Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018, Max Filippov wrote: > On Fri, Sep 28, 2018 at 2:08 PM, Andy Lutomirski wrote: > >> On Sep 28, 2018, at 1:26 PM, Thomas Gleixner wrote: > >> > >>> On Fri, 28 Sep 2018, Max Filippov wrote: > >>> > On Fri, Sep 28, 2018 at 11:02 AM, Andy Lutomirski > wrote: > The

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, Sep 28, 2018 at 2:08 PM, Andy Lutomirski wrote: > > >> On Sep 28, 2018, at 1:26 PM, Thomas Gleixner wrote: >> >>> On Fri, 28 Sep 2018, Max Filippov wrote: >>> On Fri, Sep 28, 2018 at 11:02 AM, Andy Lutomirski wrote: There may be a much nicer solution. Unless I missed som

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

> On Sep 28, 2018, at 1:26 PM, Thomas Gleixner wrote: > >> On Fri, 28 Sep 2018, Max Filippov wrote: >> >>> On Fri, Sep 28, 2018 at 11:02 AM, Andy Lutomirski >>> wrote: >>> There may be a much nicer solution. Unless I missed something, only >>> mips and xtensa even have the possibility of c

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018, Max Filippov wrote: > On Fri, Sep 28, 2018 at 11:02 AM, Andy Lutomirski wrote: > > There may be a much nicer solution. Unless I missed something, only > > mips and xtensa even have the possibility of cmpxchg being missing. > > We could just make those arches supply a futex-d

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, Sep 28, 2018 at 11:02 AM, Andy Lutomirski wrote: > There may be a much nicer solution. Unless I missed something, only > mips and xtensa even have the possibility of cmpxchg being missing. > We could just make those arches supply a futex-detecting helper. In case of xtensa availability o

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018, Andy Lutomirski wrote: > On Fri, Sep 28, 2018 at 7:53 AM Martin Schwidefsky > wrote: > > On Fri, 28 Sep 2018 07:11:44 -0700 Andy Lutomirski > > wrote: > > > > > There’s another way to skin this cat: keep KERNEL_DS but pass a valid > > > pointer. > > > I don’t suppose you r

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, Sep 28, 2018 at 7:53 AM Martin Schwidefsky wrote: > > On Fri, 28 Sep 2018 07:11:44 -0700 > Andy Lutomirski wrote: > > > > On Sep 28, 2018, at 1:42 AM, Thomas Gleixner wrote: > > > > > >> On Fri, 28 Sep 2018, Martin Schwidefsky wrote: > > >> On Fri, 28 Sep 2018 09:12:10 +0200 > > >> Geert

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018 07:11:44 -0700 Andy Lutomirski wrote: > > On Sep 28, 2018, at 1:42 AM, Thomas Gleixner wrote: > > > >> On Fri, 28 Sep 2018, Martin Schwidefsky wrote: > >> On Fri, 28 Sep 2018 09:12:10 +0200 > >> Geert Uytterhoeven wrote: > >>> I don't know if that has happened, and whet

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018, Andy Lutomirski wrote: > > On Sep 28, 2018, at 1:42 AM, Thomas Gleixner wrote: > > > >> On Fri, 28 Sep 2018, Martin Schwidefsky wrote: > >> On Fri, 28 Sep 2018 09:12:10 +0200 > >> Geert Uytterhoeven wrote: > >>> I don't know if that has happened, and whether it would work on

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

> On Sep 28, 2018, at 1:42 AM, Thomas Gleixner wrote: > >> On Fri, 28 Sep 2018, Martin Schwidefsky wrote: >> On Fri, 28 Sep 2018 09:12:10 +0200 >> Geert Uytterhoeven wrote: >>> I don't know if that has happened, and whether it would work on s390 now. >> >> commit 03b8c7b623c80af264c4c8d6111e

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018, Martin Schwidefsky wrote: > On Fri, 28 Sep 2018 09:12:10 +0200 > Geert Uytterhoeven wrote: > > I don't know if that has happened, and whether it would work on s390 now. > > commit 03b8c7b623c80af264c4c8d6111e5c6289933666 > Author: Heiko Carstens > Date: Sun Mar 2 13:09:47

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018 09:12:10 +0200 Geert Uytterhoeven wrote: > Hi Thomas, > > On Fri, Sep 28, 2018 at 8:21 AM Thomas Gleixner wrote: > > On Thu, 27 Sep 2018, Andy Lutomirski wrote: > > > I have a couple questions here: > > > > > > - Is this actually okay on all architectures? That is, are t

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018, Geert Uytterhoeven wrote: > > On Fri, Sep 28, 2018 at 8:21 AM Thomas Gleixner wrote: > > On Thu, 27 Sep 2018, Andy Lutomirski wrote: > > > I have a couple questions here: > > > > > > - Is this actually okay on all architectures? That is, are there > > >cases where we'll

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

Hi Thomas, On Fri, Sep 28, 2018 at 8:21 AM Thomas Gleixner wrote: > On Thu, 27 Sep 2018, Andy Lutomirski wrote: > > I have a couple questions here: > > > > - Is this actually okay on all architectures? That is, are there > >cases where we'll screw up if we fail a USER_DS access this early?

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Fri, 28 Sep 2018, Thomas Gleixner wrote: > On Thu, 27 Sep 2018, Andy Lutomirski wrote: > > I have a couple questions here: > > > > - Is this actually okay on all architectures? That is, are there > >cases where we'll screw up if we fail a USER_DS access this early? > >s390 stands out

Re: [PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

On Thu, 27 Sep 2018, Andy Lutomirski wrote: > I have a couple questions here: > > - Is this actually okay on all architectures? That is, are there >cases where we'll screw up if we fail a USER_DS access this early? >s390 stands out as the obvious special case (where USER_DS is not >t

[PATCH] futex: Set USER_DS for the futex_detect_cmpxchg() test

futex_detect_cmpxchg() checks whether cmpxchg is available by trying it on the NULL pointer and seeing what the error code is (EFAULT vs ENOSYS). This happens with KERNEL_DS set, which is impolite: while the NULL *user* pointer is definitely invalid when there is no user program running, the NULL