linux-kernel - Re: [PATCH v5 06/10] Uprobes: Support SDT markers having reference count (semaphore)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20180628195106.GA3952@redhat.com>
Date:   Thu, 28 Jun 2018 21:51:06 +0200
From:   Oleg Nesterov <oleg@...hat.com>
To:     Ravi Bangoria <ravi.bangoria@...ux.ibm.com>
Cc:     srikar@...ux.vnet.ibm.com, rostedt@...dmis.org,
        mhiramat@...nel.org, peterz@...radead.org, mingo@...hat.com,
        acme@...nel.org, alexander.shishkin@...ux.intel.com,
        jolsa@...hat.com, namhyung@...nel.org,
        linux-kernel@...r.kernel.org, corbet@....net,
        linux-doc@...r.kernel.org, ananth@...ux.vnet.ibm.com,
        alexis.berlemont@...il.com, naveen.n.rao@...ux.vnet.ibm.com,
        linux-arm-kernel@...ts.infradead.org, linux-mips@...ux-mips.org,
        linux@...linux.org.uk, ralf@...ux-mips.org, paul.burton@...s.com
Subject: Re: [PATCH v5 06/10] Uprobes: Support SDT markers having reference
 count (semaphore)

I have to admit that after a quick glance I can't understand this patch
at all... I'll try to read it again tomorrow, but could you at least explain
how find_node_in_range/build_probe_list can work if off_type==REF_CTR_OFFSET?

On 06/28, Ravi Bangoria wrote:
>
> -find_node_in_range(struct inode *inode, loff_t min, loff_t max)
> +find_node_in_range(struct inode *inode, int off_type, loff_t min, loff_t max)
>  {
>  	struct rb_node *n = uprobes_tree.rb_node;
>  
>  	while (n) {
>  		struct uprobe *u = rb_entry(n, struct uprobe, rb_node);
> +		loff_t offset = uprobe_get_offset(u, off_type);
>  
>  		if (inode < u->inode) {
>  			n = n->rb_left;
>  		} else if (inode > u->inode) {
>  			n = n->rb_right;
>  		} else {
> -			if (max < u->offset)
> +			if (max < offset)
>  				n = n->rb_left;
> -			else if (min > u->offset)
> +			else if (min > offset)
>  				n = n->rb_right;
>  			else
>  				break;

To simplify, lets forget about uprobe->inode (which acts as a key too). So uprobes_tree
is a binary tree sorted by uprobe->offset key and that is why the binary search works.

But it is not sorted by uprobe->ref_ctr_offset. So for example n->rb_left can have the
n->ref_ctr_offset key that is greater than the n's ref_ctr_offset. So how we can use the
binary search if REF_CTR_OFFSET?

I must have missed something, I assume you tested this patch and it works somehow...

Oleg.