Re: Linux 2.2.19pre10
Hi Alan,
On Tue, 13 Feb 2001, Alan Cox wrote:
>> Yes, I understand that. But I never got any note that my fix is
>> broken and I still do not understand what's the concern.
>
> Unless Im misreading the code the segment you poke at has
> potentially been freed before it is written too.
Oh yes I was blind, shame on me. Here comes a fixed version.
Greetings
Christoph
--- 2.2.19-pre10/ipc/shm.c.orig Tue Feb 13 14:35:25 2001
+++ 2.2.19-pre10/ipc/shm.c Tue Feb 13 14:34:49 2001
@@ -337,6 +337,8 @@
if (current->euid == shp->u.shm_perm.uid ||
current->euid == shp->u.shm_perm.cuid ||
capable(CAP_SYS_ADMIN)) {
+ /* Do not find it any more */
+ shp->u.shm_perm.key = IPC_PRIVATE;
shp->u.shm_perm.mode |= SHM_DEST;
if (shp->u.shm_nattch <= 0)
killseg (id);
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
Hi Alan, On Tue, 13 Feb 2001, Alan Cox wrote: >> No, I do not think that it's minor. We had to bring down running >> application servers to be able to start another one, because the >> new one couldn't create or attach the systemwide os-monitoring >> segment and thus refused to start. That's very bad behaviour. > > Well I'll take corrected fixes, but Im not going to hold up a release for it Yes, I understand that. But I never got any note that my fix is broken and I still do not understand what's the concern. We are holding the BKL while doing this. And if shm_close does not get called with it we should probably acquire it. Greetings Christoph - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
> Yes, I understand that. But I never got any note that my fix is broken > and I still do not understand what's the concern. Unless Im misreading the code the segment you poke at has potentially been freed before it is written too. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
> > Possibly but its a minor item that doesnt really matter anyway so leaving it > > is fine > > No, I do not think that it's minor. We had to bring down running > application servers to be able to start another one, because the new > one couldn't create or attach the systemwide os-monitoring > segment and thus refused to start. That's very bad behaviour. Well I'll take corrected fixes, but Im not going to hold up a release for it - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
Hi Alan, On Mon, 12 Feb 2001, Alan Cox wrote: >> First, I'm glad I wasn't hallucinating, and that the mail did >> indeed get seen by someone. >> >> Second, instead of reverting, can't we simply move those two lines >> up a bit: > > Possibly but its a minor item that doesnt really matter anyway so leaving it > is fine No, I do not think that it's minor. We had to bring down running application servers to be able to start another one, because the new one couldn't create or attach the systemwide os-monitoring segment and thus refused to start. That's very bad behaviour. Greetings Christoph - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
Hi Alan, On Mon, 12 Feb 2001, Alan Cox wrote: First, I'm glad I wasn't hallucinating, and that the mail did indeed get seen by someone. Second, instead of reverting, can't we simply move those two lines up a bit: Possibly but its a minor item that doesnt really matter anyway so leaving it is fine No, I do not think that it's minor. We had to bring down running application servers to be able to start another one, because the new one couldn't create or attach the systemwide os-monitoring segment and thus refused to start. That's very bad behaviour. Greetings Christoph - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
Possibly but its a minor item that doesnt really matter anyway so leaving it is fine No, I do not think that it's minor. We had to bring down running application servers to be able to start another one, because the new one couldn't create or attach the systemwide os-monitoring segment and thus refused to start. That's very bad behaviour. Well I'll take corrected fixes, but Im not going to hold up a release for it - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
Hi Alan, On Tue, 13 Feb 2001, Alan Cox wrote: No, I do not think that it's minor. We had to bring down running application servers to be able to start another one, because the new one couldn't create or attach the systemwide os-monitoring segment and thus refused to start. That's very bad behaviour. Well I'll take corrected fixes, but Im not going to hold up a release for it Yes, I understand that. But I never got any note that my fix is broken and I still do not understand what's the concern. We are holding the BKL while doing this. And if shm_close does not get called with it we should probably acquire it. Greetings Christoph - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
Yes, I understand that. But I never got any note that my fix is broken and I still do not understand what's the concern. Unless Im misreading the code the segment you poke at has potentially been freed before it is written too. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10
Hi Alan,
On Tue, 13 Feb 2001, Alan Cox wrote:
Yes, I understand that. But I never got any note that my fix is
broken and I still do not understand what's the concern.
Unless Im misreading the code the segment you poke at has
potentially been freed before it is written too.
Oh yes I was blind, shame on me. Here comes a fixed version.
Greetings
Christoph
--- 2.2.19-pre10/ipc/shm.c.orig Tue Feb 13 14:35:25 2001
+++ 2.2.19-pre10/ipc/shm.c Tue Feb 13 14:34:49 2001
@@ -337,6 +337,8 @@
if (current-euid == shp-u.shm_perm.uid ||
current-euid == shp-u.shm_perm.cuid ||
capable(CAP_SYS_ADMIN)) {
+ /* Do not find it any more */
+ shp-u.shm_perm.key = IPC_PRIVATE;
shp-u.shm_perm.mode |= SHM_DEST;
if (shp-u.shm_nattch = 0)
killseg (id);
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.19pre10:x!
> Ah, but it does matter ! We break compatibility with other systems (and > our manpages, and possibly standards) if we don't mark the segment > IPC_PRIVATE upon removal - This being so midbogglingly critical a bug that nobody noticed until 2.2.18 and the bug existed since 1.2 or earlier. I think I can wait for 2.2.20 - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://vger.kernel.org/lkml/
Re: Linux 2.2.19pre10
> First, I'm glad I wasn't hallucinating, and that the mail did indeed get > seen by someone. > > Second, instead of reverting, can't we simply move those two lines up a > bit: Possibly but its a minor item that doesnt really matter anyway so leaving it is fine - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://vger.kernel.org/lkml/
Linux 2.2.19pre10
2.2.19pre10 o Update aic7xxx driver to 5.1.33 (Doug Ledford) o Revert shm change - its unsafe (Richard Nelson) o Update sunrpc code, add rpc ping congestion (Trond Myklebust) checks o Fix wrong kfree in cosa driver (Jan Kasprzak) o NFS client fixes(Trond Myklebust) o Better dcache/inode hashes (Dave Miller) o Fix missing skb->protocol init in AX.25 (Thomas Osterried) o EEpro100 reporting fix as per 2.4 (Ion Badulescu) o Starfire ethernet driver(Don Becker, Ion Badulescu, Jeff Garzik, ...) o Memory handling fixes for ISDN core code(Kai Germaschewski) o ISDN module locking fixes (Kai Germaschewski) o Fix ISDN modem profile reading (Kai Germaschewski) o Fix missing mark_bh calls in isdn (Kai Germaschewski) o Fix problems make xconfig has with config (Andrzej Krzysztofowicz) o Clean up isdn to user new __init etc(Kai Germaschewski) 2.2.19pre9 o Merge all the pending NFS server fixes (Neil Brown) o Neil becomes NFS server maintainer (Neil Brown) o Update to aic7xxx 5.1.32(Doug Ledford) o Fix cs89x0 media selection (Frank Copeland) o Tidy APM stuff, make buggy bios selector tighter(Stephen Rothwell) o Fix i2o config typo (YOSHIMURA Keitaro) o Network updates, fix possible classifier hang (Dave Miller) o Sparc updates (nfs compat, syscalls)(Dave Miller) o Sparc watchdog driver (Eric Brower) o Remove experimental tag on QoS code (Dave Miller) o Move dumpable extra logic into binfmt avoiding (Solar Designer) other changes to arch code. Back out old stuff o Fix sysctl miscastings from signed/unsigned (Greg Kroah-Hartmann) o Alpha OSF syscall remove error printk o Don't trust IRQ routing on the ruffian ARC (Ivan Kokshaysky) 2.2.19pre8 o Add support for ICS1893 PHY to sis900 (L C Chang) o Fix typo in nautilus code (Tom Vier) o Clean up usb bandwidth messages (Randy Dunlap) o USB ACM loosen up end point rules (Randy Dunlap) o Fix tty module count corruptions(Maciej Rozycki) o i2o block updates (Boji Kannanthanam) o menuconfig updates (Kirk Reiser) o Fix dmi/apm ordering bug(Keith Owens, Neale Banks) o Alpha SMP build fix (Herbert Xu) o Fix igmp bugs (Stefan Jonsson) o Fix USB config.in problems (Greg Kroah-Hartmann) o Update Cort Dougan's info (Cort Dougan) o Update to 2.4.0 style A20 gate handler (Randy Dunlap) o Fix unneeded compat defines on S/390 ctc(Kurt Roeckx) o Macintosh HID driver fixes (Cort Dougan) o Fix ppc config/input layer and ksyms(Cort Dougan) o ISDN updates(Kai Germaschewski) o TGAfb as a module (Andrzej Krzysztofowicz) o Syscall table updates for sparc64 (Ben Collins) o 8139too driver updates (Jens David) o Tighten packet length checks in masq/tproxy (Julian Anastasov) o Fix udp port selection hang (Dave Miller) 2.2.19pre7 o Remove dead arm files (Russell King) o Fix VIA rhine build failure for a few folks (Peter Monta) o ARM ptrace fixes(Russell King) o Fix ymfpci setup for legacy devices (Pete Zaitcev) o xspeed dsl needs pci(Lars Holmberg) o Typo fix(Dave Miller) o Update ftdi usb serial driver (Greg Kroah-Hartmann) o Update keyspan usb serial drivers (Greg Kroah-Hartmann) o Sparc updates (Dave Miller) o Remove incorrect lp printk (Tim Waugh) o Fix ppa panic on timeout(Tim Waugh) o Maestro3 needs ac97 codec (Oleg Krivosheev) o Fix kwhich versus old bash (Pete Zaitcev) o Fix ip checksum compiler behaviour assumption (Dave
Linux 2.2.19pre10
2.2.19pre10 o Update aic7xxx driver to 5.1.33 (Doug Ledford) o Revert shm change - its unsafe (Richard Nelson) o Update sunrpc code, add rpc ping congestion (Trond Myklebust) checks o Fix wrong kfree in cosa driver (Jan Kasprzak) o NFS client fixes(Trond Myklebust) o Better dcache/inode hashes (Dave Miller) o Fix missing skb-protocol init in AX.25 (Thomas Osterried) o EEpro100 reporting fix as per 2.4 (Ion Badulescu) o Starfire ethernet driver(Don Becker, Ion Badulescu, Jeff Garzik, ...) o Memory handling fixes for ISDN core code(Kai Germaschewski) o ISDN module locking fixes (Kai Germaschewski) o Fix ISDN modem profile reading (Kai Germaschewski) o Fix missing mark_bh calls in isdn (Kai Germaschewski) o Fix problems make xconfig has with config (Andrzej Krzysztofowicz) o Clean up isdn to user new __init etc(Kai Germaschewski) 2.2.19pre9 o Merge all the pending NFS server fixes (Neil Brown) o Neil becomes NFS server maintainer (Neil Brown) o Update to aic7xxx 5.1.32(Doug Ledford) o Fix cs89x0 media selection (Frank Copeland) o Tidy APM stuff, make buggy bios selector tighter(Stephen Rothwell) o Fix i2o config typo (YOSHIMURA Keitaro) o Network updates, fix possible classifier hang (Dave Miller) o Sparc updates (nfs compat, syscalls)(Dave Miller) o Sparc watchdog driver (Eric Brower) o Remove experimental tag on QoS code (Dave Miller) o Move dumpable extra logic into binfmt avoiding (Solar Designer) other changes to arch code. Back out old stuff o Fix sysctl miscastings from signed/unsigned (Greg Kroah-Hartmann) o Alpha OSF syscall remove error printk o Don't trust IRQ routing on the ruffian ARC (Ivan Kokshaysky) 2.2.19pre8 o Add support for ICS1893 PHY to sis900 (L C Chang) o Fix typo in nautilus code (Tom Vier) o Clean up usb bandwidth messages (Randy Dunlap) o USB ACM loosen up end point rules (Randy Dunlap) o Fix tty module count corruptions(Maciej Rozycki) o i2o block updates (Boji Kannanthanam) o menuconfig updates (Kirk Reiser) o Fix dmi/apm ordering bug(Keith Owens, Neale Banks) o Alpha SMP build fix (Herbert Xu) o Fix igmp bugs (Stefan Jonsson) o Fix USB config.in problems (Greg Kroah-Hartmann) o Update Cort Dougan's info (Cort Dougan) o Update to 2.4.0 style A20 gate handler (Randy Dunlap) o Fix unneeded compat defines on S/390 ctc(Kurt Roeckx) o Macintosh HID driver fixes (Cort Dougan) o Fix ppc config/input layer and ksyms(Cort Dougan) o ISDN updates(Kai Germaschewski) o TGAfb as a module (Andrzej Krzysztofowicz) o Syscall table updates for sparc64 (Ben Collins) o 8139too driver updates (Jens David) o Tighten packet length checks in masq/tproxy (Julian Anastasov) o Fix udp port selection hang (Dave Miller) 2.2.19pre7 o Remove dead arm files (Russell King) o Fix VIA rhine build failure for a few folks (Peter Monta) o ARM ptrace fixes(Russell King) o Fix ymfpci setup for legacy devices (Pete Zaitcev) o xspeed dsl needs pci(Lars Holmberg) o Typo fix(Dave Miller) o Update ftdi usb serial driver (Greg Kroah-Hartmann) o Update keyspan usb serial drivers (Greg Kroah-Hartmann) o Sparc updates (Dave Miller) o Remove incorrect lp printk (Tim Waugh) o Fix ppa panic on timeout(Tim Waugh) o Maestro3 needs ac97 codec (Oleg Krivosheev) o Fix kwhich versus old bash (Pete Zaitcev) o Fix ip checksum compiler behaviour assumption (Dave
Re: Linux 2.2.19pre10
First, I'm glad I wasn't hallucinating, and that the mail did indeed get seen by someone. Second, instead of reverting, can't we simply move those two lines up a bit: Possibly but its a minor item that doesnt really matter anyway so leaving it is fine - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://vger.kernel.org/lkml/
Re: Linux 2.2.19pre10:x!
Ah, but it does matter ! We break compatibility with other systems (and our manpages, and possibly standards) if we don't mark the segment IPC_PRIVATE upon removal - This being so midbogglingly critical a bug that nobody noticed until 2.2.18 and the bug existed since 1.2 or earlier. I think I can wait for 2.2.20 - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://vger.kernel.org/lkml/
