Netdev List
 help / color / mirror / Atom feed
From: Jakub Kicinski <kuba@kernel.org>
To: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Rajat Gupta <rajat.gupta@oss.qualcomm.com>,
	netdev@vger.kernel.org, davem@davemloft.net, edumazet@google.com,
	pabeni@redhat.com, horms@kernel.org, jiri@resnulli.us,
	yimingqian591@gmail.com, keenanat2000@gmail.com,
	2045gemini@gmail.com, rollkingzzc@gmail.com
Subject: Re: [PATCH net] net/sched: fix pedit partial COW leading to page cache corruption
Date: Fri, 22 May 2026 17:55:07 -0700	[thread overview]
Message-ID: <20260522175507.02b4fe83@kernel.org> (raw)
In-Reply-To: <CAM0EoM=Db_AD1tqcG5w18pJKh8ysbUUHyzgP4c3thHTmeXLVyw@mail.gmail.com>

On Fri, 22 May 2026 13:01:49 -0400 Jamal Hadi Salim wrote:
> > > I must be missing something. What's the problem with changing the patch
> > > to pull headers instead? I mean - if we agree that this is where we'll
> > > end up - we should just do it now. It's the long standing kernel policy
> > > to "fix things right" instead of creating temporary fixes which then
> > > have to be reworked in -next  
> >
> > I may be the one missing things. You main concern is with this:
> > +       /*
> > +        * If the skb has shared frags the user is likely using zero-copy
> > +        * (e.g. sendfile).  Those page frags may point to page-cache pages;
> > +        * writing into them would silently corrupt the page cache.
> > +        * Linearize so pedit operates on a private copy.
> > +        * TL;DR: if you want zero-copy, don't use pedit.
> > +        */
> > +       if (skb_has_shared_frag(skb)) {
> > +               if (__skb_linearize(skb))
> > +                       goto bad;
> > +       }
> > +
> >
> > i.e you want that gone, correct?
> > And my concern was whether removing this still exposes things to
> > exploit even if temporarily.
> > Likely it wont. I have time, let me test the exploit with that code
> > ifdef'ed out.  
> 
> Ok, it fixes the issue even i ifdef that out.
> We still need the little patchlet i sent ...

More and more I feel like I'm completely missing the plot but 
for the portion of the problem that's "we are writing to frags"
the fix I was trying to describe is:

diff --git a/net/sched/act_pedit.c b/net/sched/act_pedit.c
index bc20f08a2789..3a74cef58e17 100644
--- a/net/sched/act_pedit.c
+++ b/net/sched/act_pedit.c
@@ -414,7 +414,7 @@ TC_INDIRECT_SCOPE int tcf_pedit_act(struct sk_buff *skb,
 	for (i = parms->tcfp_nkeys; i > 0; i--, tkey++) {
 		int offset = tkey->off;
 		int hoffset = 0;
-		u32 *ptr, hdata;
+		u32 *ptr;
 		u32 val;
 		int rc;
 
@@ -456,10 +456,9 @@ TC_INDIRECT_SCOPE int tcf_pedit_act(struct sk_buff *skb,
 			goto bad;
 		}
 
-		ptr = skb_header_pointer(skb, hoffset + offset,
-					 sizeof(hdata), &hdata);
-		if (!ptr)
+		if (!pskb_may_pull(skb, hoffset + offset + sizeof(*ptr)))
 			goto bad;
+		ptr = (u32 *)(skb->data + hoffset + offset);
 		/* just do it, baby */
 		switch (cmd) {
 		case TCA_PEDIT_KEY_EX_CMD_SET:
@@ -474,8 +473,6 @@ TC_INDIRECT_SCOPE int tcf_pedit_act(struct sk_buff *skb,
 		}
 
 		*ptr = ((*ptr & tkey->mask) ^ val);
-		if (ptr == &hdata)
-			skb_store_bits(skb, hoffset + offset, ptr, 4);
 	}
 
 	goto done;

  reply	other threads:[~2026-05-23  0:55 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-18  1:30 [PATCH net] net/sched: fix pedit partial COW leading to page cache Rajat Gupta
2026-05-18 13:10 ` Han Guidong
2026-05-18 13:31 ` Jamal Hadi Salim
2026-05-19  3:39   ` [PATCH net] net/sched: fix pedit partial COW leading to page cache corruption Rajat Gupta
2026-05-19 11:18     ` Toke Høiland-Jørgensen
2026-05-19 15:10     ` Han Guidong
2026-05-20  9:12       ` Jamal Hadi Salim
2026-05-20 10:04         ` Han Guidong
2026-05-20 10:36         ` Han Guidong
2026-05-20 11:40           ` Jamal Hadi Salim
2026-05-20  9:23     ` Jamal Hadi Salim
2026-05-20 20:00       ` Jamal Hadi Salim
2026-05-21  9:53         ` Jamal Hadi Salim
2026-05-21 10:15           ` Jamal Hadi Salim
2026-05-21 14:35             ` Jakub Kicinski
2026-05-21 15:16               ` Jamal Hadi Salim
2026-05-21 15:46                 ` Jakub Kicinski
2026-05-22 11:47                   ` Jamal Hadi Salim
2026-05-22 15:46                     ` Jakub Kicinski
2026-05-22 16:37                       ` Jamal Hadi Salim
2026-05-22 17:01                         ` Jamal Hadi Salim
2026-05-23  0:55                           ` Jakub Kicinski [this message]
2026-05-23 12:07                             ` Jamal Hadi Salim
2026-05-23 12:13                               ` Jamal Hadi Salim
2026-05-23 16:46                                 ` Jakub Kicinski
2026-05-23 16:57                                   ` Jamal Hadi Salim
     [not found]               ` <CAKa-r6soz=iMBiYG0Grhhc12yhdw9vMNV+XjjEPCmtgKK6+rhA@mail.gmail.com>
2026-05-21 15:56                 ` Jakub Kicinski
2026-05-22 11:49                 ` Jamal Hadi Salim
2026-05-22 12:00                   ` Toke Høiland-Jørgensen
2026-05-22 14:49                   ` Davide Caratti
2026-05-22  7:49             ` Han Guidong

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260522175507.02b4fe83@kernel.org \
    --to=kuba@kernel.org \
    --cc=2045gemini@gmail.com \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=horms@kernel.org \
    --cc=jhs@mojatatu.com \
    --cc=jiri@resnulli.us \
    --cc=keenanat2000@gmail.com \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=rajat.gupta@oss.qualcomm.com \
    --cc=rollkingzzc@gmail.com \
    --cc=yimingqian591@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox