Opened 4 weeks ago

Last modified 2 weeks ago

#33093 merge_ready defect

Use IF_BUG_ONCE in buf_flush_to_tls()

Reported by: nickm Owned by: nickm
Priority: Medium Milestone: Tor: 0.4.2.x-final
Component: Core Tor/Tor Version:
Severity: Normal Keywords: assert, tor-connection security-low consider-backport-after-0433 035-backport 041-backport 042-backport
Cc: Actual Points: .1
Parent ID: Points:
Reviewer: ahf Sponsor:

Description

See parent: the BUG() messages in this function caused dgoulet to run out of disk.

Child Tickets

Change History (8)

comment:1 Changed 4 weeks ago by nickm

Status: assignedneeds_review

Branch is bug33093_logging_035 with PR at https://github.com/torproject/tor/pull/1693 . Simple fix, merges cleanly to master.

comment:2 Changed 4 weeks ago by teor

Keywords: security-low consider-backport-after-0433 added

comment:3 Changed 4 weeks ago by ahf

Reviewer: ahf
Status: needs_reviewmerge_ready

Looks good.

comment:4 Changed 4 weeks ago by nickm

Milestone: Tor: 0.4.3.x-finalTor: 0.4.2.x-final

Merging to master; marking for possible backport. I think that this is a good candidate for backporting, since it is unlikely to actually break anything.

comment:5 Changed 3 weeks ago by dgoulet

Parent ID: #32472

Merged to master. Unparenting in order to close the parent that was waiting on this fix to call it "good enough".

comment:6 Changed 3 weeks ago by teor

Keywords: 043-must removed

comment:7 in reply to:  description ; Changed 2 weeks ago by cypherpunks

Replying to nickm:

See parent: the BUG() messages in this function caused dgoulet to run out of disk.

Doesn't #32472 say the opposite happened? (Running out of disk space caused the BUG().)

comment:8 in reply to:  7 Changed 2 weeks ago by teor

Replying to cypherpunks:

Replying to nickm:

See parent: the BUG() messages in this function caused dgoulet to run out of disk.

Doesn't #32472 say the opposite happened? (Running out of disk space caused the BUG().)

The bug report is a bit unclear. Some people learned English later in life, or learned a different dialect of English. So they write sentences in a different order.

But it's really unlikely that:

  • a disk full error stops tor writing TLS to the network,
  • then tor logs a lot of log lines,
  • which end up being written to (a full) disk.

It's far more likely that a TLS error (or tor bug) has caused large quantities of logs, and they filled up the disk.

Note: See TracTickets for help on using tickets.