[DTrace-devel] [PATCH v2] Implement the pcap action

Eugene Loh eugene.loh at oracle.com
Tue Nov 7 21:20:58 UTC 2023


I assume Alan's comments were integrated.  Meanwhile...

On 11/7/23 01:59, Kris Van Hees via DTrace-devel wrote:
> diff --git a/libdtrace/dt_cg.c b/libdtrace/dt_cg.c
> @@ -1882,8 +1905,112 @@ dt_cg_act_panic(dt_pcb_t *pcb, dt_node_t *dnp, dtrace_actkind_t kind)
>   static void
>   dt_cg_act_pcap(dt_pcb_t *pcb, dt_node_t *dnp, dtrace_actkind_t kind)
>   {
> -	dnerror(dnp, D_UNKNOWN, "pcap() is not implemented (yet)\n");
> -	/* FIXME: Needs implementation */
> +	dtrace_hdl_t	*dtp = pcb->pcb_hdl;
> +	dt_regset_t	*drp = pcb->pcb_regs;
> +	dt_irlist_t	*dlp = &pcb->pcb_ir;
> +	dt_node_t	*addr = dnp->dn_args;
> +	dt_node_t	*proto = addr->dn_list;
> +	uint_t		time_off, size_off, data_off, off;
> +	uint_t		lbl_lenok = dt_irlist_label(dlp);
> +	uint64_t	pcapsz = dtp->dt_options[DTRACEOPT_PCAPSIZE];
> +	int		lenreg;
> +
> +	TRACE_REGSET("pcap(): Begin ");
> +
> +	/*
> +	 * Create records for timestamp, size, protocol id, and packet data.
> +	 * The protocol id is stored immediately from the passed in argument.
> +	 */
> +	time_off = dt_rec_add(dtp, dt_cg_fill_gap, DTRACEACT_PCAP,
> +			      sizeof(uint64_t), sizeof(uint64_t), NULL, 0);
> +	size_off = dt_rec_add(dtp, dt_cg_fill_gap, DTRACEACT_PCAP,
> +			      sizeof(uint32_t), sizeof(uint32_t), NULL, 0);
> +	dt_cg_store_val(pcb, proto, DTRACEACT_PCAP, NULL, 0);
> +	data_off = dt_rec_add(dtp, dt_cg_fill_gap, DTRACEACT_PCAP,
> +			      dtp->dt_options[DTRACEOPT_PCAPSIZE], 1, NULL, 0);

Use pcapsz in place of dtp->dt_options[DTRACEOPT_PCAPSIZE] here.

> +
> +	/* Store timestamp (ns since boot time). */
> +	if (dt_regset_xalloc_args(drp) == -1)
> +		longjmp(yypcb->pcb_jmpbuf, EDT_NOREG);
> +	dt_regset_xalloc(drp, BPF_REG_0);
> +	emit(dlp,  BPF_CALL_HELPER(BPF_FUNC_ktime_get_ns));
> +	dt_regset_free_args(drp);
> +	emit(dlp,  BPF_STORE(BPF_DW, BPF_REG_9, time_off, BPF_REG_0));
> +	dt_regset_free(drp, BPF_REG_0);
> +
> +	/*
> +	 * Determine and store the packet data size.
> +	 * The size of the data to be copied (and reported as size of the
> +	 * packet capture) is the lesser of the linear data size
> +	 * (skb->len - skb->data_len) and the capture size.
> +	 */
> +	dt_cg_node(addr, dlp, drp);
> +
> +	off = dt_cg_ctf_offsetof("struct sk_buff", "len", NULL);
> +
> +	if (dt_regset_xalloc_args(drp) == -1)
> +		longjmp(yypcb->pcb_jmpbuf, EDT_NOREG);
> +
> +	emit(dlp,  BPF_MOV_REG(BPF_REG_3, addr->dn_reg));
> +	emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_3, off));
> +	emit(dlp,  BPF_MOV_IMM(BPF_REG_2, sizeof(uint32_t)));
> +	emit(dlp,  BPF_LOAD(BPF_DW, BPF_REG_1, BPF_REG_FP, DT_STK_SP));
> +	dt_regset_xalloc(drp, BPF_REG_0);
> +	emit(dlp,  BPF_CALL_HELPER(dtp->dt_bpfhelper[BPF_FUNC_probe_read_kernel]));
> +	dt_regset_free(drp, BPF_REG_0);
> +	dt_regset_free_args(drp);

Shouldn't make any difference, but we normally free_args immediately 
after the function call.  That makes sense since that's when the BPF 
verifier invalidates those regs.  And I thought that's what was done in 
v1.  So why have free(%r0) and free_args() been exchanged?

> +
> +	if ((lenreg = dt_regset_alloc(drp)) == -1)
> +		longjmp(yypcb->pcb_jmpbuf, EDT_NOREG);
> +
> +	emit(dlp,  BPF_LOAD(BPF_DW, lenreg, BPF_REG_FP, DT_STK_SP));
> +	emit(dlp,  BPF_LOAD(BPF_W, lenreg, lenreg, 0));
> +
> +	off = dt_cg_ctf_offsetof("struct sk_buff", "data_len", NULL);
> +
> +	if (dt_regset_xalloc_args(drp) == -1)
> +		longjmp(yypcb->pcb_jmpbuf, EDT_NOREG);
> +
> +	emit(dlp,  BPF_MOV_REG(BPF_REG_3, addr->dn_reg));
> +	emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_3, off));
> +	emit(dlp,  BPF_MOV_IMM(BPF_REG_2, sizeof(uint32_t)));
> +	emit(dlp,  BPF_LOAD(BPF_DW, BPF_REG_1, BPF_REG_FP, DT_STK_SP));
> +	dt_regset_xalloc(drp, BPF_REG_0);
> +	emit(dlp,  BPF_CALL_HELPER(dtp->dt_bpfhelper[BPF_FUNC_probe_read_kernel]));
> +	emit(dlp,  BPF_LOAD(BPF_DW, BPF_REG_0, BPF_REG_FP, DT_STK_SP));
> +	emit(dlp,  BPF_LOAD(BPF_W, BPF_REG_0, BPF_REG_0, 0));
> +	dt_regset_free(drp, BPF_REG_0);
> +	dt_regset_free_args(drp);

Same thing about free_args() should be right after the function call.

Also, here you put some result in %r0 and then free(%r0).  Typically 
you'd get lucky but the whole reason for the xalloc/free is in case %r0 
is spilled/filled.  If such a thing is really possible, then this code 
will be overwriting %r0 (filling the reg) before the %r0 result is used.

> +
> +	emit(dlp,  BPF_ALU64_REG(BPF_SUB, lenreg, BPF_REG_0));
> +	emit(dlp,  BPF_BRANCH_IMM(BPF_JLE, lenreg, pcapsz, lbl_lenok));
> +	emit(dlp,  BPF_MOV_IMM(lenreg, pcapsz));
> +	emitl(dlp, lbl_lenok,
> +		   BPF_STORE(BPF_W, BPF_REG_9, size_off, lenreg));
> +
> +	/* Copy the packet data to the output buffer. */
> +	off = dt_cg_ctf_offsetof("struct sk_buff", "data", NULL);
> +
> +	if (dt_regset_xalloc_args(drp) == -1)
> +		longjmp(yypcb->pcb_jmpbuf, EDT_NOREG);
> +	emit(dlp,  BPF_MOV_REG(BPF_REG_3, addr->dn_reg));
> +	dt_regset_free(drp, addr->dn_reg);
> +	emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_3, off));
> +	emit(dlp,  BPF_MOV_IMM(BPF_REG_2, sizeof(uint64_t)));
> +	emit(dlp,  BPF_LOAD(BPF_DW, BPF_REG_1, BPF_REG_FP, DT_STK_SP));
> +	dt_regset_xalloc(drp, BPF_REG_0);
> +	emit(dlp,  BPF_CALL_HELPER(dtp->dt_bpfhelper[BPF_FUNC_probe_read_kernel]));
> +	emit(dlp,  BPF_LOAD(BPF_DW, BPF_REG_3, BPF_REG_FP, DT_STK_SP));
> +	emit(dlp,  BPF_LOAD(BPF_DW, BPF_REG_3, BPF_REG_3, 0));
> +	emit(dlp,  BPF_MOV_REG(BPF_REG_2, lenreg));

I already commented on this:  it's a problem if lenreg is %r1 %r2 %r4 or 
%r5:  the BPF verifier will complain since those regs were invalidated 
with the last func call.  And if lenreg is %r3, the verifier will not 
complain, but the value one is getting is the value that was just 
written into %r3.

I think the thing to do is to free_args() immediately after the function 
call, then xalloc_args(), then fill the regs starting with %r2 first.

> +	dt_regset_free(drp, lenreg);
> +	emit(dlp,  BPF_MOV_REG(BPF_REG_1, BPF_REG_9));
> +	emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, data_off));
> +	emit(dlp,  BPF_CALL_HELPER(dtp->dt_bpfhelper[BPF_FUNC_probe_read_kernel]));
> +	dt_regset_free_args(drp);
> +	dt_regset_free(drp, BPF_REG_0);
> +
> +	TRACE_REGSET("pcap(): End   ");
>   }
>   
> diff --git a/libdtrace/dt_consume.c b/libdtrace/dt_consume.c
> @@ -1427,28 +1427,26 @@ dt_print_mod(dtrace_hdl_t *dtp, FILE *fp, const char *format, caddr_t addr)
>   }
>   
>   int
> -dt_print_pcap(dtrace_hdl_t *dtp, FILE *fp, dtrace_recdesc_t *rec,
> +dt_print_pcap(dtrace_hdl_t *dtp, FILE *fp, dtrace_recdesc_t *rec, uint_t nrecs,
>   	      const caddr_t buf)
>   {
> -	caddr_t	addr;
> -	uint64_t time, proto, pktlen, maxlen;
> -	const char *filename;
> +	caddr_t			data;
> +	uint64_t		time, proto, pktlen;
> +	uint64_t		maxlen = dtp->dt_options[DTRACEOPT_PCAPSIZE];
> +	const char		*filename;
>   
> -	addr = (caddr_t)buf + rec->dtrd_offset;
> +	if (nrecs < 4)
> +		return dt_set_errno(dtp, EDT_PCAP);
>   
>   	if (dt_read_scalar(buf, rec, &time) < 0 ||
> -	    dt_read_scalar(buf + sizeof(uint64_t), rec, &pktlen) < 0)
> +	    dt_read_scalar(buf, rec + 1, &pktlen) < 0)
>   		return dt_set_errno(dtp, EDT_PCAP);
>   
> -	if (pktlen == 0) {
> -		/*
> -		 * skb must have been NULL, skip without capturing.
> -		 */
> +	/* If pktlen is 0, the skb must have been ULL, so don't capture it. */

ULL => NULL

> +	if (pktlen == 0)
>   		return 0;
> -	}
> -	maxlen = DT_PCAPSIZE(dtp->dt_options[DTRACEOPT_PCAPSIZE]);


So get rid of DT_PCAPSIZE in pcap.h.

>   
> -	if (dt_read_scalar(buf, rec + 1, &proto) < 0)
> +	if (dt_read_scalar(buf, rec + 2, &proto) < 0)
>   		return dt_set_errno(dtp, EDT_PCAP);



More information about the DTrace-devel mailing list