From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate.crashing.org (gate.crashing.org [63.228.1.57]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 3z9Txm3YDRzDqy4 for ; Tue, 2 Jan 2018 07:56:12 +1100 (AEDT) Message-ID: <1514840151.2743.127.camel@kernel.crashing.org> Subject: Re: [PATCH] powerpc/mm: restore SEGV_ACCERR for segv on mapped region From: Benjamin Herrenschmidt To: John Sperbeck , Paul Mackerras , Michael Ellerman , linuxppc-dev@lists.ozlabs.org Date: Tue, 02 Jan 2018 07:55:51 +1100 In-Reply-To: <20180101052458.132027-1-jsperbeck@google.com> References: <20180101052458.132027-1-jsperbeck@google.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Sun, 2017-12-31 at 21:24 -0800, John Sperbeck wrote: > A recent refactoring of the powerpc page fault handler caused > access to protected memory regions to indicate SEGV_MAPERR instead > of the traditional SEGV_ACCERR in the si_code field of a user-space > signal handler. This can confuse debug libraries that temporarily > change the protection of memory regions, and expect to use > SEGV_ACCERR as an indication to restore access to a region. > > This change restores the previous behavior. The following program > exhibits the issue: Good catch ! Acked-by: Benjamin Herrenschmidt > > $ ./repro read || echo "FAILED" > $ ./repro write || echo "FAILED" > $ ./repro exec || echo "FAILED" > > #include > #include > #include > #include > #include > #include > #include > > static void segv_handler(int n, siginfo_t *info, void *arg) { > _exit(info->si_code == SEGV_ACCERR ? 0 : 1); > } > > int main(int argc, char **argv) > { > void *p = NULL; > struct sigaction act = { > .sa_sigaction = segv_handler, > .sa_flags = SA_SIGINFO, > }; > > assert(argc == 2); > p = mmap(NULL, getpagesize(), > (strcmp(argv[1], "write") == 0) ? PROT_READ : 0, > MAP_PRIVATE|MAP_ANONYMOUS, -1, 0); > assert(p != MAP_FAILED); > > assert(sigaction(SIGSEGV, &act, NULL) == 0); > if (strcmp(argv[1], "read") == 0) > printf("%c", *(unsigned char *)p); > else if (strcmp(argv[1], "write") == 0) > *(unsigned char *)p = 0; > else if (strcmp(argv[1], "exec") == 0) > ((void (*)(void))p)(); > return 1; /* failed to generate SEGV */ > } > > Signed-off-by: John Sperbeck > --- > arch/powerpc/mm/fault.c | 7 ++++++- > 1 file changed, 6 insertions(+), 1 deletion(-) > > diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c > index 4797d08581ce..6e1e39035380 100644 > --- a/arch/powerpc/mm/fault.c > +++ b/arch/powerpc/mm/fault.c > @@ -145,6 +145,11 @@ static noinline int bad_area(struct pt_regs *regs, unsigned long address) > return __bad_area(regs, address, SEGV_MAPERR); > } > > +static noinline int bad_access(struct pt_regs *regs, unsigned long address) > +{ > + return __bad_area(regs, address, SEGV_ACCERR); > +} > + > static int do_sigbus(struct pt_regs *regs, unsigned long address, > unsigned int fault) > { > @@ -490,7 +495,7 @@ static int __do_page_fault(struct pt_regs *regs, unsigned long address, > > good_area: > if (unlikely(access_error(is_write, is_exec, vma))) > - return bad_area(regs, address); > + return bad_access(regs, address); > > /* > * If for any reason at all we couldn't handle the fault,