cupsd crashes in _ppdCreateFromIPP2() when a printer reports a media size with no PPD name (e.g. jpn_chou2)
I maintainer di solito rispondono entro 1 giorno
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 2/5
- Tempo stimato
- 1-3 ore
- Idoneità per principianti
- 78/100
Direzione di ricerca
Tutti e otto gli usi di NULL sono in _ppdCreateFromIPP2() in cups/ppd-cache.c: i due valori predefiniti ppdname e le chiamate copy/snprintf("%s.Borderless") nei cicli media-col-database, media-size-supported e media-supported; il writer *PageSize più sotto protegge già con && pwg->ppd, quindi usalo come modello. pwg_ppdize_name() e la voce incriminata _PWG_MEDIA_MM("jpn_chou2_111.1x146mm", NULL, NULL, ...) stanno in cups/pwg-media.c. Riproduci fuori dal dispositivo passando l'attr.txt allegato attraverso ippReadIO() e _ppdCreateFromIPP(); fatto significa nessun segfault e il PPD elenca tutte le 20 dimensioni, inclusa Jpn_chou2_111.1x146mm.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
Disclaimer:
Found with Claude while debugging my printing software; I reproduced it and checked the analysis myself. Explanation below was written by Claude, too.
Describe the bug
cupsd segfaults while creating a temporary queue for a DNS-SD discovered driverless printer when one of the printer's media sizes maps to a pwg_media_t entry that has no PPD name. _ppdCreateFromIPP2() (cups/ppd-cache.c) copies pwg->ppd with strlcpy() / snprintf("%s") (cupsCopyString() on master) without checking it for NULL.
In our case the printer reports a 111×146 mm size (custom_111x146mm_111x146mm in media-supported, x-dimension 11100 / y-dimension 14600 in media-col-database and media-size-supported). pwgMediaForSize() matches that, within its tolerance, to jpn_chou2_111.1x146mm, which is defined in cups/pwg-media.c without a PPD name: _PWG_MEDIA_MM("jpn_chou2_111.1x146mm", NULL, NULL, 111.1, 146).
There are eight such uses in _ppdCreateFromIPP2(): the two ppdname defaults (media-col-default / media-default) and the copy / snprintf("%s.Borderless") in the media-col-database, media-size-supported and media-supported loops. The *PageSize writer further down in the same file already checks && pwg->ppd, so elsewhere the code already treats this field as possibly NULL.
The same code is still there in 2.4.20, on the current 2.4.x branch and on master.
To Reproduce
- On a Mac, share an HP Deskjet F2400 (USB) over the network. The shared queue reports the 111×146 mm size described above (full attributes attached as
attr.txt). - On a Linux machine in the same network running cupsd with Avahi, the printer shows up as a temporary queue in
lpstat -l -e(network). - Run
lpoptions -p HP_Deskjet_F2400_series_Mac -l(or anything that callscupsCopyDestInfo()on it). That sends CUPS-Create-Local-Printer, and cupsd builds a PPD from the printer's attributes. - cupsd crashes with SIGSEGV. It is restarted, and the next query crashes it again.
Backtrace from the cupsd core dump (2.4.15, aarch64, cupsd without symbols):
#0 ?? () from libc.so.6
#1 strlcpy () from libc.so.6
#2 _ppdCreateFromIPP2 () from libcups.so.2
#3 ?? () (cupsd)
Same crash in an off-device repro: read the printer's Get-Printer-Attributes response with ippReadIO() and pass it to _ppdCreateFromIPP(). Against a 2.4.15 libcups with debug info:
#1 _cups_strlcpy (dst=..., src=0x0, size=...)
#3 _ppdCreateFromIPP2 (...) at ppd-cache.c:3721
Line 3721 is the strlcpy(temp.media, pwg->ppd, ...) in the media-col-database loop. The repro crashes on 2.4.15 and 2.4.20.
Expected behavior
cupsd creates the temporary queue. Sizes whose PWG entry has no PPD name get one made from the PWG name, or are skipped, instead of crashing.
A fix that works for us: fall back to a name built from the PWG name.
static const char *
pwg_ppd_name(pwg_media_t *pwg, char *buf, size_t bufsize)
{
if (pwg->ppd)
return (pwg->ppd);
pwg_ppdize_name(pwg->pwg, buf, bufsize);
return (buf);
}
Use pwg_ppd_name(pwg, ppdtemp, sizeof(ppdtemp)) in place of pwg->ppd at the eight places (ppdtemp being a local char[PPD_MAX_NAME]). With that, the PPD generated for the Deskjet has all 20 sizes including Jpn_chou2_111.1x146mm, and cupsd no longer crashes. A patch against 2.4.15 is available if useful.
System Information:
- OS and its version: custom Yocto Linux (yocto-5.3.4 / whinlatter), Raspberry Pi 5, aarch64, glibc 2.42, systemd
- Application:
lpoptions(CUPS CLI); also our own client callingcupsCopyDestInfo() - CUPS version: 2.4.15 (crash also reproduced with 2.4.20)
- Printer: HP Deskjet F2400, USB on a Mac and shared over the network (driverless temporary queue on the Linux side)
- Built with the OpenEmbedded recipe, configure options:
--enable-dbus --with-dbusdir=/etc/dbus-1 --enable-browsing --disable-gssapi --enable-debug --disable-relro --enable-libusb --with-system-groups=lpadmin,root,sys,wheel --with-cups-group=lp --with-domainsocket=/run/cups/cups.sock --with-dnssd=avahi --with-systemd=/usr/lib/systemd/system --with-tls=openssl
Additional context
- Attached: attr.txt, the
ipptool --ippserveroutput for the shared queue (see REPORTING_ISSUES.md). Hostname, location, UUID and the printer's serial number are replaced; all media attributes are unchanged. - I searched existing issues and PRs and didn't find this reported. #760 is related (PPD generation failing for a temporary queue) but is a different problem.
- Lingua principale
- C
- Stelle
- 1.8k
- Fork
- 362
- Merge medio
- 19h 12m
- PR unite (30g)
- 8
Preparare l'ambiente
- Include un Dockerfile o un file Docker Compose
- Nessun modello di pull request
- Leggi la guida per i contributori
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di OpenPrinting/cups
-
NULL pointer dereference in `print_job()` during MIME auto-detectionForse già presa Una pull request collegata a questa issue è aperta o già unita. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
OpenPrinting/cups#1826 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
OAuth (bearer) quota check denies every user when the allowed-user list is emptyForse già presa @RonitMehta08 l’ha presa 3 giorni fa. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 66/100
OpenPrinting/cups#1795 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
cupsdGetPrivateAttrs: @ACL never grants access, and @OWNER is looked up as an operating-system group for non-ownersForse già presa @user-Param l’ha presa 10 giorni fa. Aperta
Difficoltà 3/5 1-2 giorni Idoneità per principianti 35/100
OpenPrinting/cups#1796 ·
I maintainer di solito rispondono entro 1 giorno
-
OAuth/OIDC token-validation gaps: nonce not enforced when absent, at_hash alg, inbound bearer not signature/aud/exp-validated + unverified email used for authzForse di nuovo libera @michaelrsweet l’ha presa 148 giorni fa e non c’è nessuna pull request aperta. Apertainvestigating
OpenPrinting/cups#1609 · 1 assegnatario ·
I maintainer di solito rispondono entro 1 giorno
-
enhancement
Difficoltà 4/5 3-5 giorni Idoneità per principianti 52/100
OpenPrinting/cups#1572 ·
I maintainer di solito rispondono entro 1 giorno
Tutte le issue di OpenPrinting/cups
Issue simili
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 80/100
ESPresense/ESPresense#2680 · 2 commenti ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 74/100
ImageMagick/ImageMagick#9134 ·
I maintainer di solito rispondono entro 1 giorno
-
Builder: a structure opened without calculation data shows the Diffraction (fibre) view by defaultApertabug
Difficoltà 2/5 1-3 ore Idoneità per principianti 65/100
FriendsofECCE/ECCE#1245 · 1 commento ·
-
Arch: ARM/AArch64 Doc
Difficoltà 1/5 Meno di un'ora Idoneità per principianti 90/100
RT-Thread/rt-thread#12082 · 1 commento ·
I maintainer di solito rispondono entro 2 giorni
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
apache/nuttx-apps#4108 · 4 commenti ·
I maintainer di solito rispondono entro 1 giorno