Hi Alexander On Mon, 2005-04-25 at 21:57 +0200, Alexander Nyberg wrote: > mån 2005-04-25 klockan 21:12 +0200 skrev [email protected]: > > From: Andree Leidenfrost <[email protected]>, Paolo 'Blaisorblade' Giarrusso <[email protected]> > > > > Fix the error path, which is triggered when the processor misses the fpx regs > > (i.e. the "fxsr" cpuinfo feature). For instance by VIA C3 Samuel2. Tested and > > obvious, please merge ASAP. > > > > Signed-off-by: Paolo 'Blaisorblade' Giarrusso <[email protected]> > > --- > > > > linux-2.6.12-paolo/arch/um/os-Linux/sys-i386/registers.c | 7 ++++--- > > 1 files changed, 4 insertions(+), 3 deletions(-) > > > > diff -puN arch/um/os-Linux/sys-i386/registers.c~uml-fix-no_fpx_regs_handling arch/um/os-Linux/sys-i386/registers.c > > --- linux-2.6.12/arch/um/os-Linux/sys-i386/registers.c~uml-fix-no_fpx_regs_handling 2005-04-25 21:03:11.000000000 +0200 > > +++ linux-2.6.12-paolo/arch/um/os-Linux/sys-i386/registers.c 2005-04-25 21:08:07.000000000 +0200 > > @@ -105,14 +105,15 @@ void init_registers(int pid) > > panic("check_ptrace : PTRACE_GETREGS failed, errno = %d", > > err); > > > > + errno = 0; > > err = ptrace(PTRACE_GETFPXREGS, pid, 0, exec_fpx_regs); > > if(!err) > > return; > > + if(errno != EIO) > > + panic("check_ptrace : PTRACE_GETFPXREGS failed, errno = %d", > > + errno); > > Looks like you mean "if (err != EIO)" here No. The patch is correct. ptrace will always return -1 in case of an error. The actual error code is in errno, hence this is what needs to be compared to EIO. Please also see the ptrace manpage. > > have_fpx_regs = 0; > > - if(err != EIO) > > - panic("check_ptrace : PTRACE_GETFPXREGS failed, errno = %d", > > - err); > > > > err = ptrace(PTRACE_GETFPREGS, pid, 0, exec_fp_regs); > > if(err) Best regards Andree -- Andree Leidenfrost Sydney - Australia
Attachment:
signature.asc
Description: This is a digitally signed message part
- References:
- Prev by Date: Re: [PATCH] private mounts
- Next by Date: Re: [PATCH] private mounts
- Previous by thread: Re: [patch 1/1] uml: fix handling of no fpx_regs [critical, for 2.6.12]
- Next by thread: [patch 1/1] uml ubd: handle readonly status
- Index(es):