]" + +.BR "[Back to the original terminal]" .nf -# dhclient -4 --no-pid -# dhclient -6 --no-pid # ip address show 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever - inet6 ::1/128 scope host + inet6 ::1/128 scope host proto kernel_lo valid_lft forever preferred_lft forever -2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 65520 qdisc pfifo_fast state UNKNOWN group default qlen 1000 - link/ether fa:c1:2a:27:92:a9 brd ff:ff:ff:ff:ff:ff - inet 192.168.1.118/24 brd 192.168.1.255 scope global eth0 +2: enp9s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 65520 qdisc fq_codel state UNKNOWN group default qlen 1000 + link/ether 9a:44:5e:54:d3:2c brd ff:ff:ff:ff:ff:ff + inet 88.198.0.164/27 brd 88.198.0.191 scope global noprefixroute enp9s0 valid_lft forever preferred_lft forever - inet6 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17/128 scope global + inet6 2a01:4f8:222:904::2/64 scope global nodad valid_lft forever preferred_lft forever - inet6 2a02:6d40:3ca5:2001:f8c1:2aff:fe27:92a9/64 scope global dynamic mngtmpaddr - valid_lft 3594sec preferred_lft 3594sec - inet6 fe80::f8c1:2aff:fe27:92a9/64 scope link + inet6 fe80::9844:5eff:fe54:d32c/64 scope link nodad proto kernel_ll valid_lft forever preferred_lft forever .fi .RE @@ -893,72 +895,69 @@ NDP/DHCPv6: .BR "Start and connect a guest with basic port forwarding" .RS .nf -$ ./passt -f -t 2222:22 -Outbound interface: eth0 -ARP: - address: 28:16:ad:39:a9:ea +$ passt -f -t 2222:22 +UNIX domain socket bound at /tmp/passt_1.socket +No IPv6 nameserver available for NDP/DHCPv6 +Template interface: enp9s0 (IPv4), enp9s0 (IPv6) +MAC: + host: 9a:55:9a:55:9a:55 + NAT to host 127.0.0.1: 88.198.0.161 DHCP: - assign: 192.168.1.118 - mask: 255.255.255.0 - router: 192.168.1.1 - search: - redhat.com + assign: 88.198.0.164 + mask: 255.255.255.224 + router: 88.198.0.161 +DNS: + 185.12.64.1 + 185.12.64.2 + NAT to host ::1: fe80::1 NDP/DHCPv6: - assign: 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17 - router: fe80::62e3:27ff:fe33:2b01 - search: - redhat.com -UNIX domain socket bound at /tmp/passt_1.socket - -You can now start qrap: - ./qrap 5 qemu-system-x86_64 ... -net socket,fd=5 -net nic,model=virtio -or directly qemu, patched with: - qemu/0001-net-Allow-also-UNIX-domain-sockets-to-be-used-as-net.patch -as follows: - qemu-system-x86_64 ... -net socket,connect=/tmp/passt_1.socket -net nic,model=virtio - + assign: 2a01:4f8:222:904::2 + router: fe80::1 + our link-local: fe80::1 +Inbound forwarding rules (HOST): + TCP [*]:2222 => 22 + +You can now start qemu (>= 7.2, with commit 13c6be96618c): + kvm ... -device virtio-net-pci,netdev=s -netdev stream,id=s,server=off,addr.type=unix,addr.path=/tmp/passt_1.socket .fi -.BR " [From another terminal]" -.nf -$ ./qrap 5 qemu-system-x86_64 test.qcow2 -m 1024 -display none -nodefaults -nographic -net socket,fd=5 -net nic,model=virtio -Connected to /tmp/passt_1.socket -.fi -.BR " [Back to the original terminal]" +.BR "[From another terminal]" .nf -passt: DHCP: ack to request -passt: from 52:54:00:12:34:56 -passt: NDP: received NS, sending NA -passt: NDP: received RS, sending RA -passt: DHCPv6: received SOLICIT, sending ADVERTISE -passt: NDP: received NS, sending NA -passt: DHCPv6: received REQUEST/RENEW/CONFIRM, sending REPLY -passt: NDP: received NS, sending NA +$ qemu-system-x86_64 -m 128 test.qcow2 -display none -nodefaults -nographic -device virtio-net-pci,netdev=s -netdev stream,id=s,server=off,addr.type=unix,addr.path=/tmp/passt_1.socket +.fi +.BR "[Back to the original terminal]" +.nf +accepted connection from PID 1301736 +New guest MAC address observed: 52:54:00:12:34:56 +DHCP: offer to discover + from 52:54:00:12:34:56 +DHCP: ack to request + from 52:54:00:12:34:56 +NDP: received RS, sending RA .fi -.BR " [From yet another terminal]" + +.BR "[From yet another terminal]" .nf $ ssh -p 2222 root@localhost root@localhost's password: .fi -.BR " [...]" +.BR "[...]" .nf # ip address show -1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 +1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever -2: ens2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 65520 qdisc pfifo_fast state UP group default qlen 1000 +2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 52:54:00:12:34:56 brd ff:ff:ff:ff:ff:ff - inet 192.168.1.118/24 brd 192.168.1.255 scope global noprefixroute ens2 + inet 88.198.0.164/27 scope global eth0 valid_lft forever preferred_lft forever - inet6 2a02:6d40:3ca5:2001:b81d:fa4a:8cdd:cf17/128 scope global noprefixroute + inet6 2a01:4f8:222:904:5054:ff:fe12:3456/64 scope global dynamic flags 100 valid_lft forever preferred_lft forever - inet6 2a02:6d40:3ca5:2001:b019:9ae2:a2fe:e6b4/64 scope global dynamic noprefixroute - valid_lft 3588sec preferred_lft 3588sec - inet6 fe80::1f98:d09f:9309:9e77/64 scope link noprefixroute + inet6 fe80::5054:ff:fe12:3456/64 scope link valid_lft forever preferred_lft forever .fi .RE diff --git a/qrap.1 b/qrap.1 deleted file mode 100644 index 98cc877..0000000 --- a/qrap.1 +++ /dev/null @@ -1,77 +0,0 @@ -.\" SPDX-License-Identifier: GPL-2.0-or-later -.\" Copyright (c) 2020-2021 Red Hat GmbH -.\" Author: Stefano Brivio <sbrivio(a)redhat.com> -.TH qrap 1 - -.SH NAME -.B qrap -\- temporary qemu wrapper connecting UNIX domain socket to file descriptor - -.SH SYNOPSIS -.B qrap -[\fIFDNUM\fR \fIQEMU_CMD\fR] [\fIQEMU_ARG\fR]... -.SH DESCRIPTION -\fBqrap\fR is a wrapper, designed specifically for usage with \fBqemu\fR(1) and -\fBpasst\fR(1), connecting a UNIX domain socket to a file descriptor, and -running \fBqemu\fR(1) with given arguments. - -A running instance of \fBpasst\fR(1) is probed checking for UNIX domain sockets, -answering a dummy ARP request, with names starting from -\fI/tmp/passt_1.socket\fR up to \fI/tmp/passt_64.socket\fR. - -If first and second arguments are not a socket number and a command, that is, -respectively, \fIFDNUM\fR and \fIQEMU_CMD\fR, \fBqrap\fR will try to locate a -qemu executable in \fBPATH\fR and patch the command line to specify a network -device and a \fInetdev\fR back-end for usage with \fBpasst\fR(1). - -If \fBqrap\fR patches the command line, it will remove any potentially -conflicting network device, that is, any \fI-netdev\fR or \fI-net\fR option, or -any \fI-device\fR option with argument starting with \fIvirtio-net-pci,\fR, -\fIvirtio-net-ccw,\fR, \fIe1000,\fR, \fIe1000e,\fR, or \fIrtl8139,\fR. - -.SH EXAMPLES - -.BR "Start qemu explicitly passing its command name and a socket number" -.RS -.nf -$ ./qrap 5 kvm test.qcow2 -m 1024 -display none -nodefaults -nographic -net socket,fd=5 -net nic,model=virtio -.fi -.RE - -.BR "Start qemu without command name and socket number, patching arguments" -.RS -.nf -$ ./qrap test.qcow2 -m 1024 -display none -nodefaults -nographic -.fi -.RE - -.SH NOTES - -This wrapper is temporary: qemu commit 13c6be96618c ("net: stream: add unix -socket") introduces native AF_UNIX socket support, and it should be included in -qemu starting from the 7.2 release. It will be around for a little bit longer to -give users enough time to switch. - -.SH AUTHOR - -Stefano Brivio <sbrivio(a)redhat.com> - -.SH REPORTING BUGS - -No public bug tracker is available at this time. For the moment being, report -issues to Stefano Brivio <sbrivio(a)redhat.com>. - -.SH COPYRIGHT - -Copyright (c) 2020-2021 Red Hat GmbH. - -\fBqrap\fR is free software: you can redistribute is and/or modify it under the -terms of the GNU General Public License as published by the Free Software -Foundation, either version 2 of the License, or (at your option) any later -version. - -.SH SEE ALSO - -\fBqemu\fR(1), \fBpasst\fR(1). - -High-level documentation is available at https://passt.top/passt/about/. diff --git a/qrap.c b/qrap.c deleted file mode 100644 index d596706..0000000 --- a/qrap.c +++ /dev/null @@ -1,429 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0-or-later - -/* PASST - Plug A Simple Socket Transport - * - * qrap.c - qemu wrapper connecting UNIX domain socket to file descriptor - * - * Copyright (c) 2020-2021 Red Hat GmbH - * Author: Stefano Brivio <sbrivio(a)redhat.com> - * - * TODO: Drop this implementation once qemu commit 13c6be96618c ("net: stream: - * add unix socket") is included in a release (7.2), and once we can reasonably - * assume existing users switched to it. - */ - -#include <stdio.h> -#include <stdint.h> -#include <stdlib.h> -#include <string.h> -#include <unistd.h> -#include <sys/types.h> -#include <sys/socket.h> -#include <errno.h> -#include <linux/limits.h> -#include <limits.h> -#include <fcntl.h> -#include <net/if_arp.h> -#include <netinet/in.h> -#include <netinet/ip6.h> -#include <netinet/if_ether.h> -#include <time.h> - -#include <linux/icmpv6.h> - -#include "util.h" -#include "ip.h" -#include "passt.h" -#include "arp.h" - -static char *qemu_names[] = { - "kvm", - "qemu-kvm", -#ifdef ARCH - ( "qemu-system-" ARCH ), -#endif - "/usr/libexec/qemu-kvm", - NULL, -}; - -/** - * struct drop_arg - Drop matching arguments on command line - * @name: Option name - * @val: Substring in option value, NULL matches any value - */ -static const struct drop_arg { - char *name; - char *val; -} drop_args[] = { - { "-netdev", NULL }, - { "-net", NULL }, - { "-device", "virtio-net-pci," }, - { "-device", "{\"driver\":\"virtio-net-pci\"," }, - { "-device", "virtio-net-ccw," }, - { "-device", "{\"driver\":\"virtio-net-ccw\"," }, - { "-device", "e1000," }, - { "-device", "{\"driver\":\"e1000\"," }, - { "-device", "e1000e," }, - { "-device", "{\"driver\":\"e1000e\"," }, - { "-device", "rtl8139," }, - { "-device", "{\"driver\":\"rtl8139\"," }, - { 0 }, -}; - -/** - * struct pci_dev - PCI devices to add on command line depending on machine name - * @mach: Machine name - * @name: Device ("-device") name to insert - * @template: Prefix for device specification (first part of address) - * @template_post: Suffix for device specification (last part of address) - * @template_json: Device prefix for when JSON is used - * @template_json_post: Device suffix for when JSON is used - * @base: Base used for PCI addresses - * @first: First usable PCI address - * @last: Last usable PCI address - */ -static const struct pci_dev { - char *mach; - char *name; - char *template; - char *template_post; - char *template_json; - char *template_json_post; - int base; - int first; - int last; -} pci_devs[] = { - { - "pc-q35", "virtio-net-pci", - "bus=pci.", ",addr=0x0", - "\"bus\":\"pci.", ",\"addr\":\"0x0\"", - 10, 3, /* 2: hotplug bus */ 31 - }, - { - "pc-", "virtio-net-pci", - "bus=pci.0,addr=0x", "", - "\"bus\":\"pci.0\",\"addr\":\"0x", "", - 16, 2, /* 1: ISA bridge */ 31 - }, - { - "s390-ccw", "virtio-net-ccw", - "devno=fe.0.", "", - "\"devno\":\"fe.0.", "", - 16, 1, 16 - }, - { 0 }, -}; - -#define DEFAULT_FD 5 - -/** - * usage() - Print usage and exit - * @name: Executable name - */ -void usage(const char *name) -{ - fprintf(stderr, "Usage: %s [FDNUM QEMU_CMD] [QEMU_ARG]...\n", name); - fprintf(stderr, "\n"); - fprintf(stderr, "If first and second arguments aren't a socket number\n" - "and a path, %s will try to locate a qemu binary\n" - "and directly patch the command line\n", name); - - exit(EXIT_FAILURE); -} - -/** - * main() - Entry point and main loop - * @argc: Argument count - * @argv: File descriptor number, then qemu with arguments - * - * Return: 0 once interrupted, non-zero on failure - */ -int main(int argc, char **argv) -{ - int i, s, qemu_argc = 0, addr_map = 0, has_dev = 0, has_json = 0, retry_on_reset, rc; - struct timeval tv = { .tv_sec = 0, .tv_usec = (long)(500 * 1000) }; - char *qemu_argv[ARG_MAX], dev_str[ARG_MAX]; - struct sockaddr_un addr = { - .sun_family = AF_UNIX, - }; - const struct pci_dev *dev = NULL; - long fd; - struct { - uint32_t vnet_len4; - struct ethhdr eh4; - struct arphdr ah; - struct arpmsg am; - - uint32_t vnet_len6; - struct ethhdr eh6; - struct ipv6hdr ip6hr; - struct icmp6hdr ihr; - struct in6_addr target; - } __attribute__((__packed__)) probe = { - .vnet_len4 = htonl(42), - { - .h_dest = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }, - .h_source = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }, - .h_proto = htons(ETH_P_ARP), - }, - { .ar_hrd = htons(ARPHRD_ETHER), - .ar_pro = htons(ETH_P_IP), - .ar_hln = ETH_ALEN, - .ar_pln = 4, - .ar_op = htons(ARPOP_REQUEST), - }, - { - .sha = { 0 }, .sip = { 0 }, .tha = { 0 }, .tip = { 0 }, - }, - .vnet_len6 = htonl(78), - { - .h_dest = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }, - .h_source = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }, - .h_proto = htons(ETH_P_IPV6), - }, - { - .version = 6, - .payload_len = htons(24), - .nexthdr = IPPROTO_ICMPV6, - .hop_limit = 255, - .saddr = IN6ADDR_LOOPBACK_INIT, - .daddr = IN6ADDR_ANY_INIT, - }, - { - .icmp6_type = 135, - .icmp6_code = 0, - }, - IN6ADDR_ANY_INIT, - }; - char probe_r; - - if (argc >= 3) { - const char *path = getenv("PATH"); - errno = 0; - fd = strtol(argv[1], NULL, 0); - if (fd >= 3 && fd < INT_MAX && !errno && path) { - char env_path[ARG_MAX + 1], *p, command[ARG_MAX]; - - strncpy(env_path, path, ARG_MAX); - /* cppcheck-suppress strtokCalled */ - p = strtok(env_path, ":"); - while (p) { - snprintf(command, ARG_MAX, "%s/%s", p, argv[2]); - if (!access(command, X_OK)) - goto valid_args; - - /* cppcheck-suppress strtokCalled */ - p = strtok(NULL, ":"); - } - } - } - - fd = DEFAULT_FD; - - for (i = 1; i < argc - 1; i++) { - if (strcmp(argv[i], "-machine")) - continue; - - for (dev = pci_devs; dev->mach; dev++) { - if (strstr(argv[i + 1], dev->mach) == argv[i + 1]) - break; - } - } - - if (!dev || !dev->mach) - dev = pci_devs; - - for (qemu_argc = 1, i = 1; i < argc; i++) { - const struct drop_arg *a; - - for (a = drop_args; a->name; a++) { - if (!strcmp(argv[i], a->name)) { - if (!a->val) - break; - - if (i + 1 < argc && - strstr(argv[i + 1], a->val) == argv[i + 1]) - break; - } - } - if (a->name) { - i++; - continue; - } - - if (!strcmp(argv[i], "-device") && i + 1 < argc) { - const char *template = NULL; - const char *p; - - has_dev = 1; - - if ((p = strstr(argv[i + 1], dev->template))) { - template = dev->template; - } else if ((p = strstr(argv[i + 1], dev->template_json))) { - template = dev->template_json; - has_json = 1; - } - - if (template) { - long n; - - n = strtol(p + strlen(template), NULL, dev->base); - if (!errno) - addr_map |= (1 << n); - } - } - - qemu_argv[qemu_argc++] = argv[i]; - } - - for (i = dev->first; i < dev->last; i++) { - if (!(addr_map & (1 << i))) - break; - } - if (i == dev->last) { - fprintf(stderr, "Couldn't find free address for device\n"); - usage(argv[0]); - } - - if (has_dev) { - qemu_argv[qemu_argc++] = "-device"; - if (!has_json) { - if (dev->base == 16) { - snprintf(dev_str, ARG_MAX, - "%s,%s%x%s,netdev=hostnet0,x-txburst=4096", - dev->name, dev->template, i, dev->template_post); - } else if (dev->base == 10) { - snprintf(dev_str, ARG_MAX, - "%s,%s%d%s,netdev=hostnet0,x-txburst=4096", - dev->name, dev->template, i, dev->template_post); - } - } else { - if (dev->base == 16) { - snprintf(dev_str, ARG_MAX, - "{\"driver\":\"%s\",%s%x\"%s,\"netdev\":\"hostnet0\",\"x-txburst\":4096}", - dev->name, dev->template_json, i, dev->template_json_post); - } else if (dev->base == 10) { - snprintf(dev_str, ARG_MAX, - "{\"driver\":\"%s\",%s%d\"%s,\"netdev\":\"hostnet0\",\"x-txburst\":4096}", - dev->name, dev->template_json, i, dev->template_json_post); - } - } - qemu_argv[qemu_argc++] = dev_str; - } - - qemu_argv[qemu_argc++] = "-netdev"; - if (!has_json) { - qemu_argv[qemu_argc++] = "socket,fd=" STR(DEFAULT_FD) ",id=hostnet0"; - } else { - qemu_argv[qemu_argc++] = "{\"type\":\"socket\",\"fd\":\"" STR(DEFAULT_FD) "\",\"id\":\"hostnet0\"}"; - } - qemu_argv[qemu_argc] = NULL; - -valid_args: - for (i = 1; i < UNIX_SOCK_MAX; i++) { - retry_on_reset = 50; - -retry: - s = socket(AF_UNIX, SOCK_STREAM, 0); - if (s < 0) { - perror("socket"); - exit(EXIT_FAILURE); - } - - if (setsockopt(s, SOL_SOCKET, SO_RCVTIMEO, &tv, sizeof(tv))) - perror("setsockopt SO_RCVTIMEO"); - if (setsockopt(s, SOL_SOCKET, SO_SNDTIMEO, &tv, sizeof(tv))) - perror("setsockopt SO_SNDTIMEO"); - - snprintf(addr.sun_path, UNIX_PATH_MAX, UNIX_SOCK_PATH, i); - - errno = 0; - - if (connect(s, (const struct sockaddr *)&addr, sizeof(addr))) { - rc = errno; - perror("connect"); - } else if (send(s, &probe, sizeof(probe), 0) != sizeof(probe)) { - rc = errno; - perror("send"); - } else if (recv(s, &probe_r, 1, MSG_PEEK) <= 0) { - rc = errno; - perror("recv"); - } else { - break; - } - - /* FIXME: in a KubeVirt environment, libvirtd invokes qrap three - * times in a strict sequence when a virtual machine needs to - * be started, namely, when: - * - the domain XML is saved - * - the domain is started (for "probing") - * - the virtual machine is started for real - * and it often happens that the qemu process is still running - * when qrap is invoked again, so passt will refuse the new - * connection because the previous one is still active. This - * overlap seems to be anywhere between 0 and 3ms. - * - * If we get a connection reset, retry a few times, to allow for - * the previous qemu instance to terminate and, in turn, for the - * connection to passt to be closed. - * - * This should be fixed in libvirt instead. It probably makes - * sense to check this behaviour once native libvirt support is - * there, and this implies native qemu support too, so at that - * point qrap will have no reason to exist anymore -- that is, - * this FIXME will probably remain until the tool itself is - * obsoleted. - */ - if (retry_on_reset && rc == ECONNRESET) { - retry_on_reset--; - /* cppcheck-suppress usleepCalled */ - usleep(50 * 1000); - goto retry; - } - - fprintf(stderr, "Probe of %s failed\n", addr.sun_path); - - close(s); - } - - if (i == UNIX_SOCK_MAX) { - perror("connect"); - exit(EXIT_FAILURE); - } - - tv.tv_usec = 0; - if (setsockopt(s, SOL_SOCKET, SO_RCVTIMEO, &tv, sizeof(tv))) - perror("setsockopt, SO_RCVTIMEO reset"); - if (setsockopt(s, SOL_SOCKET, SO_SNDTIMEO, &tv, sizeof(tv))) - perror("setsockopt, SO_SNDTIMEO reset"); - - fprintf(stderr, "Connected to %s\n", addr.sun_path); - - if (dup2(s, (int)fd) < 0) { - perror("dup"); - exit(EXIT_FAILURE); - } - - close(s); - - if (qemu_argc) { - char **name; - - for (name = qemu_names; *name; name++) { - qemu_argv[0] = *name; - execvp(*name, qemu_argv); - if (errno != ENOENT) { - perror("execvp"); - usage(argv[0]); - } - } - if (errno == ENOENT) - fprintf(stderr, "Couldn't find qemu command\n"); - } else { - execvp(argv[2], argv + 2); - } - - perror("execvp"); - - return EXIT_FAILURE; -} diff --git a/tap.c b/tap.c index cf87591..dfa66c7 100644 --- a/tap.c +++ b/tap.c @@ -1377,8 +1377,6 @@ static void tap_backend_show_hints(const struct ctx *c) info("\nYou can now start qemu (>= 7.2, with commit 13c6be96618c):"); info(" kvm ... -device virtio-net-pci,netdev=s -netdev stream,id=s,server=off,addr.type=unix,addr.path=%s", c->sock_path); - info("or qrap, for earlier qemu versions:"); - info(" ./qrap 5 kvm ... -net socket,fd=5 -net nic,model=virtio"); break; case MODE_VU: info("You can start qemu with:"); diff --git a/test/distro/opensuse b/test/distro/opensuse index eab722b..9ceb0bd 100644 --- a/test/distro/opensuse +++ b/test/distro/opensuse @@ -39,7 +39,7 @@ hostb ./passt -s __STATEDIR__/passt.socket -P __PIDFILE__ & sleep 1 host echo hout DNS6 sed -n 's/^nameserver $[^:]*:$$[^%]*$.*/\1\2/p' /etc/resolv.conf | head -1 -hout GUEST_FILES ls -1 *.c *.h *.sh passt.1 qrap.1 Makefile README.md | tr '\n' ' '; echo +hout GUEST_FILES ls -1 *.c *.h *.sh passt.1 Makefile README.md | tr '\n' ' '; echo test OpenSUSE Leap 15.1 -- 2.43.0

2 1

[PATCH] tap: Fix EAGAIN/EWOULDBLOCK check in tap_pasta_input()
by Hayato Kiwata 17 Jul '26

17 Jul '26

On Linux, EAGAIN and EWOULDBLOCK have the same value. Commit d2a1dc744b10 ("tap: Restructure in tap_pasta_input()") added a check for EWOULDBLOCK on purpose: > - Check for EWOULDBLOCK as well as EAGAIN for the benefit of any future > ports where those might not have the same value However, the condition uses && instead of ||, so it can never be true on a platform where EAGAIN and EWOULDBLOCK are different. Use || instead so that the check works as intended on such platforms. Fixes: d2a1dc744b10 ("tap: Restructure in tap_pasta_input()") Signed-off-by: Hayato Kiwata <dev(a)haytok.jp> --- tap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tap.c b/tap.c index d418961..cf87591 100644 --- a/tap.c +++ b/tap.c @@ -1329,7 +1329,7 @@ static void tap_pasta_input(struct ctx *c, const struct timespec *now) continue; } - if (errno == EAGAIN && errno == EWOULDBLOCK) + if (errno == EAGAIN || errno == EWOULDBLOCK) break; /* all done for now */ die("Error on tap device, exiting"); -- 2.39.5 (Apple Git-154)

3 2

[PATCH] pasta: Add --pass-fds option to preserve specific file descriptors
by Richard Lawrence 17 Jul '26

17 Jul '26

Introduce a new long-only command line option '--pass-fds' for pasta mode. This option accepts a comma-separated list of file descriptor numbers. It parses, validates, sorts, and deduplicates these file descriptors, and then skips closing them by calling close_range() on the gaps between them inside close_open_files(). The option is documented in the usage help and the passt.1 man page. Co-authored-by: rarensu <40872193+rarensu(a)users.noreply.github.com> Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot](a)users.noreply.github.com> Signed-off-by: Richard Lawrence <rlawrence(a)tamu.edu> --- conf.c | 9 +++++- passt.1 | 5 ++++ util.c | 93 ++++++++++++++++++++++++++++++++++++++++++++++++++------- 3 files changed, 96 insertions(+), 11 deletions(-) diff --git a/conf.c b/conf.c index 5b6cc2b..0032d02 100644 --- a/conf.c +++ b/conf.c @@ -732,7 +732,9 @@ pasta_opts: " Don't copy all addresses to namespace\n" " --ns-mac-addr ADDR Set MAC address on tap interface\n" " --no-splice Disable inbound socket splicing\n" - " --splice-only Only enable loopback forwarding\n"); + " --splice-only Only enable loopback forwarding\n" + " --pass-fds FDS Comma-separated list of fds to pass to\n" + " the spawned command\n"); passt_exit(status); } @@ -1316,6 +1318,7 @@ void conf(struct ctx *c, int argc, char **argv) {"stats", required_argument, NULL, 31 }, {"conf-path", required_argument, NULL, 'c' }, {"chroot-fallback", no_argument, NULL, 32 }, + {"pass-fds", required_argument, NULL, 33 }, { 0 }, }; const char *optstring = "+dqfel:hs:c:F:I:p:P:m:a:n:M:g:i:o:D:S:H:461t:u:T:U:"; @@ -1557,6 +1560,10 @@ void conf(struct ctx *c, int argc, char **argv) case 32: c->chroot_fallback = true; break; + case 33: + if (c->mode != MODE_PASTA) + die("--pass-fds is for pasta mode only"); + break; case 'd': c->debug = 1; c->quiet = 0; diff --git a/passt.1 b/passt.1 index 1570f6a..93c9b24 100644 --- a/passt.1 +++ b/passt.1 @@ -755,6 +755,11 @@ of local traffic in pasta\fR in the \fBNOTES\fR for more details. Do not create a tap device in the namespace. In this mode, \fIpasta\fR only forwards loopback traffic between namespaces. +.TP +.BR \-\-pass\-fds " " \fIfds\fR +Pass a comma-separated list of file descriptors to the spawned command. +These file descriptors will be kept open. + .SH EXAMPLES .SS \fBpasta diff --git a/util.c b/util.c index 4bc5d6f..c73fd74 100644 --- a/util.c +++ b/util.c @@ -932,29 +932,102 @@ const char *str_ee_origin(const struct sock_extended_err *ee) void close_open_files(int argc, char **argv) { const struct option optfd[] = { { "fd", required_argument, NULL, 'F' }, + { "pass-fds", required_argument, NULL, 33 }, { 0 }, }; long fd = -1; - int name, rc; + const char *pass_fds_arg = NULL; + int name, rc = 0; + int old_opterr = opterr; + opterr = 0; do { name = getopt_long(argc, argv, "-:F:", optfd, NULL); if (name == 'F') fd = conf_tap_fd(optarg); + else if (name == 33) + pass_fds_arg = optarg; } while (name != -1); + opterr = old_opterr; - if (fd == -1) { - rc = close_range(STDERR_FILENO + 1, ~0U, CLOSE_RANGE_UNSHARE); - } else if (fd == STDERR_FILENO + 1) { /* Still a single range */ - rc = close_range(STDERR_FILENO + 2, ~0U, CLOSE_RANGE_UNSHARE); - } else { - rc = close_range(STDERR_FILENO + 1, fd - 1, - CLOSE_RANGE_UNSHARE); - if (!rc) - rc = close_range(fd + 1, ~0U, CLOSE_RANGE_UNSHARE); + #define MAX_PASS_FDS 1024 + unsigned int keep_fds[MAX_PASS_FDS]; + int keep_fds_cnt = 0; + + if (fd != -1 && fd > STDERR_FILENO) { + keep_fds[keep_fds_cnt++] = (unsigned int)fd; + } + + if (pass_fds_arg) { + const char *p = pass_fds_arg; + while (*p) { + char *endptr; + unsigned long val = strtoul(p, &endptr, 10); + if (p == endptr) { + die("Invalid --pass-fds option: %s", pass_fds_arg); + } + if (val > INT_MAX || val <= STDERR_FILENO) { + die("Invalid file descriptor in --pass-fds: %lu", val); + } + if (keep_fds_cnt >= MAX_PASS_FDS) { + die("Too many file descriptors in --pass-fds"); + } + keep_fds[keep_fds_cnt++] = (unsigned int)val; + if (*endptr == ',') { + p = endptr + 1; + if (*p == '\0') { + die("Invalid --pass-fds option: trailing comma"); + } + } else if (*endptr == '\0') { + p = endptr; + } else { + die("Invalid character in --pass-fds option: %s", pass_fds_arg); + } + } } + /* Sort keep_fds in ascending order */ + for (int i = 1; i < keep_fds_cnt; i++) { + unsigned int key = keep_fds[i]; + int j = i - 1; + while (j >= 0 && keep_fds[j] > key) { + keep_fds[j + 1] = keep_fds[j]; + j--; + } + keep_fds[j + 1] = key; + } + + /* Remove duplicates */ + int write_idx = 0; + for (int i = 0; i < keep_fds_cnt; i++) { + if (i == 0 || keep_fds[i] != keep_fds[i - 1]) { + keep_fds[write_idx++] = keep_fds[i]; + } + } + keep_fds_cnt = write_idx; + + /* Close ranges */ + unsigned int current = STDERR_FILENO + 1; + for (int i = 0; i < keep_fds_cnt; i++) { + unsigned int f = keep_fds[i]; + if (current < f) { + int call_rc = close_range(current, f - 1, CLOSE_RANGE_UNSHARE); + if (call_rc != 0) + rc = call_rc; + } + current = f + 1; + } + if (keep_fds_cnt == 0 || keep_fds[keep_fds_cnt - 1] < ~0U) { + int call_rc = close_range(current, ~0U, CLOSE_RANGE_UNSHARE); + if (call_rc != 0) + rc = call_rc; + } + + /* Note: We assume that the parent process did not set FD_CLOEXEC on the + * passed file descriptors, so we do not explicitly clear it here. + */ + if (rc) { if (errno == ENOSYS || errno == EINVAL) { /* This probably means close_range() or the -- 2.52.0

3 2